The English Language.

LATE FELLOW OF KING'S COLLEGE, CAMBRIDGE; FELLOW OF THE ROYAL COLLEGE OF
PHYSICIANS, LONDON; MEMBER OF THE ETHNOLOGICAL SOCIETY,
NEW YORK; LATE PROFESSOR OF THE ENGLISH LANGUAGE
AND LITERATURE, UNIVERSITY COLLEGE, LONDON.

THIRD EDITION,

REVISED AND GREATLY ENLARGED.

LONDON:

TAYLOR, WALTON, AND MABERLY,

1850.

THE REV. WILLIAM BUTCHER, M.A.,

THE AUTHOR.

PREFACE

SECOND EDITION.

The first edition of the present work was laid before the public, with the intention of representing in a form as systematic as the extent of the subject would allow, those views concerning the structure and relations of the English language, which amongst such scholars as had studied them with the proper means and opportunities, were then generally received; and which, so being received, might take their stand as established and recognized facts. With the results of modern criticism, as applied to his native tongue, it was conceived that an educated Englishman should be familiar. To this extent the special details of the language were exhibited; and to this extent the work was strictly a Grammar of the English Language.

But besides this, it was well known that the current grammarians, and the critical philologists, had long ceased to write alike upon the English, or {vi}indeed upon any other, language. For this reason the sphere of the work became enlarged; so that, on many occasions, general principles had to be enounced, fresh terms to be defined, and old classifications to be remodelled. This introduced extraneous elements of criticism, and points of discussion which, in a more advanced stage of English philology, would have been superfluous. It also introduced elements which had a tendency to displace the account of some of the more special and proper details of the language. There was not room for the exposition of general principles, for the introduction of the necessary amount of preliminary considerations, and for the minutiæ of an extreme analysis. Nor is there room for all this at present. A work that should, at one and the same time, prove its principles, instead of assuming them, supply the full and necessary preliminaries in the way of logic, phonetics, and ethnology, and, besides this, give a history of every variety in the form of every word, although, perhaps, a work that one man might write, would be a full and perfect Thesaurus of the English Language, and, would probably extend to many volumes. For, in the English language, there are many first principles to be established, and much historical knowledge to be applied. Besides which, the particular points both of etymology and syntax are far more numerous than is imagined. Scanty as is the amount of declension and conjugation in current use, there are to be found in every department of our grammars, {vii}numerous isolated words which exhibit the fragments of a fuller inflection, and of a more highly developed etymology. This is well-known to every scholar who has not only viewed our language as a derivative of the Anglo-Saxon, and observed that there are similar relations between many other languages (e. g. the Italian and Latin, the German and Mœso-Gothic, &c.), but who has, also, generalized the phenomena of such forms of relationship and derivation, and enabled himself to see in the most uninflected languages of the nineteenth century, the fragments of a fuller and more systematic inflection, altered by time, but altered in a uniform and a general manner.

The point, however, upon which, in the prefaces both of the first edition of the present work and of his English Grammar, the writer has most urgently insisted is the disciplinal character of grammatical studies in general, combined with the fact, that the grammatical study of one's own language is almost exclusively disciplinal. It is undoubtedly true, that in schools something that is called English Grammar is taught: and it is taught pretty generally. It is taught so generally that, I believe, here are only two classes of English boys and girls who escape it—those who are taught nothing at all in any school whatever, and those who are sent so early to the great classical schools (where nothing is taught but Latin and Greek), as to escape altogether the English part of their scholastic education. But {viii}what is it that is thus generally taught? not the familiar practice of speaking English—that has been already attained by the simple fact of the pupil having been born on English soil, and of English parents. Not the scientific theory of the language—that is an impossibility with the existing text-books. Neither, then, of these matters is taught. Nevertheless labour is expended, and time is consumed. What is taught? Something undoubtedly. The facts, that language is more or less regular (i. e. capable of having its structure exhibited by rules); that there is such a thing as grammar; and that certain expressions should be avoided, are all matters worth knowing. And they are all taught even by the worst method of teaching. But are these the proper objects of systematic teaching? Is the importance of their acquisition equivalent to the time, the trouble, and the displacement of more valuable subjects, which are involved in their explanation? I think not. Gross vulgarity of language is a fault to be prevented; but the proper prevention is to be got from habit—not rules. The proprieties of the English language are to be learned, like the proprieties of English manners, by conversation and intercourse; and the proper school for both, is the best society in which the learner is placed. If this be good, systematic teaching is superfluous; if bad, insufficient. There are undoubted points where a young person may doubt as to the grammatical propriety of a certain expression. In this case let him ask some one older, and more instructed. Grammar, {ix}as an art, is, undoubtedly, the art of speaking and writing correctly—but then, as an art, it is only required for foreign languages. For our own we have the necessary practice and familiarity.

The claim of English grammar to form part and parcel of an English education stands or falls with the value of the philological and historical knowledge to which grammatical studies may serve as an introduction, and with the value of scientific grammar as a disciplinal study. I have no fear of being supposed to undervalue its importance in this respect. Indeed in assuming that it is very great, I also assume that wherever grammar is studied as grammar, the language which the grammar so studied should represent, must be the mother-tongue of the student; whatever that mother-tongue may be—English for Englishmen, Welsh for Welshmen, French for Frenchmen, German for Germans, &c. This study is the study of a theory; and for this reason it should be complicated as little as possible by points of practice. For this reason a man's mother-tongue is the best medium for the elements of scientific philology, simply because it is the one which he knows best in practice.

Now if, over and above the remarks upon the English language, and the languages allied to it, there occur in the present volume, episodical discussions of points connected with other languages, especially the Latin and Greek, it is because a greater portion of the current ideas on philological subjects {x}is taken from those languages than from our own. Besides which, a second question still stands over. There is still the question as to the relative disciplinal merits of the different non-vernacular languages of the world. What is the next best vehicle for philological philosophy to our mother-tongue, whatever that mother-tongue maybe? Each Athenian who fought at Salamis considered his own contributions to that great naval victory the greatest; and he considered them so because they were his own. So it is with the language which we speak, and use, and have learned as our own. Yet each same Athenian awarded the second place of honour to Themistocles. The great classical languages of Greece and Rome are in the position of Themistocles. They are the best when the question of ourselves and our possessions is excluded. They are the best in the eyes of an indifferent umpire. More than this; if we take into account the studies of the learned world, they are second only to the particular mother-tongue of the particular student, in the way of practical familiarity. Without either affirming or denying that, on the simple scores of etymological regularity, etymological variety, and syntactic logic, the Sanskrit may be their equal, it must still be admitted that this last-named language has no claims to a high value as a practical philological discipline upon the grounds of its universality as a point of education; nor will it have. Older than the Greek, it may (or may not) be; more multiform than the Latin, it may (or may not) be: but equally rich in the attractions {xi}of an unsurpassed literature, and equally influential as a standard of imitation, it neither has been nor can be. We may admit all that is stated by those who admire its epics, or elucidate its philosophy; we may admire all this and much more besides, but we shall still miss the great elements of oratory and history, that connect the ancient languages of Greece and Italy with the thoughts, and feelings, and admiration of recent Europe.

The same sort of reasoning applies to the Semitic languages. One element they have, in their grammatical representation, which gives them a value in philological philosophy, in the abstract, above all other languages—the generality of the expression of their structure. This is symbolic, and its advantage is that it exhibits the naturally universal phenomena of their construction in a universal language. Yet neither this nor their historical value raises them to the level of the classical languages.

Now, what has just been written has been written with a view towards a special inference, and as the preliminary to a practical deduction; and it would not have been written but for some such ulterior application. If these languages have so high a disciplinal value, how necessary it is that the expression of their philological phenomena should be accurate, scientific, and representative of their true growth and form? How essential that their grammars should exhibit nothing that may hereafter be unlearned? Pace grammaticorum dixerim, this is not the case. Bad {xii}as is Lindley Murray in English, Busby and Lilly are worse in Greek and Latin. This is the comparison of the men on the low rounds of the ladder. What do we find as we ascend? Is the grammatical science of even men like Mathiæ and Zump much above that of Wallis? Does Buttmann's Greek give so little to be unlearned as Grimm's German? By any one who has gone far in comparative philology, the answer will be given in the negative.

This is not written in the spirit of a destructive criticism. If an opinion as to the fact is stated without reserve, it is accompanied by an explanation, and (partially, perhaps) by a justification. It is the business of a Greek and Latin grammarian to teach Greek and Latin cito, tute, ac jucunde,—cito, that is, between the years of twelve and twenty-four; tute, that is, in a way that quantities may be read truly, and hard passages translated accurately; jucunde, that is, as the taste and memory of the pupil may determine. With this view the grammar must be artificial. Granted. But then it should profess to be so. It should profess to address the memory only, not the understanding. Above all it should prefer to leave a point untaught, than to teach it in a way that must be unlearned.

In 1840, so little had been done by Englishmen for the English language, that in acknowledging my great obligations to foreign scholars, I was only able to speak to what might be done by my own countrymen. Since then, however, there has been a good {xiii}beginning of what is likely to be done well. My references to the works of Messrs. Kemble, Garnet, and Guest, show that my authorities are now as much English as German. And this is likely to be the case. The details of the syntax, the illustrations drawn from our provincial dialects, the minute history of individual words, and the whole system of articulate sounds can, for the English, only be done safely by an Englishman: or, to speak more generally, can, for any language, only be dealt with properly by the grammarian whose mother-tongue is that language. The Deutsche Grammatik of Grimm is the work not of an age nor of a century, but, like the great history of the Athenian, a κτῆμα εἰς ἀεί. It is the magazine from whence all draw their facts and illustrations. Yet it is only the proper German portion that pretends to be exhaustive. The Dutch and Scandinavians have each improved the exhibition of their own respective languages. Monument as is the Deutsche Grammatik of learning, industry, comprehensiveness, and arrangement, it is not a book that should be read to the exclusion of others: nor must it be considered to exhibit the grammar of the Gothic languages, in a form unsusceptible of improvement. Like all great works, it is more easily improved than imitated. One is almost unwilling to recur to the old comparison between Aristotle, who absorbed the labour of his predecessors, and the Eastern sultans, who kill-off their younger brothers. But such is the case with Grimm and his fore-runners in philology. Germany, that, in {xiv}respect to the Reformation, is content to be told that Erasmus laid the egg which Luther hatched, must also acknowledge that accurate and systematic scholars of other countries prepared the way for the Deutsche Grammatik,—Ten Kate in Holland; Dowbrowsky, a Slavonian; and Rask, a Dane.

Nor are there wanting older works in English that have a value in Gothic philology. I should be sorry to speak as if, beyond the writers of what may be called the modern school of philology, there was nothing for the English grammarian both to read and study. The fragments of Ben Jonson's English Grammar are worth the entireties of many later writers. The work of Wallis is eminently logical and precise. The voice of a mere ruler of rules is a sound to flee from; but the voice of a truly powerful understanding is a thing to be heard on all matters. It is this which gives to Cobbett and Priestley, to Horne Tooke as a subtle etymologist, and to Johnson as a practical lexicographer, a value in literary history, which they never can have in grammar. It converts unwholesome doctrines into a fertile discipline of thought.

The method of the present work is mixed. It is partly historical, and partly logical. The historical portions exhibit the way in which words and inflections have been used; the logical, the way in which they ought to be used. Now I cannot conceal from either my readers or myself the fact that philological criticism at the present moment is of an essentially {xv}historical character. It has been by working the historical method that all the great results both in general and special scholarship have been arrived at; and it is on historical investigation that the whole induction of modern philology rests. All beyond is à priori argument; and, according to many, à priori argument out of place. Now, this gives to the questions in philology, to questions concerning the phenomena of concord, government, &c. a subordinate character. It does so, however, improperly. Logic is in language what it is in reasoning,—a rule and standard. But in its application to reasoning and to language there is this difference. Whilst illogical reasoning, and illogical grammar are equally phenomena of the human mind, even as physical disease is a phenomenon of the human body, the illogical grammar can rectify itself by its mere continuance, propagation, and repetition. In this respect the phenomena of language stand apart from the other phenomena of either mind or organized matter. No amount of false argument can make a fallacy other than a fallacy. No amount of frequency can make physical disease other than a predisposing cause to physical disorganization. The argument that halts in its logic, is not on a par with the argument that is sound. Such also is the case with any bodily organ. No prevalence of sickness can ever evolve health. Language, however, as long as it preserves the same amount of intelligibility is always language. Provided it serve as a medium, it does its proper work; {xvi}and as long as it does this, it is, as far as its application is concerned, faultless. Now there is a limit in logical regularity which language is perpetually overstepping; just as there is a logical limit which the reasoning of common life is perpetually overstepping, and just as there is a physiological limit which the average health of men and women may depart from. This limit is investigated by the historical method; which shows the amount of latitude in which language may indulge and yet maintain its great essential of intelligibility. Nay, more, it can show that it sometimes transgresses the limit in so remarkable a manner, as to induce writers to talk about the corruption of a language, or the pathology of a language, with the application of many similar metaphors. Yet it is very doubtful whether all languages, in all their stages, are not equally intelligible, and, consequently, equally what they ought to be, viz., mediums of intercourse between man and man; whilst, in respect to their growth, it is almost certain that so far from exhibiting signs of dissolution, they are, on the contrary, like the Tithonus of mythology, the Strulbrugs of Laputa, or, lastly, such monsters as Frankenstein, very liable to the causes of death, but utterly unable to die. Hence, in language, whatever is, is right; a fact which, taken by itself, gives great value to the historical method of inquiry, and leaves little to the à priori considerations of logic.

But, on the other hand, there is a limit in logical regularity, which language never oversteps: and as {xvii}long as this is the case, the study of the logical standard of what language is in its normal form must go hand in hand with the study of the processes that deflect it. The investigation of the irregularities of language—and be it remembered that almost all change implies original irregularity—is analogous to the investigation of fallacies in logic. It is the comparison between the rule and the practice, with this difference, that in language the practice can change the rule, which in logic is impossible. I am sure that these remarks are necessary in order to anticipate objections that may be raised against certain statements laid down in the syntax. I often write as if I took no account of the historical evidence, in respect to particular uses of particular words. I do so, not because I undervalue that department of philology, but because it is out of place. To show that one or more writers, generally correct, have used a particular expression is to show that they speak, in a few instances, as the vulgar speak in many. To show that the vulgar use one expression for another is to show that two ideas are sufficiently allied to be expressed in the same manner: in other words, the historical fact is accompanied by a logical explanation; and the historical deviation is measured by a logical standard.

I am not desirous of sacrificing a truth to an antithesis, but so certain is language to change from logical accuracy to logical licence, and, at the same time, so certain is language, when so changed, to be {xviii}just as intelligible as before, that I venture upon asserting that, not only whatever is, is right, but also, that in many cases, whatever was, was wrong. There is an antagonism, between logic and practice; and the phenomena on both sides must be studied.

CONTENTS.

AN INTRODUCTION

THE ENGLISH LANGUAGE

PART I.

CHAPTER I.

§ 1. The first point to be remembered in the history of the English Language, is that it was not the original language of any of the British Islands altogether or of any portion of them. Indeed, of the whole of Great Britain it is not the language at the present moment. Welsh is spoken in Wales, Manks in the Isle of Man, Scotch Gaelic in the Highlands of Scotland, and Irish Gaelic in Ireland. Hence, the English that is now spoken was once as foreign to our country as it is at present to the East Indies; and it is no more our primitive vernacular tongue, than it is the primitive vernacular tongue for North America, Jamaica, or Australia. Like the English of Sydney, or the English of Pennsylvania, the English of Great Britain spread itself at the expense of some earlier and more aboriginal language, which it displaced and superseded. {2}

§ 2. The next point involves the real origin and the real affinities of the English Language. Its real origin is on the continent of Europe, and its real affinities are with certain languages there spoken. To speak more specifically, the native country of the English Language is Germany; and the Germanic languages are those that are the most closely connected with our own. In Germany, languages and dialects allied to each other and allied to the mother-tongue of the English have been spoken from times anterior to history; and these, for most purposes of philology, may be considered as the aboriginal languages and dialects of that country.

§ 3. Accredited details of the different immigrations from Germany into Britain.—Until lately the details of the different Germanic invasions of England, both in respect to the particular tribes by which they were made, and the order in which they succeeded each other, were received with but little doubt, and as little criticism.

Respecting the tribes by which they were made, the current opinion was, that they were chiefly, if not exclusively, those of the Jutes, the Saxons, and the Angles.

The particular chieftains that headed each descent were also known, as well as the different localities upon which they descended. These were as follows:—

§ 4. First settlement of invaders from Germany.—The account of this gives us the year 449 for the first permanent Germanic tribes settled in Britain. Ebbsfleet, in the Isle of Thanet, was the spot where they landed; and the particular name that these tribes gave themselves was that of Jutes. Their leaders were Hengist and Horsa. Six years after their landing they had established the kingdom of Kent; so that the county of Kent was the first district where the original British was superseded by the mother-tongue of the present English, introduced from Germany.

§ 5. Second settlement of invaders from Germany.—In the year 477 invaders from Northern Germany made the second permanent settlement in Britain. The coast of Sussex was the spot whereon they landed. The particular name that these tribes gave themselves was that of Saxons. Their leader {3}was Ella. They established the kingdom of the South Saxons (Sussex); so that the county of Sussex was the second district where the original British was superseded by the mother-tongue of the present English, introduced from Northern Germany.

§ 6. Third settlement of invaders from Germany.—In the year 495 invaders from Northern Germany made the third permanent settlement in Britain. The coast of Hampshire was the spot whereon they landed. Like the invaders last mentioned, these tribes were Saxons. Their leader was Cerdic. They established the kingdom of the West Saxons (Wessex); so that the county of Hants was the third district where the original British was superseded by the mother-tongue of the present English, introduced from Northern Germany.

§ 7. Fourth settlement of invaders from Germany.—A.D. 530, certain Saxons landed in Essex, so that the county of Essex was the fourth district where the original British was superseded by the mother-tongue of the present English, introduced from Northern Germany.

§ 8. Fifth settlement of invaders from Germany.—These were Angles in Norfolk and Suffolk. This settlement, of which the precise date is not known, took place during the reign of Cerdic in Wessex. The fifth district where the original British was superseded by the mother-tongue of the present English was the counties of Norfolk and Suffolk; the particular dialect introduced being that of the Angles.

§ 9. Sixth settlement of invaders from Germany.—In the year 547 invaders from Northern Germany made the sixth permanent settlement in Britain. The south-eastern counties of Scotland, between the rivers Tweed and Forth, were the districts where they landed. They were of the tribe of the Angles, and their leader was Ida. The south-eastern parts of Scotland constituted the sixth district where the original British was superseded by the mother-tongue of the present English, introduced from Northern Germany.

§ 10. It would be satisfactory if these details rested upon cotemporary evidence; in which case the next question would {4}be that of the relations of the immigrant tribes to each other as Germans, i.e. the extent to which the Jute differed from (or agreed with) the Angle, or the Saxon, and the relations of the Angle and the Saxon to each other. Did they speak different languages?—different dialects of a common tongue!—or dialects absolutely identical? Did they belong to the same or to different confederations? Was one polity common to all? Were the civilizations similar?

Questions like these being answered, and a certain amount of mutual difference being ascertained, it would then stand over to inquire whether any traces of this original difference were still to be found in the modern English. Have any provincial dialects characteristics which are Jute rather than Angle? or Angle rather than Saxon?

It is clear that the second of these questions is involved in the answer given to the first.

§ 11. The accredited relations of the Jutes, Angles, and Saxons to each other as Germans.—These are as follows:—

2. That that of Angles, was the present Dutchy of Sleswick; so that they were the southern neighbours of the Jutes.

3. That that of the Saxons was a small tract north of the Elbe, and some distinct point—more or less extensive—between the Elbe and Rhine.

4. That, although there were, probably, dialectal differences between the languages, the speech of all the three tribes was mutually intelligible.

§ 12. Assuming, then, the accuracy of our historical facts, the inference is, that, without expecting to find any very prominent and characteristic differences between the different inhabitants of England arising out of the original differences between the Germanic immigrants, we are to look for what few there are in the following quarters—

1. For the characteristic differentiæ of the Jutes, in Kent, part of Sussex, and the Isle of Wight.

3. For those of the Angles in Norfolk, Suffolk, Yorkshire, Durham, and Northumberland.

1. The differentiæ of the people of Kent, part of Sussex, and the Isle of Wight (if any), are to be explained by the differentiæ of the original Jute immigrants—

2. Those of the rest of Sussex, Wessex, Essex, and Middlesex, by those of the Saxons—

Such is our reasoning, and such a sketch of our philological researches—assuming that the opinions just exhibited, concerning the dates, conductors, localities, and order, are absolute and unimpeachable historical facts.

§ 13. Criticism of the aforesaid details.—As a preliminary to this part of the subject, the present writer takes occasion to state once for all, that nearly the whole of the following criticism is not his own (except, of course, so far as he adopts it—which he does), but Mr. Kemble's, and that it forms the introduction to his valuable work on the Saxons in England.

1. The evidence to the details just given, is not historical, but traditional.—a. Bede, from whom it is chiefly taken, wrote more than 300 years after the supposed event, i.e., the landing of Hengist and Horsa, in A.D. 449.

b. The nearest contemporary author is Gildas, and he lived at least 100 years after it.

2. The account of Hengist's and Horsa's landing, has elements which are fictional rather than historical—a. Thus "when we find Hengist and Horsa approaching the coasts of Kent in three keels, and Ælli effecting a landing in Sussex with the same number, we are reminded of the Gothic tradition which carries a migration of Ostrogoths, Visigoths, and Gepidæ, also in three vessels, to the mouths of the Vistula."

b. The murder of the British chieftains by Hengist is told totidem verbis, by Widukind, and others of the Old Saxons in Thuringia.

c. Geoffry of Monmouth relates also, how "Hengist obtained from the Britons as much land as could be enclosed {6}by an ox-hide; then, cutting the hide into thongs, enclosed a much larger space than the granters intended, on which he erected Thong Castle—a tale too familiar to need illustration, and which runs throughout the mythus of many nations. Among the Old Saxons, the tradition is in reality the same, though recorded with a slight variety of detail. In their story, a lap-full of earth is purchased at a dear rate from a Thuringian; the companions of the Saxon jeer him for his imprudent bargain; but he sows the purchased earth upon a large space of ground, which he claims, and, by the aid of his comrades, ultimately wrests it from the Thuringians."

3. There is direct evidence in favour of there having been German tribes in England anterior to A.D. 447.—a. At the close of the Marcomannic war, Marcus Antoninus transplanted a number of Germans into Britain.—Dio Cassius, lxxi. lxiii.

c. The Notitia utriusque imperii, of which the latest date is half a century earlier than the epoch of Hengist, mentions, as an officer of State, the Comes littoris Saxonici per Britannias; his government extending along the coast from Portsmouth to the Wash.

I conclude with the following extract:—"We are ignorant what fasti or even mode of reckoning the revolutions of seasons prevailed in England, previous to the introduction of Christianity. We know not how any event before the year 600 was recorded, or to what period the memory of man extended. There may have been rare annals: there may have been poems: if such there were they have perished, and have left no trace behind, unless we are to attribute to them such scanty notices as the Saxon Chronicle adds to Beda's account. From such sources, however, little could have been gained of accurate information either as to the real internal state, the domestic progress, or development of a people. The dry bare entries of the Chronicles in historical periods may supply the means of judging what sort of annals were likely to exist before the general introduction of the Roman alphabet and parchment, while, in all probability, runes supplied the place of letters, and {7}stones, or the beech-wood, from which their name is derived, of books. Again, the traditions embodied in the epic, are pre-eminently those of kings and princes; they are heroical, devoted to celebrate the divine or half-divine founders of a race, the fortunes of their warlike descendants, the manners and mode of life of military adventurers, not the obscure progress, household peace, and orderly habits of the humble husband-man. They are full of feasts and fighting, shining arms and golden goblets: the gods mingle among men almost their equals, share in the same pursuits, are animated by the same passions of love, and jealousy, and hatred; or, blending the divine with the mortal nature, become the founders of races, kingly, because derived from divinity itself. But one race knows little of another, or its traditions, and cares as little for them. Alliances or wars alone bring them in contact with one another, and the terms of intercourse between the races will, for the most part, determine the character under which foreign heroes shall be admitted into the national epos, or whether they shall be admitted at all. All history, then, which is founded in any degree upon epical tradition (and national history is usually more or less so founded) must be to that extent imperfect, if not inaccurate; only when corrected by the written references of contemporaneous authors, can we assign any certainty to its records.

"Let us apply these observations to the early events of Saxon history: of Kent, indeed, we have the vague and uncertain notices which I have mentioned; even more vague and uncertain are those of Sussex and Wessex. Of the former, we learn that in the year 477, Ælli, with three sons, Cymen, Wlencing, and Cissa, landed in Sussex; that in the year 485 they defeated the Welsh, and that in 491 they destroyed the population of Anderida. Not another word is there about Sussex before the arrival of Augustine, except a late assertion of the military pre-eminence of Ælli among the Saxon chieftains. The events of Wessex are somewhat better detailed; we learn that in 495 two nobles, Cerdic and Cyneríc, came to England, and landed at Cerdices-ora, where, on the {8}same day, they fought a battle: that in 501 they were followed by a noble named Port, who, with his two sons, Bieda and Mægla, made a forcible landing at Portsmouth: and that in 508, they gained a great battle over a British king, whom they slew, together with five thousand of his people. In 514 Stuff and Wihtgár, their nephews, brought them a reinforcement of three ships; in 519, they again defeated the Britons, and established the kingdom of Wessex. In 527, a new victory is recorded; in 530, the Isle of Wight was subdued and given to Wihtgár; and in 534, Cerdic died, and was succeeded by Cyneríc, who reigned twenty-six years. In 544, Wihtgár died. A victory of Cyneríc, in 552 and 556, and Ceawlin's accession to the throne of Wessex are next recorded. Wars of the West-Saxon kings are noted in 568, 571, 577, 584. From 590 to 595, a king of that race, named Ceól, is mentioned: in 591, we learn the expulsion of Ceawlin from power; in 593, the deaths of Ceawlin, Cwichelm, and Crida, are mentioned, and in 597, the year of Augustine's arrival, we learn that Ceólwulf ascended the throne of Wessex.

"Meagre as these details are, they far exceed what is related of Northumberland, Essex, or East-Anglia. In 547, we are told that Ida began to reign in the first of these kingdoms, and that he was succeeded in 560, by Ælli: that after a reign of thirty years, he died in 588, and was succeeded by Æþelríc, who again, in 593, was succeeded by Æþelfriþ. This is all we learn of Northumbria; of Mercia, Essex, East-Anglia, and the innumerable kingdoms that must have been comprised under these general appellations, we hear not a single word.

"If this be all that we can now recover of events, a great number of which must have fallen within the lives of those to whom Augustine preached, what credit shall we give to the inconsistent accounts of earlier actions? How shall we supply the almost total want of information respecting the first settlements? What explanation have we to give of the alliance between Jutes, Angles, and Saxon, which preceded the invasions of England? What knowledge will these records {9}supply of the real number and quality of the chieftains, the language and blood of the populations who gradually spread themselves from the Atlantic to the Frith of Forth; of the remains of Roman cultivation, or the amount of British power with which they had to contend? of the vicissitudes of good and evil fortune which visited the independent principalities before they were swallowed up in the kingdoms of the heptarchy, or the extent of the influence which they retained after the event! On all these several points we are left entirely in the dark; and yet these are facts which it most imports us to know, if we would comprehend the growth of a society which endured for at least 700 years in England, and formed the foundation of that in which we live."—The Saxons in England. Vol. I, pp. 28-32.

§ 14. Inference.—As it is nearly certain, that the year 449 is not the date of the first introduction of German tribes into Britain, we must consider that the displacement of the original British began at an earlier period than the one usually admitted, and, consequently, that it was more gradual than is usually supposed.

Perhaps, if we substitute the middle of the fourth, instead of the middle of the fifth century, as the epoch of the Germanic immigrations into Britain, we shall not be far from the truth.

CHAPTER II.

GERMANIC ORIGIN OF THE ENGLISH LANGUAGE.—THE IMMIGRANT TRIBES, AND THEIR RELATIONS TO EACH OTHER.

§ 15. By referring to §§ 3-12, it may be seen that out of the numerous tribes and nations of Germany, three in particular have been considered as the chief, if not the exclusive, sources of the present English, viz.: the Angles, the Saxons and the Jutes.

To criticise the evidence which derives the English in general from the Angles, the particular inhabitants of Sussex, Essex, Middlesex and Wessex, from the Saxons, and the Anglo-Saxon language from the Angle and Saxon would be superfluous; whilst to doubt the truth of the main facts which it attests would exhibit an unnecessary and unhealthy scepticism. That the Angles and Saxons formed at least seven-tenths of the Germanic invaders may be safely admitted. The Jute element, however, requires further notice.

§ 16. The Jutes.—Were any of the German immigrants Jutes? If so, what were their relations to the other German tribes?

a. Were there Jutes in England? That there was a Jute element in England is to be maintained, not upon the tradition that one of the three ships of Hengist and Horsa was manned by Jutes, but from the following extract from the Anglo-Saxon Chronicle:—

Still there is a flaw in it; since it is quite possible that the term Iútnacynn may have been no true denomination of a section of the Germans of England, but only the synonym of a different word, Wiht-sætan. Alfred writes—comon hi of þrym folcum þam strangestan Germaniæ; þæt of Seaxum, and of Angle, and of Geatum. Of Geatum fruman sindon Cantware and Wiht-sætan, þæt is seo þeód se Wiht þæt ealond on eardað—they came of three folk, the strongest of Germany; that of Saxons and of Angles, and of Geats. Of Geats originally are the Kent people and Wiht-set; that is the people which Wiht the Island live on.

a. The word in question is a compound=Wight=the name of the isle, + sætan=people; as Somer-set, and Dor-set.

c. The wiht- in Wiht-sætan is, undoubtedly, no such element as the vit- in Vit-land=Jut-land; since it represents the older Celtic term, known to us in the Romanized form Vectis.

Putting all this together, it becomes possible (nay probable) that the whole doctrine of a Jute element in the Anglo-Saxon migration may have arisen out of the fact of there being a portion of the people of Southern England neighbours of the Saxons, and bearing the name Wiht-sætan; a fact which, taken along with the juxtaposition of the Vit-landers (Jut-landers) and Saxons on the Continent, suggested to the writers of a long later age the doctrine of a Jute migration.

§ 17. As this last objection impugns the evidence rather than the fact, the following question finds place:— {12}

What were the Jutes of Germany? At present they are the natives of Jutland, and their language is Danish rather than German.

Neither is there reason to suppose that during the third and fourth centuries it was otherwise.

§ 18. This last circumstance detracts from the likelihood of the fact; since in no part of Kent, Sussex, Hants, nor even in the Isle of Wight—a likely place for a language to remain unchanged—have any traces of the old Jute been found.

§ 19. On the other hand the fact of Jutes, even though Danes, being members of a Germanic confederation is not only probable, but such was actually the case; at least for continental wars—subactis, cum Saxonibus, Euciis (Eutiis), qui se nobis (i.e., the Franks), propriâ voluntate tradiderunt ... usque in Oceani littoribus dominio nostro porrigitur.—Theodebert to the Emperor Justinian.—

§ 20. Inference.—Of the three following views—(1.) that the Jutes of Jutland in the fourth and fifth centuries spoke Saxon; (2.) that they spoke Danish at home, but lost their language after three or four centuries' residence in England; and (3.) that a later historian was induced by the similarity between the term Wiht-sætan, as applied to the people of the Isle of Wight, and Wit-land, as applied to Jutland, combined with the real probability of the fact supposed, to assume a Jute origin for the Saxons of the parts in question, the third is, in the mind of the present writer, the most probable.

§ 21. It has already been stated that concerning the Angles and Saxons, no reasonable man will put the question which was put in respect to the Jutes, viz., had they any real place among the Germanic invaders of England? Respecting, however, their relations to each other, and their respective geographical localities whilst occupants of Germany, anterior to {13}their immigration into Britain, there is much that requires investigation. What were the Saxons of Germany—what the Angles?

§ 22. Difficulties respecting the identification of the Saxons.—There are two senses of the word Saxon, one of which causes difficulty by being too limited; the other by being too wide.

a. The limited sense of the word Saxon.—This is what we get from Ptolemy, the first author who names the Saxons, and who gives them a limited locality at the mouth of the Elbe, bounded by the Sigulones, the Sabalingi, the Kobandi, the Chali, the Phundusii, the Harudes, and other tribes of the Cimbric Peninsula, of which the Saxons just occupied the neck, and three small islands opposite—probably Fohr, Sylt, and Nordstand.

Now a sense of the word Saxon thus limited, would restrict the joint conquerors of Britain to the small area comprized between the Elbe and Eyder, of which they do not seem even to have held the whole.

b. The wide sense of the word Saxon.—The reader need scarcely be reminded that the present kingdom of Saxony is as far inland as the northern frontier of Bohemia. Laying this, however, out of the question, as the effect of an extension subsequent to the invasion of Britain, we still find Saxons in ancient Hanover, ancient Oldenburg, ancient Westphalia, and (speaking roughly) over the greater part of the country drained by the Weser, and of the area inclosed by the eastern feeders of the Lower Rhine, the Elbe, and the range of the Hartz.

Now as it is not likely that the limited Saxon area of Ptolemy should have supplied the whole of our Saxon population, so on the other hand, it is certain, that of a considerable portion of the Saxon area in its wider extent tribes other than the Saxons of England, were occupants.

§ 23. Difficulties respecting the word Angle.—The reader is referred to an extract from the Anglo-Saxon Chronicle, in § 16, where it is stated, that "from the Angles' land (which has since always stood waste betwixt the Jutes and the {14}Saxons) came the East-Angles, Middle-Angles, Mercians, and all the Northumbrians."

Thus to bring the great Angle population from an area no larger than the county of Rutland, is an objection—but it is not the chief one.

The chief objection to the Angles of England being derived from the little district of Anglen, in Sleswick, lies in the fact of there being mention of Angli in another part of Germany.

§ 24. This exposition of the elements of uncertainty will be followed by an enumeration of—

1. Those portions of the Germanic populations, which from their geographical position, are the likeliest, à priori, to have helped to people England.

2. Those portions of the Germanic population, which although not supposed to have contributed in any notable degree to the population of Britain, had such continental relations to the Angles and Saxons, as to help in fixing their localities.

These two scenes of facts, give us what may be called our preliminary apparatus criticus.

§ 25. Between the northern limits of the Celtic populations of Gaul and the southern boundary of the Scandinavians of Jutland, we find the area which is most likely to have given origin to the Germans of England. This is best considered under two heads.

b. That of the rivers, i.e., the communications between the ocean and the inland country.

This double division is sufficient, since it is not likely that Britain was peopled by any tribes which were not either maritime, or the occupants of a river.

On the other hand, it is necessary, since although the à priori view is in favour of the coast having supplied the British immigration, the chances of its having proceeded from the interior by the way of the large rivers Rhine, Weser, and Elbe, must also be taken into consideration. {15}

§ 26. The Menapians.—Locality, from the country of the Morini on the French side of the Straits of Dover, to the Scheldt. It is generally considered that these were not Germans but Celts. The fact, however, is by no means ascertained. If Germans, the Menapians were the tribes nearest to Britain. Again, supposing that the present Flemings of Belgium are the oldest inhabitants of the country, their origin is either wholly, or in part, Menapian. Mentioned by Cæsar.

§ 27. The Batavians.—Mentioned by Cæsar; locality, from the Maas to the Zuyder Zee. Conterminous with the Menapians on the south, and with the Frisians on the north. If the present Dutch of Holland be the inhabitants of the country from the time of Cæsar downwards, their origin is Batavian.

§ 28. The Frisians.—First known to the Romans during the campaign of Drusus—"tributum Frisiis transrhenano populo—Drusus jusserat modicum;"^[3] Tacitus, Ann. iv. 72. Extended, according to Ptolemy, as far north as the Ems—τὴν δὲ παρωκεανῖτιν κατέχουσιν ... οἱ Φρίσσιοι, μέχρι τοῦ Αμισίου ποταμοῦ.

Now, as the dialect of the modern province of Friesland differs in many important points from the Dutch of Holland and Flanders; and as there is every reason to believe that the same, or greater difference, existed between the old Frisians and the old Batavians, assuming each to have been the mother-tongues of the present Frisian and Dutch respectively, we may consider that in reaching the parts to the north of the Zuyder-Zee, we have come to a second sub-division of the Germanic dialects; nevertheless, it is not the division to which either the Angles or the Saxons belong, as may be ascertained by the difference of dialect, or rather language.

§ 29. The Chauci.—Connected with the Frisii.—Falling into two divisions—the lesser (?) Chauci, from the Ems to the Weser; the greater (?) Chauci from the Weser to the Elbe—μετὰ δὲ τούτους (the Frisians), {16}Καῦχοι οἱ μικροὶ μέχρι τοῦ Οὐισούργιος ποταμοῦ, εἶτα Καῦχοι οἱ μειζοῦς, μέχρι τοῦ Ἄλβιος ποταμοῦ.

Tacitus describes the Chauci thus:—"Tam immensum terrarum spatium non tenent tantum Chauci, sed et implent; populus inter Germanos nobilissimus."

The Frisians, as has been stated, represent a separate subdivision of the German dialects, as opposed to the ancient Batavian, and the modern Dutch and Flemish. Did the Chauci represent a third, or were they part of the Frisian division?

The latter is the more likely, and that for the following reasons—Vestiges of Frisian dialects are to be found on the Continent, in Oldenburgh, and also in the island of Heligoland.

More important still is the North-Frisian dialect. North of the Elbe, in the Dutchy of Sleswick, and from the Eyder to Tondern, we find a tract of land called, by Saxo Grammaticus, Frisia Minor, and by other writers, Frisia Eydorensis.

Now, as there are no grounds for considering these North Frisians as other than indigenous to the tract in question, we get an additional reason for looking upon the intermediate line of coast as Frisian rather than either Angle or Saxon—or, at least, such parts of it as are not expressly stated to be otherwise.

§ 30. Inference.—As the whole coast south of the Elbe seems to have been occupied by tribes speaking either Frisian or Batavian dialects, and as neither of these sub-divisions represents the language of the Angles and Saxons, the original localities of those invaders must be sought for either north of the Elbe, or inland, along the course of the rivers, i.e.—inland.

§ 31. The Saxons and Nordalbingians.—North of the Elbe, and south of the Eyder (as stated in § 22), we meet the Saxons of Ptolemy; but that in a very circumscribed locality.

In the ninth century, the tribes of these parts are divided into three divisions:—

a. The Holtsati=the people of Holstein. Here holt=wood, whilst sat is the -set in Somer-set and Dor-set. {17}

Besides the names of these three particular divisions the tribes between the Elbe and Eyder were called by the general name of Nordalbingii=i.e. people to the north of the Elbe.

§ 32. The people of Anglen—North of the Nordalbingii; Anglen being the name of a district between the Schlie and Flensburg.

§ 34. The Saxons of Holstein, how large their area?—There is no reason for considering the Nordalbingian Holtsati, Thiedmarsi and Stormarii as other than Saxons; although the fact of the Northfrisians to the north, and of the Frisians of Hanover to the south of them, is a slight complication of the primâ facie view.

Neither is it necessary to identify the two divisions, and to consider the Saxons as Frisians, or the Frisians as Saxons, as is done by some authors.

It is only necessary to perceive the complication which the existence of the Northfrisians introduces, and to recognise the improbability of parts of the present dutchies of Holstein and Sleswick having constituted the whole of the Anglo-Saxon area.

In other words, we have to ascertain in what direction the Germanic population represented by the Saxons at the mouth of the Elbe extended itself—for some further extension there undoubtedly must have been.

§ 35. This brings us to the other series of preliminary facts, viz.: the consideration of the more important tribes of the middle and lower courses of the three great rivers, the Rhine, the Weser, and the Elbe.

§ 36. The Germans of the Middle Rhine.—Of the Germans of the Lower and Middle Rhine, it is only necessary to mention one—

The Franks.—We shall see that, taking the two terms in their widest sense, the Franks and the Saxons were in contact, a fact which makes it necessary to notice at least some portion of the Frank area. {18}

a. Salian Franks.—If the element Sal- represent the -sel, in the name of the Dutch river Y-ssel, the locality of the Salian Franks was Overyssel and Guelderland, whilst their ethnological relations were most probably with the Batavians.

b. Chamavi.—In the Tabula Peutingeriana we find—Chamavi qui Elpranci (leg. et Franci). They were conterminous with the Salii—Ὑπεδεξάμην μὲν μοῖραν τοῦ Σαλίων ἔθνους, Χαμάβους δὲ ἐξήλασα.—Julian, Op. p. 280.—D.N.

The following extract is more important, as it shows that a Roman communication at least took place between the Rhine and Britain: Χαμάβων γὰρ μὴ βουλευομένων, ἀδύνατόν ἐστιν τὴν τῆς Βρεταννίκης νήσου σιτοπομπίαν ἐπὶ τὰ Ῥωμάϊκα φρούρία διαπέμπεσθαι.—Eunap. in Except. leg. ed., Bonn, p. 42.—D.N.

The name Chamavi is still preserved in that of the district of Hameland, near Deventer.—D.N. and G.D.S.

The Bructeri, Sigambri, and Ripuarian Franks bring us to the Franks of the Middle Rhine, a portion of the division which it is not necessary to follow.

§ 37. The Thuringians.—First mentioned in the beginning of the fourth century. Locality, between the Hartz, the Werra a feeder of the Weser, and the Sala a feeder of the Elbe. As early as the sixth century the Thuringians and Saxons are conterminous, and members of the same confederation against the Franks.—D.N.

§ 38. The Catti.—Locality, the valley of the Fulda, forming part of the Upper Weser. Conterminous with the Thuringi (from whom they were separated by the river Werra) on the east, and the Franks on the west. The modern form of the word Catti is Hesse, and the principality of Hesse is their old locality.—G.D.S.

§ 39. Geographical conditions of the Saxon area.—Southern and northern limits.—The Saxons were in league with the Thuringians and Jutes against the Franks.

By the Jutes they were limited on the north, by the Thuringians on the south-east, and by the Franks on the south-west; the middle portion of the southern frontier being formed by the Catti between the Franks and Thuringians. {19}

Western limit.—This is formed by the Batavians and Frisians of the sea-coast, i.e., by the Batavians of Holland, Guelderland, and Overyssel, and, afterwards, by the Frisians of West and East Friesland, and of Oldenburg.

Here, however, the breadth of the non-Saxon area is uncertain. Generally speaking, it is broadest in the southern, and narrowest in the northern portion. The Frisian line is narrower than the Batavian, whilst when we reach the Elbe the Saxons appear on the sea-coast. Perhaps they do so on the Weser as well.

§ 40. Eastern limit.—Preliminary remark.—Before the eastern limit of the Saxons is investigated, it will be well to indicate the extent to which it differs from the southern.

a. The Thuringians, Catti (or Hessians), and Franks, on the southern boundary of the Saxon area were Germans. Hence the line of demarcation between their language was no broad and definite line, like that between the English and the Welsh, but rather one representing a difference of dialect, like that between the Yorkshire and the Lowland Scotch. Hence, too, we ought not only not to be surprised, if we find dialects intermediate to the Frank and Saxon, the Saxon and Thuringian, &c., but we must expect to find them.

b. The same is the case with the Batavian and Frisian frontier.—We really find specimens of language which some writers call Saxon, and others Dutch (Batavian).

The eastern frontier, however, will be like the frontier between England and Wales, where the line of demarcation is broad and definite, where there are no intermediate and transitional dialects, and where the two contiguous languages belong to different philological classes.—The languages to the east of the Saxon area will be allied to the languages of Russia, Poland, and Bohemia; i.e., they will be not Germanic but Slavonic.

Note.—The northern frontier of the Saxon area is intermediate in character to the western and southern on one hand, and to the eastern on the other; the Danish of the Cimbric Peninsula being—though not German—Gothic. {20}

We begin at the northern portion of the Saxon area, i.e., the south-eastern corner of the Cimbric Peninsula, and the parts about the Town of Lubeck; where the Dutchies of Mecklenburg Schwerin and Holstein join. The attention of the reader is particularly directed to the dates.

§ 41. Slavonians of Holstein, Mecklenburg, and Lauenburg.—The Polabi—From po=on, and Labe=the Elbe. Name Slavonic. Germanized by the addition of the termination—ing, and so become Po-lab-ing-i; just as in Kent we find the Kent-ing-s. Conterminous with the Nordalbingian Stormarii, from whom they are divided by the river Bille, a small confluent of the Elbe. Capital Ratzeburg. First mentioned by writers subsequent to the time of Charlemagne.—D.N.

§ 42. The Wagrians.—North of the Polabi, and within the Cimbric Peninsula, divided from the Danes by the Eyder, from the Non-Danish Nordalbingians by the Trave. Capital Oldenburg. The Isle of Femern was Wagrian. Authorities—chiefly writers of and subsequent to the time of Charlemagne. In one of these we learn that the town of Haðum (Sleswick) lies between the Angles, the Saxons, and the Wends.

Now, Wend is the German designation of the Slavonians; so that there must have been Slavonians in the Cimbric Peninsula at least as early as the ninth century.—D.N.

§ 43. Obotriti, written also Obotritæ, Abotriti, Abotridi; Apodritæ, Abatareni, Apdrede, Afdrege, and for the sake of distinction from a people of the same name, Nort-Obtrezi, occupants of the western part of Mecklenburg, and extended as far east as the Warnow, as far south as Schwerin. Called by Adam of Bremen, Reregi. The Obotrites were allies of the Franks against the Saxons, and after the defeat and partial removal of the latter, were transplanted to some of their localities.—"Saxones transtulit" (i.e., Charlemagne), "in Franciam et pagos transalbianos Abodritis dedit."—Eginhart Ann. A.D. 804.—D.N.

§ 44. The Lini—Slavonians on the left bank of the Elbe, and the first met with on that side of the river. Occupants of Danneburg, Luchow and Wustrow, in Luneburg. By the {21}writers subsequent to the time of Charlemagne the Smeldengi (a German designation), and the Bethenici are mentioned along with the Lini (or Linones). Of this Slavonic a Paternoster may be seen in the Mithridates representing the dialect of the neighbourhood in Luchow in A.D. 1691. It is much mixed with the German. About the middle of the last century this (Cis-Albian Slavonic) dialect became extinct.—D.N.

§ 45. The Warnabi or Warnavi.—Locality. Parts about Grabow, Valley of the Elbe. This is the locality of the Varini of Tacitus, the Οὐΐρούνοι of Ptolemy, and the Werini of later writers, a tribe connected with the Angli, and generally considered as Germanic.—D.N.

§ 48. Hevelli.—Locality; the valley of the Hevel. These are the Slavonians of Brandenburg and Mittelmark.—D.N.

§ 49. Slavonians of Altmark.—In Altmark, as in Lunenburg, though on the German side of the Elbe we find the names of the places Slavonic, e.g., Klotze, Wrepke, Solpke, Blatz, Regatz, Colbitz, &c.; so that Altmark, like Lunenburg, was originally a Cis-Albian Slavonic locality.

§ 50. South of the Hevel we meet with the Sorabian, or Sorb Slavonians, the descendants of whom form at the present time part of the population of Lusatia and Silesia. It is not, however, necessary to follow these further, since the German frontier now begins to be Thuringian rather than Saxon.

§ 51. Saxon area.—From the preceding investigations we determine the area occupied by the Saxons of Germany to be nearly as follows:

a.—Ethnologically considered.—Tract bounded on the north by the North Frisian Germans and Jute Danes of Sleswick; on the north and north-east by the Slavonians of the Elbe, sometimes Trans-Albian like the Wagrians and Obotrites; sometimes Cis-Albian, like the Linones and the Slaves of Altmark; on the south by the Thuringians, Catti, and Franks; on the west by the Franks, Batavians, and Frisians.

b. Considered in relation to the ancient population that it {22}comprised.—The country of the Saxons of Ptolemy; the Angli of Tacitus; the Langobardi of Tacitus; the Angrivarii; the Dulgubini; the Ampsivarii (?); the Bructeri Minores (?); the Fosi, and Cherusci; and probably part of the Cauci. Of populations mentioned by the later writers (i.e. of those between the seventh and eleventh centuries), the following belong to this area—the Stormarii, Thietmarsi, Hotsati (=the Nordalbingii, or Nordleudi), the Ostfali, (Osterluidi), Westfali, Angarii, and Eald-Seaxan (Old Saxons).

c. Considered in relation to its modern population.—Here it coincides most closely with the kingdom of Hanover, plus parts of the Dutchies of Holstein and Oldenburg, and parts of Altmark? Brunswick? and Westphalia, and minus the Frisian portion of East Friesland, and the Slavonic part of Luneburg.

d. River system.—By extending the Saxons of Westphalia as far as Cleves (which has been done by competent judges) we carry the western limit to the neighbourhood of the Rhine. This, however, is as far as it can safely be carried. In the respect to the Upper Ems, it was probably Saxon, the lower part being Frisian. The Weser is pre-eminently the river of the Saxons, with the water-system of which their area coincides more closely than with any other physical division. The Elbe was much in the same relation to the Germans and Slavonians, as the Rhine was to the Germans and the Gauls. Roughly speaking, it is the frontier—the Cis-Albian Slaves (the Linones and the Slavonians of Altmark) being quite as numerous as the Trans-Albian Germans, (the people of Stormar, Ditmarsh, and Holstein). The Eyder was perhaps equally Danish, Frisian, and Saxon.

e. Mountains.—The watershed of the Weser on the one side, and of the Ruhr and Lippe on the other, is the chief high land contained within the Saxon area, and is noticed as being the line most likely to form a subdivision of the Saxon population, either in the way of dialect or political relations—in case such a subdivision exists, a point which will be considered in the next chapter.

CHAPTER III.

§ 52. The area occupied by the Saxons of Germany has been investigated; and it now remains to ask, how far the language of the occupants was absolutely identical throughout, or how far it fell into dialects or sub-dialects. In doing this, it may as well be asked, First, what we expect, à priori; Second, what we really find.

§ 53. To the Saxon area in Germany, there are five philological frontiers, the Slavonic, the Frisian, the Batavian, the Frank, and the Thuringian, to which may probably be added the Hessian; in each of which, except the Slavonic, we may expect that the philological phenomenon of intermixture and transition will occur. Thus—

Finally, the Saxon of the centre of the area is expected to be the Saxon of the most typical character.

§ 54. Such is what we expect. How far it was the fact is not known for want of data. What is known, however, is as follows.—There were at least two divisions of the Saxon; (1st) the Saxon of which the extant specimens are of English origin, and (2nd), the Saxon of which the extant specimens are of continental origin. We will call these at present the Saxon of England, and the Saxon of the Continent. {24}

§ 55. Respecting the Saxon of England and the Saxon of the Continent, there is good reason for believing that the first was spoken in the northern, the second in the southern portion of the Saxon area, i.e., the one in Hanover and the other in Westphalia, the probable boundaries between them being the line of highlands between Osnaburg and Paderborn.

§ 56. Respecting the Saxon of England and the Saxon of the Continent, there is good reason for believing that, whilst the former was the mother-tongue of the Angles and the conquerors of England, the latter was that of the Cherusci of Arminius, the conquerors and the annihilators of the legions of Varus.

§ 57. Respecting the Saxon of England and the Saxon of the Continent, it is a fact that whilst we have a full literature in the former, we have but fragmentary specimens of the latter—these being chiefly the following: (1) the Heliand, (2) Hildubrand and Hathubrant, (3) the Carolinian Psalms.

§ 58. The preceding points have been predicated respecting the difference between the two ascertained Saxon dialects, for the sake of preparing the reader for the names by which they are known. Supposing the nomenclature to be based upon any of the preceding facts, we might have the following nomenclature:—

Of these names the last would be the best for strictly scientific purposes, or for the purposes of investigation; since the fact upon which it is based is the most undeniable.

Such is what the nomenclature might be, or, perhaps, ought to be. What it is is another question.

§ 59. The Saxon of England is called Anglo-Saxon; a term against which no exception can be raised.

§ 60. The Saxon of the Continental used to be called Dano-Saxon, and is called Old Saxon.

§ 61. Why called Dano-Saxon.—When the poem called Heliand was first discovered (and that in an English library), the difference in language between it and the common Anglo-Saxon composition was accounted for by the assumption of a Danish intermixture.

§ 62. Why called Old Saxon.—When the Continental origin of the Heliand was recognised, the language was called Old Saxon, because it represented the Saxon of the mother-country, the natives of which were called Old Saxons by the Anglo-Saxons themselves. Still the term is exceptionable; the Saxon of the Heliand is most probably a sister-dialect of the Anglo-Saxon, rather the Anglo-Saxon itself is a continental locality. Exceptionable, however, as it is, it will be employed.

2. Confessionis Formulæ, e Codice Essensi.—Lacomblet, Archiv, für Geschichte des Niederrhins, 1, 4-9.

9. The Carolinian Psalms.—A translation of the Psalms, referred to the time of Charlemagne; sometimes considered to be old Batavian.

10. Heliand, a Gospel Harmony in alliterative metre, and the chief Old Saxon composition extant. {26}

CHAPTER IV.

§ 65. The last chapter has limited the Anglo-Saxon area to the northern part of the Saxon area in general. Further details, however, upon this point, may stand over until the general affinities of the English language have been considered.

§ 66. Over and above those languages of Germany and Holland which were akin to the dialects of the Angles and the Saxons, cognate languages were spoken in Denmark, Sweden, Norway, Iceland, and the Feroe isles, i.e., in Scandinavia.

§ 67. The general collective designation for the Germanic tongues of Germany and Holland, and for the Scandinavian languages of Denmark, Sweden, Norway, Iceland, and the Feroe Isles, is taken from the name of those German tribes who, during the decline of the Roman Empire, were best known to the Romans as the Goths; the term Gothic for the Scandinavian and Germanic languages, collectively, being both current and convenient.

§ 68. Of this great stock of languages the Scandinavian is one branch; the Germanic, called also Teutonic, another.

§ 69. The Scandinavian branch of the Gothic stock comprehends, 1. The dialects of Scandinavia Proper, i.e., of Norway and Sweden; 2. of the Danish isles and Jutland; 3. of Iceland; 4. of the Feroe Isles. On the side of Lapland the languages of this branch come in contact with the Laplandic and Finlandic; whilst in Sleswick they are bounded by the Low German. {29}

§ 71. It is in the Mœso-Gothic that the most ancient specimen of any Gothic tongue has been preserved. It is also the Mœso-Gothic that was spoken by the conquerors of ancient Rome; by the subjects of Hermanic, Alaric, Theodoric, Genseric (?), Euric, Athanaric, and Totila.

This history of this language, and the meaning of the term by which it is designated, is best explained by the following passages:—

a. A.D. 482. "Trocondo et Severino consulibus—Theodoricus cognomento Valamer utramque Macedoniam, Thessaliamque depopulatus est, Larissam quoque metropolim depredatus, Fausto solo consule (A.D. 485)—Idem Theodoricus rex Gothorum Zenonis Augusti munificentia pene pacatus, magisterque præsentis militiæ factus, consul quoque designatus, creditam sibi Ripensis Daciæ partem Mœsiæque inferioris, cum suis satellitibus pro tempore tenuit."—Marcellini Comitis Chronicon, D.N.

b. "Frederichus ad Theodoricum regem, qui tunc apud Novam Civitatem provinciæ Mœsiæ morabatur, profectus est."—Vita S. Severini, D.N.

c. "Zeno misit ad Civitatem Novam, in quâ erat Theodoricus dux Gothorum, filius Valameris, et eum invitavit in solatium sibi adversus Basiliscum."—Anon. Valesii, p. 663, D.N.

d. Civitas Nova is Nicopolis on the Danube; and the nation thus spoken of is the Gothic nation in the time of Zeno. At this time they are settled in the Lower Mœsia, or Bulgaria.

How they got here from the northern side of the Danube we find in the history of the reign of Valens. When pressed by intestine wars, and by the movements of the Huns, they were assisted by that emperor, and settled in the parts in question. {32}

Furthermore, they were converted to Christianity; and the Bible was translated into their language by their Bishop Ulphilas.

Fragments of this translation, chiefly from the Gospels, have come down to the present time; and the Bible translation of the Arian Bishop Ulphilas, in the language of the Goths of Mœsia, during the reign of Valens, exhibits the earliest sample of any Gothic tongue.

§ 72. How Gothic tribes reached the Lower Danube is a point upon which there is a variety of opinion. The following facts, however, may serve as the basis of our reasoning.

A.D. 249-251—The Goths are found about equidistant from the Euxine Sea, and the eastern portion of the range of Mount Hæmus, in the Lower Mœsia, and at Marcianopolis. Here they gain a great battle against the Romans, in which the Emperor Decius is killed.

During the reign of Gallienus they appear as maritime warriors, and ravage Asia Minor, Greece, and Illyria.

A.D. 269—Are conquered at Naissus, on the western boundary of Mœsia Superior by Claudius.

A.D. 373—In the reign of Valens (as already stated), they were admitted to settle within the limits of the empire.

§ 73. Now, although all this explains, how a Gothic language was spoken in Bulgaria, and how remnants of it have been preserved until the nineteenth century, the manner in which the tribe who spoke it reached Marcianopolis, so as to conquer the Emperor Decius, in A.D. 249, is unexplained.

A. The Baltic doctrine. According to this the Goths migrated from the Baltic to the Mæotis, from the Mæotis to the Euxine, and from the Euxine to the Danube, along which river they moved from east to west. {33}

B. The Getic doctrine.—Here the Goths are made out to be the aborigines of the Lower Danube, of Dacia, Mœsia, and even Thrace; in which case their movement was, also, from east to west.

C. The German doctrine.—Here the migration is from west to east, along the course of the Danube, from some part of south-eastern Germany, as its starting-point, to Asia Minor as its extreme point, and to Bulgaria (Mœsia Inferior) as its point of settlement.

§ 74. Respecting the first of these views the most that can be said in its favour is, that it is laid down by Jornandes, who wrote in the fifth century, and founded his history upon the earlier writings of Ablavius and Dexippus, Gothic historians, who, in their turn took their account from the old legends of the Goths themselves—in priscis eorum carminibus, pæne historico ritu. On the other hand, the evidence is, at best, traditional, the fact improbable, and the likelihood of some such genealogy being concocted after the relationship between the Goths of the Euxine, and Germans of the Baltic had been ascertained exceedingly great.

§ 75. The second is supported by no less an authority than Grimm, in his latest work, the History of the German Language;—and the fact of so learned and comprehensive an investigator having admitted it, is, in the mind of the present writer, the only circumstance in its favour. Over and above the arguments that may be founded on a fact which will soon be noticed, the chief reasons are deduced from a list of Dacian or Getic plants in Dioscorides, which are considered to bear names significant in the German. Whether or not, the details of this line of criticism will satisfy the reader who refers to them, it is certain that they are not likely to take a more cogent form than they take in the hands of the Deutsche Grammatik.

§ 76. The third opinion is the likeliest; and if it were not for a single difficulty would, probably, never have been demurred to. The fact in question is the similarity between the words Getæ and Gothi.

The fact that a tribe called G-O-T-H-I should, when they first peopled the Mœsogothic country, have hit upon the {34}country of a people with a name so like their own as G-E-T-Æ, by mere accident, is strange. English or American colonies might be sent to some thousand places before one would be found with a name so like that of the mother-country as Get is to Got. The chances, therefore, are that the similarity of name is not accidental, but that there is some historical, ethnological, or geographical grounds to account for it. Grimm's view has been noticed. He recognises the difficulty, and accounts for it by making the Goths indigenous to the land of Getæ.

To a writer who (at one and the same time) finds difficulty in believing that this similarity is accidental and is dissatisfied with Grimm's reasoning, there seems to be no other alternative but to consider that the Goths of the Lower Danube had no existence at all in Germany under that name, that they left their country under a different^[5] one, and that they took the one by which they were known to the Romans (and through them to us), on reaching the land of the Getæ—as, in England, the Saxons of Essex and Wessex did not (since they brought their name with them), but as the East and West Kent-ings^[6] did.

This doctrine, of course, falls to the ground directly it can be shown that the Goths of Mœsia were either called Goths in Germany, or any where else, anterior to their settlement in the Geta-land.

Be this, however, as it may, the first division of the Teutonic branch of languages is the Mœso-Gothic of the Goths of the Lower Danube, in the fourth century, as preserved in the translation of Ulphilas, and in other less important fragments.

§ 77. The Old High German, called also Francic and Alemannic, was spoken in the ninth, tenth, and eleventh centuries, in Suabia, Bavaria, and Franconia. It is in the Old High German that the Krist of Otfrid, the Psalms of Notker, the Canticle of Willeram, the Glosses of Kero, the Vita Annonis, &c., are composed.

§ 78. The Low Germanic Division, to which the Anglo-Saxon belongs, is currently said to comprise six languages, or rather four languages in different stages.

§ 79. The Frisian and Dutch.—It is a current statement that the Old Frisian bears the same relation to the Modern Dutch of Holland that the Anglo-Saxon does to the English.

1. That a single language, spoken in two dialects, was originally common to both Holland and Friesland.

2. That from the northern of these dialects we have the Modern Frisian of Friesland.

The Modern Dutch has certain grammatical forms older than those of the Old Frisian; e.g., the Dutch infinitives and the Dutch weak substantives, in their oblique cases, end in -en; those of the Old Frisian in -a: the form in -en being the older.

4. The North Frisian, spoken in a few villages of Sleswick. One of the characters of the North Frisian is the possession of a Dual Number.

§ 81. In respect to its stages, we have the Old Frisian of the Asega-bog, the Middle Frisian of Gysbert Japicx, and the Modern Frisian of the present Frieslanders, Westphalians, and Heligolanders.

§ 82. The Low German and Platt-Deutsch.—The words Low German are not only lax in their application, but they are equivocal; since the term has two meanings, a general meaning when it signifies a division of the Germanic languages, comprising English, Dutch, Anglo-Saxon, Old Saxon, and Frisian, and a limited one when it means the particular dialects of the Ems, the Weser, and the Elbe. To avoid this the dialects in question will be henceforth called by their continental name of Platt-Deutsch; which although foreign, is convenient.

§ 83. The points of likeness and difference between two languages belonging to different branches of the same Gothic stock may be partially collected from the following comparison between certain Icelandic, Norse or Scandinavian, and certain Anglo-Saxon or Germanic inflections.

§ 84. The most characteristic difference between the Saxon and Icelandic lies in the peculiar position of the definite article in the latter language. In Saxon, the article corresponding with the modern word the, is þæt, se, seó, for the neuter, masculine, and feminine genders respectively; and these words, regularly declined, are prefixed to the words with which they agree, just as is the case with the English and with the majority of languages. In Icelandic, however, the article, instead of preceding, follows its noun, with which it coalesces, having previously suffered a change in form. The Icelandic article corresponding to þæt, se, seó, is hitt (N.), hinn (M.), hin (F.): from this the h is ejected, so that, instead of the regular inflection (a), we have the forms (b).

§ 85. In the Swedish, Norwegian, and Danish this peculiarity in the position of the definite article is preserved. Its origin, however, is concealed; and an accidental identity with the indefinite article has led to false notions respecting its nature. In the languages in point the i is changed into e, so that what in Icelandic is it and in, is in Danish et and en. En, however, as a separate word, is the numeral one, and also the indefinite article a; whilst in the neuter gender it is et—en Sol, a sun; et Bord, a table: Solen, the sun; Bordet, the table. From modern forms like those just quoted, it has been imagined that the definite is merely the indefinite article transposed. This it is not.

Reference will be made to this passage on more occasions than one, to show how words originally distinct may, in the process of time, take the appearance of being identical. To apply an expression of Mr. Cobbett's, en=a, and -en=the, are the same combination of letters, but not the same word. {42}

§ 86. Observe in the Icelandic forms the absence of the termination -an. Observe also the neuter termination -t, as hagr, hagt. Throughout the modern forms of the Icelandic (viz. the Swedish, Danish, and Norwegian languages) this termination is still preserved: e.g., en god Hest, a good horse; et godt Hjært, a good heart; en skön Pige, a beautiful damsel; et skarpt Svœrd, a sharp sword.

§ 87. Amongst the pronouns the following differences present themselves. The Saxon forms are, for the pronoun of the second person, þu (thou), git (ye two), ge (ye); whilst in Icelandic they are þu, þið, per, respectively. Again, in Saxon there is no reflective pronoun corresponding with the Latin se. In Icelandic we have sik, sér, sin, corresponding to the Latin se, sibi, suus. Besides this, the word sin is declined, so that like the Latin suus it becomes adjectival.

In Saxon there is of course no such an adjectival form. There the Possessives of the Third Person correspond not with the Latin suus, sua, suum; but with the Latin ejus and eorum. The English words his and her are genitive cases, not adjectives.

Further remarks upon the presence of the Reflective Pronoun sik in Icelandic, and its absence in Saxon, will appear in the sequel.

Of the Icelandic verbs the infinitives end in -a; as kalla, to call; elska, to love; whereas the Saxon termination is -an; as lufian, to love; wyrcan, to work. {44}

§ 89. The characteristic, however, of the Icelandic (indeed, of all the Scandinavian languages) is the possession of a passive form, or a passive voice, ending in -st:—Ek, þu, hann brennist=I, thou, he is burnt; Ver brennumst=We are burnt; þér brennizt=ye are burnt; þeir brennast=they are burnt. Past tense, Ek, þu, hann brendist; ver brendumst, þér brenduzt, þeir brendust. Imperat.: brenstu=be thou burnt. Infinit.: brennast=to be burnt.

In the modern Danish and Swedish, the passive is still preserved, but without the final t. In the older stages of Icelandic, on the other hand, the termination was not -st but -sc; which -sc grew out of the reflective pronoun sik. With these phenomena the Scandinavian languages give us the evolution and development of a passive voice; wherein we have the following series of changes:—1st. the reflective pronoun coalesces with the verb, whilst the sense changes from that of a reflective to that of a middle verb; 2nd. the c changes to t, whilst the middle sense passes into a passive one; 3rd. t is dropped from the end of the word, and the expression that was once reflective then becomes strictly passive.

Now the Saxons have no passive voice at all. That they should have one originating like that of the Scandinavians was impossible. Having no reflective pronoun, they had nothing to evolve it from.

§ 90. Recapitulating, we find that the characteristic differences of the greatest importance between the Icelandic and Saxon are three in number:—

§ 91. In the previous comparison the substantives were divided as follows:—1st. into those ending with a vowel; 2ndly, into those ending with a consonant. In respect to the substantives ending with a vowel (eáge, nama, tunge), it may have been observed that their cases were in A. S. almost {46}exclusively formed in -n, as eágan, tungan, &c.; whilst words like skip and smið had, throughout their whole declension, no case formed in -n; no case indeed wherein the sound of -n entered. This enables us (at least with the A. S.) to make a general assertion concerning the substantives ending in a vowel in contrast to those ending in a consonant, viz. that they take an inflection in -n.

In Icelandic this inflection in -n is concealed by the fact of -an having been changed into -a. However, as this -a represents -an, and as fragments or rudiments of -n are found in the genitive plurals of the neuter and feminine genders (augna, tungna), we may make the same general assertion in Icelandic that we make in A. S., viz. that substantives ending in a vowel take an inflection in -n.

§ 92. The points of likeness and difference between two languages, belonging to different divisions of the same Germanic branch, may be partially collected from the following comparison between certain Mœso-Gothic and certain Anglo-Saxon inflections.

§ 93. It must, however, be premised, that, although the distinction between nouns taking an inflection in -n, and nouns not so inflected, exists equally in the Mœso-Gothic and the Icelandic, the form in which the difference shows itself is different; and along with the indication of this difference may be introduced the important terms weak and strong, as applied to the declension of nouns.

Weak nouns end in a vowel; or, if in a consonant, in a consonant that has become final from the loss of the vowel that originally followed it. They also form a certain proportion of their oblique cases in -n, or an equivalent to -n—Nom. augô, gen. aug-in-s.

Strong nouns end in a consonant; or, if in a vowel, in one of the vowels allied to the semivowels y or w, and through them to the consonants. They also form their oblique cases by the addition of a simple inflection, without the insertion of n.

Furthermore, be it observed that nouns in general are weak and strong, in other words, that adjectives are weak or {47}strong, as well as substantives. Between substantives and adjectives, however, there is this difference:—

1. A substantive is either weak or strong, i.e., it has one of the two inflections, but not both. Augô=an eye, is weak under all circumstances; waurd=a word, is strong under all circumstances.

2. An adjective is both weak and strong. The Anglo-Saxon for good is sometimes god (strong), sometimes gode (weak). Which of the two forms is used depends not on the word itself, but on the state of its construction.

1. The definite sense is generally expressed by the weak form, as se blinde man=the blind man.

2. The indefinite sense is generally expressed by the strong form, as sum blind man=a blind man.

Hence, as far as adjectives are concerned, the words definite and indefinite coincide with the words weak and strong respectively, except that the former are terms based on the syntax, the latter terms based on the etymology of the word to which they apply.

These may be compared with the Saxon declensions; viz. aúgô with eáge, manna with nama, tuggô with tunge, vaúrd with leáf, fisks with smið, and brûþs with spræc.

Observe—In the neuter form blindata M. G. we have the sound of t, as in Icelandic. This becomes z (ts) in Old High German, and s in modern German.

The conjugation of the M. G. is as follows. From the Anglo-Saxon it differs most in its plural persons.

The conjugation of the auxiliary verb in Mœso-Gothic is as follows. It may be compared with the A. S. § 89.

§ 94. The points of likeness or difference between two languages, each of the Low Germanic division, may be partially collected from the following comparison between certain Old Frisian and certain Anglo-Saxon inflections.

As a general statement we may say, that in the transition letters the Frisian corresponds with the A. S. more closely than it does with any other language. It must, moreover, be remarked, that, in such pairs of words as frjund and freond, the difference (as far at least as the e and j are concerned) is a mere difference of orthography. Such also is probably the case with the words déd and dæd, and many others.

The Anglo-Saxon inflection of ^a) Substantives ending in a vowel, ^b) Substantives ending in a consonant, ^c) Adjectives with an indefinite ^d) Adjectives with a definite sense, ^e) Verbs Active ^f) and verbs auxiliar, may be seen in the comparison between the A. S. and the Icelandic. The corresponding inflections in Frisian are as follows:—

With respect to the masculine substantives terminating in a consonant, it must be observed that in A. S. there are two modes of declension; in one, the plural ends in -s; in the other, in -a. The specimen in § 83 represents the first of these modes only. From this the Frisian is essentially different. With the second it has a close alliance; e.g.:—

In the inflection of the verbs there is between the Frisian and A. S. this important difference. In A. S. the infinite ends in -an macian, to make, læran, to learn, bærnan, to burn; whilst in Frisian it ends in -a, as maka, léra, berna.

The Frisian numerals (to be compared with those of the Anglo-Saxons, p. 43), are as follows:—Én, twá, thrjú, {54}fjúwer, fíf, sex, sjúgun, achta, njugun, tian, &c. Of these the first three take an inflection, e.g., En, like Gode and the adjectives, has both a definite and an indefinite form, en, and thet ene; whilst twa and thrjú run as follows:—Nom. and Acc. Neut. twa; Masc. twene; Fem. twa; Dat. twam; Gen. twira.—Nom. and Acc. Neut. thrju; Masc. thre; Fem. thrja; Dat. thrim; Gen. thrira.

In respect to the Pronouns, there is in the Old Frisian of Friesland no dual number, as there is in Anglo-Saxon. On the other hand, however, the Frisians (whilst they have no such form as his) possess, like the Icelandic, the inflected adjectival pronoun sin, corresponding to the Latin suus: whilst, like the Anglo-Saxons, and unlike the Icelanders, they have nothing to correspond with the Latin se.

§ 95. In Frisian there is between the demonstrative pronoun used as an article, and the same word used as a demonstrative in the limited sense of the term, the following difference of declension:—

The Saxons draw no such a distinction. With them the article and demonstrative is declined as follows:—

§ 96. Specimen of Glossarial affinity.—Taken from Rask's Preface to his Frisian Grammar:—

§ 97. In this Chapter there has been, thus far, an attempt to do two things at once. Firstly, to exhibit the general likeness between stocks, branches, &c.; and secondly, to show the special affinities between certain languages allied to our {56}own, and of the Gothic Stock. What follows, consists of certain observations upon two or three points of nomenclature.

1. That it is no national name, but a name given by the Latins to the natives of the country called Germania. The word German is foreign to all the Gothic languages.

2. That it was first applied to proper Germanic tribes in the time of Julius Cæsar, and that it served to distinguish the Gothic Germans from the Celtic Gauls.

3. That, anterior to the time of Cæsar, there is no proof of it being applied as a distinctive designation to any of the tribes to whom it was afterwards limited. The first tribe to whom it was applied, was (in the opinion of the present writer) a Gallic tribe.

4. That since the time of Julius Cæsar, its application has been constant, i.e., it has always meant Gothic tribes, or Gothic languages.

5. That sometimes it has been general to the whole nation—Unde fit ut tantæ populorum multitudines arctoo sub axe oriantur, ut non immerito universa illa regio Tanai tenus usque ad occiduum, licet et propriis loca ea singula nuncupentur nominibus, generali tamen vocabulo Germania vocitetur ... Gothi, siquidem, Vandalique, Rugi, Heruli, atque Turcilingi, necnon etiam aliæ feroces ac barbaræ nationes e Germania prodierunt.—Paulus Diaconus.

6. That sometimes it has been peculiar and distinctive to certain prominent portions of the nation—equi frænis Germanicis, sellis Saxonicis falerati.

7. That the general power of the word has been, with few exceptions, limited to the Germans of Germany. We do not find either English or Scandinavian writers calling their countrymen Germani.

8. That the two German tribes most generally meant, when the word German is used in a limited sense, are the Franks and the Alemanni.

9. That by a similar latitude the words Francic and {57}Alemannic have been occasionally used as synonymous with Germanic.

10. That the origin of the word Germani, in the Latin language, is a point upon which there are two hypotheses.

a. That it is connected with the Latin word Germani=brothers, meaning either tribes akin to one another, or tribes in a degree of brotherly alliance with Rome.

b. That it grew out of some such German word as Herman, Irmin, Wehrmann, or the Herm- in Hermunduri, Hermiones, &c.

For all the facts concerning the word Germani, see the Introduction to the third edition of the Deutsche Grammar.

§ 99. Dutch.—For the purposes of Philology the meaning given to this word is inconvenient. In England, it means the language of the people of Holland.

In Germany, Holland, and Scandinavia, it means the language of the people of Germany in general; and this general power of the word is retained even with us in the expression High-Dutch, and Low-Dutch. In the present work the term is avoided as much as possible. Nevertheless, wherever it occurs it means the Dutch of Holland.

The origin of the word has been a subject of much investigation; the question, however, may be considered to be settled by the remarks of Grimm, D. G.—Introduction to the third edition.

2. In the earliest passage where it occurs, the derivative form þiudiskô corresponds with the Greek word ἐθνικῶς—The Mœso-Gothic Translation of the New Testament—Galatians, ii. 14.

3. The derivation of the word from the substantive þiudu=a people, a nation, is undoubted.

4. So also is the derivation of the modern word Dutch, in all its varied forms:—Old High-German, Diutisc; Anglo-Saxon, Þeódisc; Latin, Theodisca, Theudisca, Teutisca; Italian, Tedesco; Danish, Tyske; English, Dutch; the latter part of the word being the adjectival ending -isc=ish. {58}

5. The original meaning being of, or belonging to, the people, or of, or belonging to, the nation, secondary meanings grew out of it.

6. Of these the most remarkable are a) the power given to the word in Ulphilas (heathen), illustrated by the similarly secondary power of the Greek ἔθνικος; b) the meaning vernacular, provincial or vulgar given to it as applied to language.

8. That it was not given much before, is inferred from negative evidence. The word theotisca is not found in the Latin writers of the sixth, seventh, and eighth centuries, although there are plenty of passages where it might well have been used had it existed. The terms really used are either patrius sermo, sermo barbaricus, sermo vulgaricus, lingua rustica; or else the names of particular tribes, as lingua Anglorum, Alamannorum.

9. That it was current in the ninth century is evident from a variety of quotations:—Ut quilibet episcopus homilias aperte transferre studeat in rusticam Romanam linguam, aut þeotiscam, quo tandem cuncti possint intelligere quæ dicantur.—Synodus Turonensis. Quod in lingua Thiudisca scaftlegi, id est armorum depositio, vocatur.—Capit. Wormatiense. De collectis quas Theudisca lingua heriszuph appellat.—Conventus Silvacensis. Si barbara, quam Teutiscam dicunt, lingua loqueretur.—Vita Adalhardi, &c.—D.G., i. p. 14, Introduction.

10. That its present national sense is wholly secondary and derivative, and that originally it was no more the name of a people or a language than the word vulgate in the expression the vulgate translation of the Scriptures is the name of a people or a language.

§ 100. Teutonic.—About the tenth century the Latin writers upon German affairs began to use not only the words Theotiscus and Theotiscé, but also the words Teutonicus and Teutonicé. Upon this, Grimm remarks that the latter term sounded more learned; since Teutonicus was a classical word, an adjective derived from the Gentile name of the Teutones conquered by Manus. Be it so. It then follows that the connexion between Teutonicus and Theotiscus is a mere accident, the origin {59}of the two words being different. The worthlessness of all evidence concerning the Germanic origin of the Teutonic tribes conquered by Marius, based upon the connexion between the word Teuton and Dutch, has been pointed out by the present writer in the 17th number of the Philological Transactions.^[10] All that is proved is this, viz., that out of the confusion between the two words arose a confusion between the two nations. These last may or may not have been of the same race.

§ 101. Anglo-Saxon—In the ninth century the language of England was Angle, or English. The lingua Anglorum of Bede is translated by Alfred on englisce. The term Saxon was in use also at an early (perhaps an equally early) date—fures quos Saxonice dicimus vergeld þeóvas. The compound term Anglo-Saxon is later.—Grimm, Introduction to the third edition of D.G., p. 2.

§ 102. Icelandic, Old Norse.—Although Icelandic is the usual name for the mother-tongue of the Danish, Swedish, and Norwegian, the Norwegian philologists generally prefer the term Old Norse.

In favour of this view is the fact that Norway was the mother-country, Iceland the colony, and that much of what is called Old Icelandic was composed in Norway.

Still the reason is insufficient; since the present term Icelandic is given to the language not because Iceland was the country that produced, but because it is the country that has preserved it.

This leads to the argument in its most general form—should a language be named from the colony, or from the mother-country? The Norwegians say from the mother-country. Let us consider this.

Suppose that whilst the Latin of Virgil and Cicero in Italy had been changing into the modern Italian, in some old Roman colony (say Sardinia) it had remained either wholly {60}unaltered, or else, altered so little as for the modern Sardinian—provided he could read at all—to be able to read the authors of the Augustan age, just like those of the era of Charles Albert; no other portion of the old Roman territory—not even Rome itself—having any tongue more like to that of the Classical writers, than the most antiquated dialect of the present Italian. Suppose, too, that the term Latin had become obsolete, would it be imperative upon us to call the language of the Classics Old Italian, Old Roman, or at least Old Latin, when no modern native of Rome, Latium, or Italy could read them? Would it be wrong to call it Sardinian when every Sarde could read them? I think not. Mutatis mutandis, this is the case with Iceland and Norway.

CHAPTER V.

§ 103. The population and, to a certain extent, the language of England, have been formed of three elements, which in the most general way may be expressed as follows:—

a. Elements referable to the original British population, and derived from times anterior to the Anglo-Saxon invasion.

§ 104. Each of these requires a special analysis, but that of the second will be taken first, and will form the contents of the present chapter.

All that we have at present learned concerning the Germanic invaders of England, is the geographical area which they wholly or partially occupied, and the tribes and nations with which they were conterminous whilst in Germany. How far, however, it was simple Saxons who conquered England single-handed, or how far the particular Saxon Germans were portions of a complex population, requires further investigation. Were the Saxons one division of the German population, whilst the Angles were another? or were the Angles a section of the Saxons, so that the latter was a generic term, including the former? Again, although the Saxon invasion may be the one which has had the greatest influence, and drawn the most attention, why may there not have been separate and independent migrations, the effects and record of which, have in the lapse of time, become fused with those of the more important divisions?

Questions like these require notice, and in a more advanced state of what may be called minute ethnographical {62}philology will obtain more of it than has hitherto been their share. At present our facts are few, and our methods of investigation imperfect.

§ 105. In respect to this last, it is necessary to distinguish between the opinions based on external, and the opinions based on internal evidence. To the former class belong the testimonies of cotemporary records, or (wanting these) of records based upon transmitted, but cotemporary, evidence. To the latter belong the inferences drawn from similarity of language, name, and other ethnological data. Of such, a portion only will be considered in the present chapter; not that they have no proper place in it, but because the minuter investigation of an important section of these (i.e., the subject of the English dialects) will be treated as a separate subject elsewhere.

§ 106. The Angles; who were they, and what was their relation to the Saxons?—The first answer to this question embodies a great fact in the way of internal evidence, viz., that they were the people from whom England derives the name it bears=the Angle-land, i.e., land of the Angles. Our language too is English, i.e., Angle. Whatever, then, they may have been on the Continent, they were a leading section of the invaders here. Why then has their position in our inquiries been hitherto so subordinate to that of the Saxons? It is because their definitude and preponderance are not so manifest in Germany as we infer (from the terms England and English) it to have been in Britain. Nay more, their historical place amongst the nations of Germany, and within the German area, is both insignificant and doubtful; indeed, it will be seen from the sequel, that in and of themselves we know next to nothing about them, knowing them only in their relations, i.e., to ourselves and to the Saxons. The following, however, are the chief facts that form the foundation for our inferences.

§ 107. Although they are the section of the immigration which gave the name to England, and as such, the preponderating element in the eyes of the present English, they were not so in the eyes of the original British; who neither knew at the time of the Conquest, nor know now, of any other name for their German enemies but Saxon. And Saxon is the {63}name by which the present English are known to the Welsh, Armorican, and Gaelic Celts.

§ 108. Although they are the section of the immigration which gave the name to England, &c., they were quite as little Angles as Saxons, in the eyes of foreign cotemporary writers; since the expression Saxoniæ trans-marinæ, occurs as applied to England.

§ 109. Although they are the section of the immigration which gave the name to England, &c., the material notice of them as Germans of Germany, are limited to the following facts.

Extract from Tacitus.—This merely connects them with certain other tribes, and affirms the existence of certain religious ordinances common to them—

"Contra Langobardos paucitas nobilitat: plurimis ac valentissimis nationibus cincti, non per obsequium, sed prœliis et periclitando tuti sunt. Reudigni deinde, et Aviones, et Angli, et Varini, et Eudoses, et Suardones, et Nuithones, fluminibus aut silvis muniuntur: nec quidquam notabile in singulis, nisi quod in commune Herthum, id est, Terram matrem colunt, eamque intervenire rebus hominum, invehi populis, arbitrantur. Est in insula Oceani castum nemus, dicatumque in eo vehiculum, veste contectum, attingere uni sacerdoti concessum. Is adesse penetrali deam intelligit, vectamque bobus feminis multâ cum veneratione prosequitur. Læti tunc dies, festa loca, quæcumque adventu hospitioque dignatur. Non bella ineunt, non arma sumunt, clausum omne ferrum; pax et quies tunc tantùm nota, tunc tantùm amata, donec idem sacerdos satiatam conversatione mortalium deam templo reddat: mox vehiculum et vestes, et, si credere velis, numen ipsum secreto lacu abluitur. Servi ministrant, quos statim idem lacus haurit. Arcanus hinc terror, sanctaque ignorantia, quid sit id, quod tantùm perituri vident."^[11]

Extract from Ptolemy.—This connects the Angles with {64}the Suevi, and Langobardi, and places them on the Middle Elbe.

Ἐντὸς καὶ μεσογείων ἐθνῶν μέγιστα μέν ἐστι τό, τε τῶν Σουήβων τῶν Ἀγγειλῶν, οἵ εἰσιν ἀνατολικώτεροι τῶν Λαγγοβάρδων, ἀνατείνοντες πρὸς τὰς ἄρκτους μέχρι τῶν μέσων τοῦ Ἄλβιος ποταμοῦ.

Heading of a law referred to the age of Charlemagne.—This connects them with the Werini (Varni), and the Thuringians—"Incipit lex Angliorum et Verinorum (Varni); hoc est Thuringorum."—Zeuss, 495, and Grimm. G.D.S.

§ 110. These notices agree in giving the Angles a German locality, and in connecting them ethnologically, and philologically with the Germans of Germany. The notices that follow, traverse this view of the question, by indicating a slightly different area, and Danish rather than German affinities.

Extracts connecting them with the inhabitants of the Cimbric Peninsula.—a. The quotation from the Anglo-Saxon Chronicle of § 16.

b. From Bede; "Porro de Anglis, hoc est illa patria, quæ Angulus dicitur, et ab eo tempore usque hodie, manere desertus inter provincias Jutarum et Saxonum perhibetur."—Angl. i. 15.

c. From Alfred, "And be wæstan eald Seaxum is Albe muða þære ea and Frisland. And þanon west norð is þæt land, the man Angle, hæt and Sillende, and summe dæl Dena."^[12]—Oros. p. 20.

Also, speaking of Other's voyage,^[13] "He seglode to þæm porte þe man hæt Hæþum; se stent betwuhs Winedum and Seaxum, and Angle, and hyrð in on Dene ... and þa {65}twegen dagas ær he to Hædhum come, him wæs on þæt steorbord Gothland and Sillende and iglanda fela. On þæm landum eardodon Engle, ær hi hiðer on land comon."^[14]—Oros. p. 23.

d. From Etherwerd, writing in the eleventh century—"Anglia vetus sita est inter Saxones et Giotos, habens oppidum capitale, quod sermone Saxonico Sleswic nuncupatur, secundum vero Danos Hathaby."^[14]

§ 111. The district called Angle.—The district of Anglen, so called (where it is mentioned at all) at the present moment, is a part of the Dutchy of Sleswick, which is literally an Angle; i.e., a triangle of irregular shape, formed by the Schlie, the Flensborger Fiord, and a line drawn from Flensborg to Sleswick; every geographical name in it being, at present, Danish, whatever it may have been previously. Thus some villages end in bye (Danish=town) as Hus-bye, Herreds-bye, Ulse-bye, &c.; some in gaard (=house), as Oegaard; whilst the other Danish forms are skov=wood (shaw), hofved=head, lund=grove, &c. In short it has nothing to distinguish it from the other parts of the peninsula.

§ 112. Add to these the Danish expression, that Dan and Angul were brothers, as the exponent of a recognised relationship between the two populations, and we have a view of the evidence in favour of the Danish affinity.

§ 113. Inferences and remarks.—a. That whilst the root Angl- in Tacitus, Ptolemy, Procopius, and the Leges Anglorum, &c., is the name of a people, the root Angl- in the Anglen of Sleswick, is the name of a district; a fact which is further confirmed by the circumstance of there being in at least one other part of Scandinavia, a district with a similar name—"Hann átti bu a Halogolandi i Aungli."^[14]—Heimskringla, iii. 454.

b. That the derivation of the Angles of England from the Anglen of Sleswick is an inference of the same kind with the one respecting the Jutes (see § 20), made by the same writers, probably on the same principle, and most likely incorrectly.

c. That the Angles of England were the Angli of Tacitus, {66}Ptolemy, Procopius, and the Leges Anglorum et Werinorum, whatever these were.

§ 114. What were the Langobardi, with whom the Angles were connected by Tacitus? The most important facts to be known concerning them are, (1) that the general opinion is in favour of their having belonged to the High-German, or Mœso-Gothic division, rather than to the Low; (2) that their original locality either reached or lay beyond the Elbe; a locality, which, in the tenth century, was Slavonic, and which, in the opinion of the present writer, we have no reason to consider to have been other than Slavonic during the nine preceding ones.—That they were partially, at least, on this side of the Elbe, we learn from the following:—"Receptæ Cauchorum nationes, fracti Langobardi, gens etiam Germanis feritate ferocior; denique usque ad flumen Albim ... Romanus cum signis perductus exercitus."^[15]—Velleius Paterc. ii. 106.

§ 115. What were the Suevi, with whom the Angles were connected by Tacitus? The most important facts to be known concerning them are, (1) that the general opinion is in favour of their having belonged to the High-German or Mœso-Gothic, division, rather than to the Low; (2) that their original locality either reached or lay beyond the Elbe; a locality, which, in the tenth century, was Slavonic, and which, in the opinion of the present writer, we have no reason to consider to have been other than Slavonic during the nine preceding ones. In other words, what applies to the Langobardi applies to the Suevi also.

What the Suevi were, the Semnones were also, "Vetustissimos se nobilissimosque Suevorum Semnones memorant." Tac. Germ., 39. Speaking, too, of their great extension, he says, centum pagi ab iis habitantur.^[15]

Velleius states that there were Suevi on the west of the Middle Elbe, Ptolemy, that there were Suevi to the east of it, i.e., as far as the River Suebus (Oder?).—Καὶ τὸ τῶν Σουήβων τῶν Σεμνόνων, οἵτινες διήκουσι μετὰ τὸν Ἄλβιν ἀπὸ τοῦ εἰρημένου μέρους {67}(the middle Elbe) πρὸς ἀνατολὰς μέχρι τοῦ Σουήβου ποταμοῦ.^[16]

In the letter of Theodeberht to the Emperor Justinian, we find the North-Suevians mentioned along with the Thuringians, as having been conquered by the Franks; "Subactis Thuringis ... Norsavorum gentis nobis placata majestas colla subdidit."^[16]

§ 116. What were the Werini, with whom the Angles were connected in the Leges Anglorum et Werinorum? Without having any particular data for connecting the Werini (Varni, Οὐάρνοι) with either the High-German, or the Mœso-Gothic divisions, there are in favour of their being Slavonic in locality, the same facts as applied to the Suevi and Langobardi, with the additional one, that the name probably exists at present in the River Warnow, of Mecklenburg Schwerin, at the mouth of which (Warnemunde) the town of Rostock stands.

§ 117. What were the Thuringians, with whom the Angles are connected in the Leges Anglorum, &c.; Germanic in locality, and most probably allied to the Goths of Mœsia in language.

§ 118. Of the Reudigni, Eudoses, Nuithones, Suardones, and Aviones, too little is known in detail to make the details an inquiry of importance. Respecting them all, it may be said at once, that whatever may be the Germanic affinities involved in their connection with the Suevi, Langobardi, Angli, &c., they are traversed by the fact of their locality being in the tenth century Slavonic.

§ 119. The last tribe which will be mentioned, is that of the Angrarii, most probably another form of the Angrivarii of Tacitus, the name of the occupants of the valley of the Aller, the northern confluent of the Weser.

As this word is compound (-varii=ware=inhabitants), the root remains Angr-, a word which only requires the r to become l in order to make Angl-. As both the locality and the relation to the Saxons, make the Angrivarian locality one of the best we could assume for the Angles, the only {68}difficulty lies in the change from r to l. Unfortunately, this, in the Saxon-German, is an unlikely one.

§ 120. The last fact connected with the Angles, will be found in a more expanded form in the Chapter on the Dialects of the English Language. It relates to the distribution over the conquered parts of Britain. Their chief area was the Midland and Eastern counties, Norfolk, Suffolk, Cambridgeshire, Huntingdonshire, Leicestershire, &c., rather than the parts south of the Thames, which were Saxon, and those north of the Wash, where Danish influences have been considerable.

§ 121. The reader has now got a general view of the extent to which the position of the Angles, as a German tribe, is complicated by conflicting statements; statements which connect them with (probably) High-German Thuringians, Suevi, and Langobardi, and with (probably) Slavonic Varni, Eudoses, Suardones, &c.; whereas in England, they are scarcely distinguishable from the Low-German Saxons. In the present state of our knowledge, the only safe fact seems to be, that of the common relation of both Angle and Saxon, to the present English of England.

This brings the two sections within a very close degree of affinity, and makes it probable, that just, as at present, descendants of the Saxons are English (Angle) in Britain, so, in the third and fourth centuries, ancestors of the Angles were Saxons in Germany. Why, however, the one name preponderated on the Continent, and the other in England is difficult to ascertain.

§ 122. By considering the Angles as Saxons under another name (or vice versâ), and by treating the statement as to the existence of Jutes in Hampshire and the Isle of Wight as wholly unhistorical, we get, as a general expression for the Anglo-Germanic immigration, that it consisted of the closely allied tribes of the North-Saxon area, an expression that implies a general uniformity of population. Is there reason to think that the uniformity was absolute?

§ 123. The following series of facts, when put together, will prepare us for a fresh train of reasoning concerning the different geographical and ethnological relations of the {69}immigrants into England, during their previous habitation in Germany.

1. The termination -as is, like the -s in the modern English, the sign of the plural number in Anglo-Saxon.

2. The termination -ing denotes, in the first instance, a certain number of individuals collected together, and united with each other as a clan, tribe, family, household.

3. In doing this, it generally indicates a relationship of a personal or political character. Thus two Baningas might be connected with each other, and (as such) indicated by the same term from any of the following causes—relationship, subordination to the same chief, origin from the same locality, &c.

4. Of these personal connections, the one which is considered to be the commonest is that of descent from a common ancestor, so that the termination -ing in this case, is a real patronymic.

5. Such an ancestor need not be real; indeed, he rarely if ever is so. Like the eponymus of the classical writers, he is the hypothetical, or mythological, progenitor of the clan, sept, or tribe, as the case may be; i.e., as Æolus, Dorus, and Ion to the Æolians, Dorians, and Ionians.

Now, by admitting these facts without limitation, and by applying them freely and boldly to the Germanic population of England, we arrive at the following inferences.

1. That where we meet two (or more) households, families, tribes, clans, or septs of the same name (that name ending in -ing), in different parts of England, we may connect them with each other, either directly or indirectly; directly when we look on the second as an offset from the first; indirectly, when we derive both from some third source.

2. That when we find families, tribes, &c., of the same name, both in Britain and in Germany, we may derive the English ones from the continental.

Now neither of these views is hypothetical. On the contrary each is a real fact. Thus in respect to divisions of the population, designated by names ending in -ing, we have

7. In Kent, Devonshire, Lincolnshire, Herefordshire, Salop, and Somerset,—Beringas.

8. In Bedford, Durham, Kent, Lancashire, Lincolnshire, Norfolk, Northamptonshire, Northumberland, Salop, Sussex, and the Isle of Wight,—Billingas, &c.—the list being taken from Mr. Kemble, vol. i. p. 64.

§ 124. On the other hand, the following Anglo-Saxon names in -ing, reappear in different parts of Germany, sometimes in definite geographical localities, as the occupants of particular districts, sometimes as mentioned in poems without further notice.

1. Wælsingas,—as the Volsungar of the Iceland, and the Wælsingen of the German heroic legends.

2. Herelingas,—mentioned in the Anglo-Saxon poem known by the name of the Traveller's Song, containing a long list of the Gothic tribes, families, nations, &c.

§ 125. We will still, for argument's sake, and for the sake {71}of the illustration of an ethnological method, take these names along with the observations by which they were preceded, as if they were wholly unexceptionable; and, having done this, ask how far each is known as German. So doing, we must make two divisions:

§ 126. Patronymics which do not necessarily denote a non-Saxon element.—Of these, the following are so little known, that they may pass as Saxons, simply because we have no grounds for thinking them aught else; the Brentings, Banings, Helsings, Serings, Ardings, Hundings, Blekings, Herelings, Gytings, Scydings, Dylings. The Scyldings and Scefings, belong, in a more positive way, to the Anglo-Saxon division; since their eponymi, Scyld and Sceaf, form a portion of the Anglo-Saxon mythology.

§ 127. Patronymics indicating a non-Saxon, rather than a Saxon element.—a. The Wælsings—In the way of tradition and mythology, this is a Frank gentile name.

c. The Hocings.—This is the German form of the Chauci, and, as such, a Frisian gentile name.

Thus, then, if we still assume that the method in question is unexceptionable, we have, from the evidence of what may be called either the gentile forms, or the patronymics in -ing, reasons for believing that Frank Myrgings, Frisian Hocings, and Thuringian Dhyrings, formed part of the invasion—these, at least; possibly others besides.

And why should the reason be other than unexceptionable? Do we not in North America, believe, that, as a general rule, the families with particular names, coincide with the families so-called in England; that the names of certain places, sometimes, at least, indicate a population originating in places similarly designated here? that the Smiths and Johnstons {72}are English in origin, and that O'Connors and O'Neils are Irish? We certainly believe all this, and, in many cases, we believe it, on the ground of the identity of name only.

§ 128. Exceptions.—Still there are exceptions. Of these the most important are as follows:—

1. The termination -ing is sometimes added to an undoubtedly British root, so as to have originated within the island, rather than to have been brought from the continent, e.g., the Kent-ings=the people of Kent. In such a case, the similarity to a German name, if it exist at all, exists as an accident.

2. The same, or nearly the same, name may not only occur in different parts of one and the same division of the Germanic areas, but in different ones, e.g., the Dhyrings may denote the Thuringians of Thuringia; but they may also denote the people of a district, or town, in Belgium, designated as Dorringen.^[17]

Still as a method, the one in question should be understood; although it has been too short a time before the learned world to have borne fruit.

N.B.—What applies to the coincidence of gentile or patronymic names on the two sides of the water, applies also to dialects; e.g., if (say) the Kentish differed from the other dialects of England, just in the same way, and with the same peculiar words and forms, as (say) the Verden dialect differed from the ones of Germany, we might fairly argue, that it was from the district of Verden that the county of Kent is peopled. At present we are writing simply for the sake of illustrating certain philological methods. The question of dialect will be treated in Part VII.

§ 129. German tribes where there is no direct evidence as to their having made part of the population of England, but where the à priori probabilities are strongly in their favour. This applies to—a. The Batavians. No direct evidence, but great à priori probability.

b. The Frisians.—Great à priori probability, and {73}something more; Βριττίαν δὲ τὴν νῆσον ἔθνη τρία πολυανθρωπότατα ἔχουσι, βασιλεύς τε ἑῖς αὐτῶν ἑκάστῳ ἐφέστηκεν, ὀνόματα δὲ κεῖται τοῖς ἔθνεσι τούτοις Ἀγγίλοι τε καὶ Φρίσσονες καὶ οἱ τῇ νήσῳ ὁμώνυμοι Βρίττωνες. Τοσαύτη δὲ ἡ τῶνδε τῶν ἐθνῶν πολυανθρωπία φαίνεται οὖσα ὥστε ἀνὰ πᾶν ἔτος κατὰ πολλοὺς ἐνθένδε μετανιστάμενοι ξὺν γυναιξὶ καὶ παισὶν ἐς Φράγγους χώρουσιν.^[18]—Procop. B. G. iv. 20.

§ 130. I believe, for my own part, there were portions in the early Germanic population of Britain, which were not strictly either Angle or Saxon (Anglo-Saxon); but I do this without thinking that it bore any great ratio to the remainder, and without even guessing at what that ratio was, or whereabouts its different component elements were located—the Frisians and Batavians being the most probable. With this view, there may have been Jutes as well; notwithstanding what has been said in §§ 16-20; since the reasoning there is not so against a Jute element in toto, as against that particular Jute element, in which Beda, Alfred, and the later writers believed and believe.

§ 131. No exception against the existence of Batavian, Frisian, Frank, and other elements not strictly Anglo-Saxon, is to be taken from the absence of traces of such in the present language, and that for the following reason. Languages which differ in an older form may so far change according to a common principle, as to become identical in a newer one. E.g., the Frisian infinitive in verbs ends in -a, (as bærna=to burn), the Saxon in -an (as bærnan=to burn). Here is a difference. Let, however, the same change affect both languages; that change being the abandonment, on both sides, of the infinitive termination altogether. What follows? even that the two originally different forms bærn-a, and bærn-an, both come out bærn (burn); so that the result is the same, though the original forms were different.

CHAPTER VI.

§ 132. The languages of Great Britain at the invasion of Julius Cæsar were of the Celtic stock.

1. The British or Cambrian branch, represented by the present Welsh, and containing, besides, the Cornish of Cornwall (lately extinct) and the Armorican of the French province of Brittany. It is almost certain that the old British, the ancient language of Gaul, and the Pictish were of this branch.

2. The Gaelic or Erse Branch, represented by the present Irish Gaelic, and containing, besides, the Gaelic of the Highlands of Scotland and the Manks of the Isle of Man.

§ 133. Taken altogether the Celtic tongues form a very remarkable class. As compared with those of the Gothic stock they are marked by the following characteristics—

The scantiness of the declension of Celtic nouns.—In Irish there is a peculiar form for the dative plural, as cos=foot, cos-aibh=to feet (ped-ibus); and beyond this there is nothing else whatever in the way of case, as found in the German, Latin, Greek, and other tongues. Even the isolated form in question is not found in the Welsh and Breton. Hence {78}the Celtic tongues are preeminently uninflected in the way of declension.

§ 134.—2. The agglutinate character of their verbal inflections.—In Welsh the pronouns for we, ye, and they, are ni, chwyi, and hwynt respectively. In Welsh also the root=love is car. As conjugated in the plural number this is—

Now the -wn, -ych, and -ant, of the persons of the verbs are the personal pronouns, so that the inflection is really a verb and a pronoun in a state of agglutination; i. e., in a state where the original separate existence of the two sorts of words is still manifest. This is probably the case with languages in general. The Celtic, however, has the peculiarity of exhibiting it in an unmistakable manner; showing, as it were, an inflexion in the process of formation, and (as such) exhibiting an early stage of language.

§ 135. The system of initial mutations.—The Celtic, as has been seen, is deficient in the ordinary means of expressing case. How does it make up for this? Even thus. The noun changes its initial letter according to its relation to the other words of the sentence. Of course this is subject to rule. As, however, I am only writing for the sake of illustrating in a general way the peculiarities of the Celtic tongues, the following table, from Prichard's Eastern Origin of the Celtic Nations, is sufficient.

§ 136. When we have seen that one of the great characteristics of the Celtic tongues is to express inflection by initial changes, we may ask how far the principle of such change is common to the two branches—British or Gaelic; this and a few other details being quite sufficient to show the affinity between them.

The changes in Welsh, classified according to the relationship of the sounds are—

1. From the sharp lenes to the corresponding flats; as p to b, t to d, c to g. The changes in Irish are the same.

2. From the flat lenes to their corresponding so-called aspirates; as b to v, d to ð. This is the change in Welsh. In Irish we have the same, but only as far as b is concerned; the aspirate of d (ð) being wanting in that language. In neither Welsh nor Irish occurs the true aspirate of g. In neither Welsh nor Irish occurs the true aspirate of c; which, being wanting, is replaced by the sound of the ch in the German auch, here spelt ç.

Now the Welsh grammarians deal with the changes from sharp to flat, and from lene to aspirate, alike; since, in respect to the grammar of their language, they are enabled to state that they take place under the same circumstances. {80}Taken collectively they are called light: and words wherein p is changed to b, and those wherein b is changed to v, are equally said to assume the light sound. This the Welsh express in spelling, and write ben for pen, and vraint for braint, &c. In Irish the arrangement is different. When a so-called aspirate is substituted for a lene, the word is said to take an aspiration, and bheul is written beul. If, however, the sharp be made flat, the original sound is said to be eclipsed. In spelling, however, it is preserved; so that teine, with the t changed, is written dteine, and pronounced deine. With this view we can now ask how far the change from p to b, t to d, c to g, b to v, c to ç, takes place in Irish and Welsh under similar circumstances.

In Welsh—after all verbs, except those of the infinitive mood; as caravi gaer (for caer)=I love a fort.

In Irish—after all verbs, provided that the substantive be masculine; as ta me ag gearrad çrainn=I am cutting (at to cut) a tree. Here çrainn comes from crainn. This change in Irish extends only to the change from lene to aspirate.

In Welsh—after the possessive pronouns thy, thine, his, its, mine (but not my); as dy vâr (for bâr)=thy wrath; ei vraint (from braint)=his privilege. N. B. Although the same word (ei) means her, his, and its, it induces the light change only when it is either masculine or neuter.

In Irish—after the possessive pronouns my, thy, and his. Here the change is of the first sort only, or an aspiration; as mo vàs (bàs)=my death; do ços (cos)=thy foot; çeann (ceann)=his head. N. B. Although the same word (a) means her, his, and its, it induces the aspirate only when it is either masculine or neuter.

In Welsh—the initials of adjectives become light when their substantive is feminine.

In Irish—the initials of adjectives singular, aspirated in the oblique cases only of the masculine, are aspirated throughout in the feminine.

In Welsh—after certain adverbs called formative, used like the English words to, as, &c., in the formation of the degrees of nouns, and the moods of verbs (in other words, {81}after certain particles), initial sounds become light; as rhy vyçan (byçan)=very (over) little; ni çarav (carav)=I do not love.

In Irish—the same, in respect to the change from lene to aspirate; ro veag=very little; ni vualim (bualim)=I do not beat; do vuaileas=I struck, &c.

In Irish—the prepositions either eclipse the noun that they govern or else aspirate it. A Welsh grammarian would say that it made them light.

In Irish—masculines are aspirated in the genitive and dative singular; feminines in the nominative and dative. N.B. The difference here is less than it appears to be. The masculine dative is changed, not as a masculine, but by the effect of the particle do, the sign of the dative; the genitive, perhaps, is changed ob differentiam. This being the fact, the nominative is the only case that is changed as such. Now this is done with the feminines only. The inflection explains this.

Such the changes from sharp to flat, and from lene to aspirate. The second order of changes is remarkable, viz. from the mutes to their corresponding liquids, and, in the case of series k, to ng. This, in Welsh, is as follows:—

In Irish the combinations m + h, n + h, ng + h are wanting: t, however, under certain conditions, becomes h, as mo high (tigh)=my house. With the unaspirated liquids the change, however, coincides with that of the Welsh—ar maile (spelt mbaile)=our town; ar nia (spelt ndia)=our God; ar ngearran=our complaint. These words come respectively from baile, dia, gearran. To show that this change takes place in Irish and Welsh under similar circumstances is more than can be expected; since ð being wanting in Irish, leaves d to be changed into n.

Plural forms.—When not expressed by a change of vowel, -d (or an allied sound) both in Welsh and Irish has a plural power; as merç, merçed; hyð, hyðoð; teyrn, teyrneð=girls, stags, kings; Welsh:—gealaç, gealaçad; sgolog, sgolagad; uiseog, uiseogad=moons, farmers, larks; Irish. In each language there are plural forms in -d.

Also in -n, as dyn=a person, dynion=persons. In Irish there is the form cu=a greyhound; Plural cuin. It may be doubted, however, whether -n is not ejected in the singular rather than added in the plural.

Also in -au, Welsh (as pén-au=heads), and in -a, Irish (as cos-a=feet).

In each language there is, in respect to both case and {83}gender, an equal paucity of inflections. The Irish, however, preserves the Indo-European dative plural in b; as ços-aiv=ped-ibus.

The ordinals in Welsh are expressed by -ved; as saiþ=seven, seiþved=seventh. The ordinals in Irish are expressed by -vad, as seaçt=seven, seaçt-vad=seventh (spelt seachmhadh).

The terminations -n and -g are diminutive in Welsh; as dyn-yn=mannikin, oen-ig=lambkin. They have the same power in Irish; as cnoc-an=a hillock; duil-eog=a leaflet. In Irish, currently spoken, there is no inflection for the comparative degrees;—there is, however, an obsolete form in -d, as glass, glaiside=green, greener. In Welsh the true comparative ends in ç, as main=slender, mainaç=more slender. A form, however, exists in -ed, meaning equality, and so implying comparison, viz., mein-ed=so slender.

As expressive of an agent, the termination -r is common to both languages. Welsh, mor-ûr=a seaman; telynaur=a harpist; Irish, sealg-aire=a hunter; figead-oir=a weaver.

As expressive of "abounding in," the termination -c (or -g) is common in both languages. Welsh, boliûag=abounding in belly; toirteaç=abounding in fruit. In each language a sound of series t, is equivalent to the English -ly. Welsh, mab-aið=boy-like. Irish, duin-eata=manly.

Of the personal terminations it may be said, that those of both the Irish and Welsh are those of the other European tongues, and that they coincide and differ in the same way with those of the Gothic stock: the form in m being the one more constant. For the theory of the personal terminations, the reader is referred to the Eastern Origin of the Celtic Nations, by Dr. Prichard.

The present notices being indicative of grammatical affinities only, the glossarial points of likeness between the Welsh and Irish are omitted.

§ 137. The Celtic tongues have lately received especial illustration from the researches of Mr. Garnett. Amongst other, the two following points are particularly investigated by him:— {84}

§ 138. The ancient language of Gaul Cambrian.—The evidence in favour of the ancient language of Gaul being Cambrian rather than Gaelic, lies in the following facts:—

a. Petorritum=a four-wheeled carriage, from the Welsh, peaer=four, and rhod=a wheel. The Gaelic for four is ceathair, and the Gaelic compound would have been different.

b. Pempedula, the cinque-foil, from the Welsh pump=five, and dalen=a leaf. The Gaelic for five is cuig, and the Gaelic compound would have been different.

c. Candetum=a measure of 100 feet, from the Welsh cant=100. The Gaelic for a hundred is cead, and the Gaelic compound would have been different.

d. Epona=the goddess of horses. In the Old Armorican the root ep=horse. The Gaelic for a horse is each.

e. The evidence from the names of geographical localities in Gaul, both ancient and modern, goes the same way: Nantuates, Nantouin, Nanteuil, are derived from the Welsh nant=a valley, a word unknown in Gaelic.

f. The evidence of certain French provincial words, which are Welsh and Armorican rather than Erse or Gaelic.

g. An inscription on an ancient Celtic tablet found at Paris, A.D. 1711, and representing a bull and three birds (cranes), is TARWOS TRI GARANOS. Now, for the first two names, the Gaelic affords as good an explanation as the Welsh; the third, however, is best explained by the Welsh.

§ 139. The Pictish most probably Cambrian.—The evidence in favour of the Pictish being Cambrian rather than Gaelic lies in the following facts:—

a. When St. Columba preached, whose mother-tongue was Irish Gaelic, he used an interpreter—Adamnanus apud {85}Colgarum, 1, 11, c.32. This is a point of external evidence, and shows the difference between the Pict and Gaelic. What follows are points of internal evidence, and show the affinity between the Pict and Welsh.

b. A manuscript in the Colbertine library contains a list of Pictish kings from the fifth century downwards. These names are not only more Celtic than Gothic, but more Welsh than Gaelic. Taran=thunder in Welsh. Uven is the Welsh Owen. The first syllable in Talorg (=forehead) is the tal in Talhaiarn=iron forehead, Taliessin=splendid forehead, Welsh names. Wrgust is nearer to the Welsh Gwrgust than to the Irish Fergus. Finally, Drust, Drostan, Wrad, Necton, closely resemble the Welsh Trwst, Trwstan, Gwriad, Nwython. Cineod and Domhnall (Kenneth and Donnell), are the only true Erse forms in the list.

c. The only Pictish common name extant is the well-known compound pen val, which is in the oldest MS. of Bede peann fahel. This means caput valli, and is the name for the eastern termination of the Vallum of Antoninus. Herein pen is unequivocally Welsh, meaning head. It is an impossible form in Gaelic. Fal, on the other hand, is apparently Gaelic, the Welsh for a rampart being gwall. Fal, however, occurs in Welsh also, and means inclosure.

The evidence just indicated is rendered nearly conclusive by an interpolation, apparently of the twelfth century, of the Durham MS. of Nennius, whereby it is stated that the spot in question was called in Gaelic Cenail. Now Cenail is the modern name Kinneil, and it is also a Gaelic translation of the Pict pen val, since cean is the Gaelic for head, and fhail for rampart or wall. If the older form were Gaelic, the substitution, or translation, would have been superfluous.

d. The name of the Ochil Hills in Perthshire is better explained from the Pict uchel=high, than from the Gaelic uasal.

e. Bryneich, the British form of the province Bernicia, is better explained by the Welsh bryn=ridge (hilly country), than by any word in Gaelic.—Garnett, in Transactions of Philological Society.

CHAPTER VII.

The Greek and its dialects, both ancient and modern, constitute the Greek or Hellenic branch of the Classical stock.

The Latin in all its dialects, the old Italian languages allied to it, and the modern tongues derived from the Roman, constitute the Latin or Ausonian branch of the Classical stock.

Now, although the Greek or Hellenic dialects are of secondary importance in the illustration of the history of the English language, the Latin or Ausonian elements require a special consideration.

The French element appeared in our language as a result of the battle of Hastings (A.D. 1066), perhaps, in a slight degree, at a somewhat earlier period.

§ 141. Previous to the notice of the immediate relations of the Norman-French, or, as it was called after its introduction into England, Anglo-Norman, its position in respect to the other languages derived from the Latin may be exhibited.

The Latin language overspread the greater part of the Roman empire. It supplanted a multiplicity of aboriginal languages; just as the English of North America has supplanted the aboriginal tongues of the native Indians, and just as the Russian is supplanting those of Siberia and Kamskatcha.

Sometimes the war that the Romans carried on against the old inhabitants was a war of extermination. In this case the original language was superseded at once. In other cases their influence was introduced gradually. In this case the influence of the original language was greater and more permanent. {87}

Just as in the United States the English came in contact with an American, whilst in New Holland it comes in contact with an Australian language, so was the Latin language of Rome engrafted, sometimes on a Celtic, sometimes on a Gothic, and sometimes on some other stock. The nature of the original language must always be borne in mind.

From Italy, its original seat, the Latin was extended in the following chronological order:—

1. To the Spanish Peninsula; where it overlaid or was engrafted on languages allied to the present Biscayan (i.e., languages of the Iberic stock), mixed in a degree (scarcely determinable) with Celtic elements=Celtiberic.

2. To Gaul, or France, where it overlaid or was engrafted on languages of the Celtic stock. This took place, at least for the more extreme parts of Gaul, in the time of Julius Cæsar; for the more contiguous parts, in the earlier ages of the Republic.

3. To Dacia and Pannonia; where it overlaid or was engrafted on a language the stock whereof is undetermined. The introduction of the Latin into Dacia and Pannonia took place in the time of Trajan.

From (1stly,) the original Latin of Italy, and from the imported Latin, of (2ndly,) the Spanish Peninsula, (3rdly,) Gaul, (4thly,) Dacia and Pannonia, we have (amongst others) the following modern languages—1st Italian, 2nd Spanish and Portuguese, 3rd French, 4th Wallachian. How far these languages differ from each other is currently known. One essential cause of this difference is the difference of the original language upon which the Latin was engrafted.

§ 142. I am not doing too much for the sake of system if I classify the languages, of which the Italian, French, &c., are the representatives, as the languages of Germany were classified, viz., into divisions.

I. The Spanish and Portuguese are sufficiently like the Italian to be arranged in a single division. This may conveniently be called the Hesperian division.

II. The second division is the Transalpine. This comprises the languages of Gaul, viz., the Modern French, the {88}Anglo-Norman, and the Provençal. It also includes a language not yet mentioned, the Romanese (Rumonsch), or the language of the Grisons, or Graubünten, of Switzerland.

III. The third division is the Dacian, Pannonian, or Wallachian, containing the present languages of Wallachia and Moldavia.

In the Jahrbücher der Literatur, June, 1829, specimens are given of two of its dialects: 1, the Daco-Wallachian, north of the Danube; 2, the Macedono-Wallachian, south of the Danube. The present specimen varies from both. It is taken from the New Testament, printed at Smyrna, 1838. The Dacian division is marked by placing the article after the noun, as homul=the man=homo ille.

§ 143. Such is the general view of the languages derived from the Latin, i.e., of the languages of the Latin branch of the Classical stock.

The French languages of the Transalpine division require to be more minutely exhibited.

Between the provincial French of the north and the provincial French of the south, there is a difference, at the present day, at least of dialect, and perhaps of language. This is shown by the following specimens: the first from the canton of Arras, on the confines of Flanders; the second, from the department of Var, in Provence. The date of each is A.D. 1807.

Practically speaking, although in the central parts of France the northern and southern dialects melt each into the other, the Loire may be considered as a line of demarcation between two languages; the term language being employed because, in the Middle Ages, whatever may be their real difference, the northern tongue and the southern tongue were dealt with not as separate dialects, but as distinct languages—the southern being called Provençal, the northern Norman-French.

Of these two languages (for so they will in the following pages be called, for the sake of convenience) the southern or Provençal approaches the dialects of Spain; the Valencian of Spain and the Catalonian of Spain being Provençal rather than standard Spanish or Castilian.

The southern French is sometimes called the Langue d'Oc, and sometimes the Limousin.

It is in the Southern French (Provençal, Langue d'Oc, or Limousin) that we have the following specimen, viz., the Oath of Ludwig, sworn A.D. 842.

§ 144. The Norman-French, spoken from the Loire to the confines of Flanders, and called also the Langue d'Oyl, differed from the Provençal in (amongst others) the following circumstances.

1. It was of later origin; the southern parts of Gaul having been colonized at an early period by the Romans.

2. It was in geographical contact, not with the allied languages of Spain, but with the Gothic tongues of Germany and Holland.

It is the Norman-French that most especially bears upon the history of the English language.

The proportion of the original Celtic in the present languages of France has still to be determined. It may, however, be safely asserted, that at a certain epoch between the first and fifth centuries, the language of Gaul was more Roman and less Celtic than that of Britain.

In the northern French we must recognise not only a Celtic and a Classical, but also a Gothic element: since Clovis and Charlemagne were no Frenchmen, but Germans; their language being High-Germanic. The High-Germanic element in French has still to be determined.

In the northern French of Normandy there is a second Gothic element, viz., a Scandinavian element. By this the proper northern French underwent a further modification.

Until the time of the Scandinavians or Northmen, the present province of Normandy was called Neustria. A generation before the Norman Conquest, a Norwegian captain, named in his own country Rolf, and in France Rollo, or Rou, settled upon the coast of Normandy. What Hengist and the Germans are supposed to have been in Britain, Rollo and his Scandinavians were in France. The province took from them its name of Normandy. The Norwegian element in the Norman-French has yet to be determined. Respecting it, however, the following statements may, even in the present state of the question, be made:—

1. That a Norse dialect was spoken in Normandy at Bayeux, some time after the battle of Hastings.

3. That the names Jersey, Guernsey, and Alderney are as truly Norse names as Orkney and Shetland.

CHAPTER VIII.

§ 145. In each of the three preceding chapters a separate stock of languages has been considered; and it has been shown, in some degree, how far languages of the same stock differ from, or agree with, each other.

Furthermore, in each stock there has been some particular language that especially illustrates the English.

In the Gothic stock there has been the Anglo-Saxon; in the Celtic the Welsh; and in the Classical the Anglo-Norman.

Nevertheless, the importance of the languages of these three divisions is by no means equal. The Gothic tongues supply the basis of our investigations. The Celtic afford a few remnants of that language which the Anglo-Saxon superseded. The Anglo-Norman language exhibits certain superadded elements.

§ 146. Over and above the Gothic, Celtic, and Classical languages, there are others that illustrate the English; and some of our commonest grammatical inflections can be but half understood unless we go beyond the groups already enumerated.

The Gothic, Celtic (?),^[20] and Classical stocks are but subordinate divisions of a wider class. Each has a sufficient amount of mutual affinities to be illustrative of each other, and each is contained, along with two other groups of equal value, under a higher denomination in philology.

What is the nature of that affinity which connects languages so different as the Gothic, Celtic (?), and Classical stocks? or what is the amount of likeness between, e.g., the {94}German and Portuguese, the Greek and Islandic, the Latin and Swedish, the Anglo-Saxon and Italian? And what other languages are so connected?

What other philological groups are connected with each other, and with the languages already noticed, by the same affinities which connect the Gothic, Celtic (?), and Classical stocks? Whatever these languages may be, it is nearly certain that they will be necessary, on some point or other, for the full illustration of the English.

As both these questions are points of general, rather than of English, philology, and as a partial answer may be got to the first from attention to the degree in which the body of the present work exhibits illustrations drawn from widely different languages, the following statements are considered sufficient.

§ 147. The philological denomination of the class which contains the Gothic, Celtic (?), and Classical divisions, and, along with the languages contained therein, all others similarly allied, is Indo-European; so that the Gothic, Celtic (?), Classical and certain other languages are Indo-European.

The other divisions of the great Indo-European group of languages are as follows:—

1. The Iranian stock of languages.—This contains the proper Persian languages of Persia (Iran) in all their stages, the Kurd language, and all the languages of Asia (whatever they may be) derived from the Zend or Sanskrit.

2. The Sarmatian stock of languages.—This contains the languages of Russia, Poland, Bohemia, and of the Slavonian tribes in general. It contains also the Lithuanic languages, i.e., the Lithuanic of Lithuania, the old Prussian of Prussia (now extinct), and the Lettish or Livonic of Courland and Livonia.

3, 4, 5. The Classical, Gothic, and Celtic (?) stocks complete the catalogue of languages undoubtedly Indo-European, and at the same time they explain the import of the term. Indo-European is the name of a class which embraces the majority of the languages of Europe, and is extended over {95}Asia as far as India. Until the Celtic was shown by Dr. Prichard to have certain affinities with the Latin, Greek, Slavonic, Lithuanic, Gothic, Sanskrit, and Zend, as those tongues had with each other, the class in question was called Indo-Germanic; since, up to that time, the Germanic languages had formed its western limit.

§ 148. Meaning of the note of interrogation (?) after the word Celtic.—In a paper read before the Ethnological Society, February 28th, 1849, and published in the Edinburgh Philosophical Magazine, the present writer has given reasons for considering the claims of the Celtic to be Indo-European as somewhat doubtful; at the same time he admits, and highly values, all the facts in favour of its being so, which are to be found in Prichard's Eastern Origin of the Celtic Nations.

He believes, however, that the Celtic can only be brought in the same group with the Gothic, Slavonic, &c., by extending the value of the class.

"To draw an illustration from the common ties of relationship, as between man and man, it is clear that a family may be enlarged in two ways.

"a. A brother, or a cousin, may be discovered, of which the existence was previously unknown. Herein the family is enlarged, or increased, by the real addition of a new member, in a recognised degree of relationship.

"b. A degree of relationship previously unrecognised may be recognised, i.e., a family wherein it was previously considered that a second-cousinship was as much as could be admitted within its pale, may incorporate third, fourth, or fifth cousins. Here the family is enlarged, or increased, by a verbal extension of the term.

"Now it is believed that the distinction between increase by the way of real addition, and increase by the way of verbal extension, has not been sufficiently attended to. Yet, that it should be more closely attended to, is evident; since, in mistaking a verbal increase for a real one, the whole end and aim of classification is overlooked. The publication of Dr. Prichard's Eastern Origin of the Celtic Nations, in 1831, {96}supplied philologists with the most definite addition that has perhaps, yet been made to ethnographical philology.

"Ever since then the Celtic has been considered to be Indo-European. Indeed its position in the same group with the Iranian, Classical, Slavono-Lithuanic, and Gothic tongues, supplied the reason for substituting the term Indo-European for the previous one Indo-Germanic.

"On the other hand, it seems necessary to admit that languages are allied just in proportion as they were separated from the mother-tongue in the same stage of its development.

"If so, the Celtic became detached anterior to the evolution of the declension of nouns, whereas the Gothic, Slavonic, Classical and Iranian languages all separated subsequent to that stage."^[21]

This, along with other reasons indicated elsewhere,^[22] induces the present writer to admit an affinity between the Celtic and the other so-called Indo-European tongues, but to deny that it is the same affinity which connects the Iranian, Classical, Gothic and Slavonic groups.

PART II.

CHAPTER I.

1. Those that are of late introduction, and cannot be called original and constituent parts of the language. Such are (amongst others) the words flannel, crowd (a fiddle), from the Cambrian; and kerne (an Irish foot-soldier), galore (enough), tartan, plaid, &c., from the Gaelic branch.

2. Those that are common to both the Celtic and Gothic stocks, and are Indo-European rather than either Welsh, or Gaelic, or Saxon. Such (amongst others) are brother, mother, in Celtic brathair, mathair; the numerals, &c.

3. Those that have come to us from the Celtic, but have come to us through the medium of another language. Such are druid and bard, whose immediate source is, not the Celtic but, the Latin.

4. Celtic elements of the Anglo-Norman, introduced into England after the Conquest, and occurring in that language as remains of the original Celtic of Gaul.

5. Those that have been retained from the original Celtic of the island, and which form genuine constituents of our language. These fall into three subdivisions.

a. Proper names—generally of geographical localities; as the Thames, Kent, &c. {98}

b. Common names retained in the provincial dialects of England, but not retained in the current language; as gwethall=household stuff, and gwlanen=flannel in Herefordshire.

c. Common names retained in the current language.—The following list is Mr. Garnett's:—

§ 150. Latin of the first period.—Of the Latin introduced by Cæsar and his successors, the few words remaining are those that relate to military affairs; viz. street (strata); coln (as in Lincoln=Lindi colonia); cest (as in Gloucester=glevæ castra) from castra. The Latin words introduced between the time of Cæsar and Hengist may be called the Latin of the first period, or the Latin of the Celtic period.

§ 151. The Anglo-Saxon.—This is not noticed here, because from being the staple of the present language it is more or less the subject of the book throughout.

§ 152. The Danish, or Norse.—The pirates that pillaged Britain, under the name of Danes, were not exclusively the inhabitants of Denmark. Of the three Scandinavian nations, the Swedes took the least share, the Norwegians the greatest {99}in these invasions. Not that the Swedes were less piratical, but that they robbed elsewhere,—in Russia, for instance, and in Finland.

The language of the three nations was the same; the differences being differences of dialect. It was that which is now spoken in Iceland, having been once common to Scandinavia and Denmark. Whether this was aboriginal in Denmark, is uncertain. In Scandinavia it was imported; the tongue that it supplanted having been, in all probability, the mother-tongue of the present Laplandic.

The Danish that became incorporated with our language, under the reign of Canute and his sons, may be called the direct Danish (Norse or Scandinavian) element, in contradistinction to the indirect Danish of §§ 144, 155.

The determination of the amount of Danish in English is difficult. It is not difficult to prove a word Scandinavian. We must also show that it is not German. A few years back the current opinion was against the doctrine that there was much Danish in England. At present, the tendency is rather the other way. The following facts are from Mr. Garnett.—Phil. Trans. Vol. i.

1. The Saxon name of the present town of Whitby in Yorkshire was Streoneshalch. The present name Whitby, Hvitby, or White-town, is Danish.

2. The Saxon name of the capital of Derbyshire was Northweortheg. The present name is Danish.

4. On a monument in Aldburgh church, Holdernesse, in the East Riding of Yorkshire, referred to the age of Edward the Confessor, is found the following inscription:—

Now, in this inscription, Ulf, in opposition to the Anglo-Saxon wulf, is a Norse form; whilst hanum is a Norse dative, and by no means an Anglo-Saxon one.—Old Norse hanum, Swedish honom.

5. The use of at for to as the sign of the infinitive mood {100}is Norse, not Saxon. It is the regular prefix in Icelandic, Danish, Swedish, and Feroic. It is also found in the northern dialects of the Old English, and in the particular dialect of Westmoreland at the present day.

§ 153. Roman of the Second Period.—Of the Latin introduced under the Christianised Saxon sovereigns, many words are extant. They relate chiefly to ecclesiastical matters, just as the Latin of the Celtic period bore upon military affairs.—Mynster, a minster, monasterium; portic, a porch, porticus; cluster, a cloister, claustrum; munuc, a monk, monachus; bisceop, a bishop, episcopus; arcebisceop, archbishop, archiepiscopus; sanct, a saint, sanctus; profost, a provost, propositus; pall, a pall, pallium; calic, a chalice, calix; candel, a candle, candela; psalter, a psalter, psalterium; mæsse, a mass, missa; pistel, an epistle, epistola; prædic-ian, to preach, prædicare; prof-ian, to prove, probare.

The following are the names of foreign plants and animals:—camell, a camel, camelus; ylp, elephant, elephas; ficbeam, fig-tree, ficus; feferfuge, feverfew, febrifuga; peterselige, parsley, petroselinum.

Others are the names of articles of foreign origin, as pipor, pepper, piper; purpur, purple, purpura; pumicstan, pumice-stone, pumex. {101}

The above-given list is from Guest's English Rhythms (B. iii. c. 3). It constitutes that portion of the elements of our language which may be called the Latin of the second, or Saxon period.

§ 154. The Anglo-Norman element.—For practical purposes we may say that the French or Anglo-Norman element appeared in our language after the battle of Hastings, A.D. 1066.

Previous, however, to that period we find notices of intercourse between the two countries.

2. Ethelred II. married Emma, daughter of Richard Duke of Normandy, and the two children were sent to Normandy for education.

3. Edward the Confessor is particularly stated to have encouraged French manners and the French language in England.

6. The French article la, in the term la Drove, occurs in a deed of A.D. 975.—See Ranouard, Journal des Savans, 1830.

The chief Anglo-Norman elements of our language are the terms connected with the feudal system, the terms relating to war and chivalry, and a great portion of the law terms—duke, count, baron, villain, service, chivalry, warrant, esquire, challenge, domain, &c.

§ 155. The Norwegian, Danish, Norse, or Scandinavian element of the Anglo-Norman (as in the proper names Guernsey, Jersey, Alderney, and perhaps others) constitutes the indirect Scandinavian element of the English.

§ 156. Latin of the Third Period.—This means the Latin which was introduced between the battle of Hastings and the revival of literature. It chiefly originated with the monks, in the universities, and, to a certain extent, in the courts of law. It must be distinguished from the indirect Latin introduced as part and parcel of the Anglo-Norman. It has yet to be accurately analyzed. {102}

Latin of the Fourth Period.—This means the Latin which has been introduced between the revival of literature and the present time. It has originated in the writings of learned men in general, and is distinguished from that of the previous periods by—

2. Preserving, in the case of substantives, in many cases its original inflections; axis, axes; basis, bases—

3. Relating to objects and ideas for which the increase of the range of science in general has required a nomenclature.

§ 157. Greek.—Words derived directly from the Greek are in the same predicament as the Latin of the third period—phænomenon, phænomena; criterion, criteria, &c.; words which are only indirectly of Greek origin, being considered to belong to the language from which they were immediately introduced into the English. Such are deacon, priest, &c., introduced through the Latin; thus a word like church proves no more in regard to a Greek element in English, than the word abbot proves in respect to a Syrian one.

§ 158. The Latin of the fourth period and the Greek agree in retaining, in many cases, the Latin or Greek inflexions rather than adopting the English ones; in other words, they agree in being but imperfectly incorporated. The phænomenon of imperfect incorporation (an important one) is reducible to the following rules:—

1. That it has a direct ratio to the date of the introduction, i.e., the more recent the word the more likely it is to retain its original inflexion.

2. That it has a relation to the number of meanings belonging to the words: thus, when a single word has two meanings, the original inflexion expresses one, the English inflexion another—genius, genii, often (spirits), geniuses (men of genius).

3. That it occurs with substantives only, and that only in the expression of number. Thus, although the plural of substantives like axis and genius are Latin, the possessive cases are English. So also are the degrees of comparison, for {103}adjectives like circular, and the tenses, &c. for verbs, like perambulate.

§ 159. The following is a list of the chief Latin substantives, introduced during the latter part of the fourth period; and, preserving the Latin plural forms—

Words wherein the Latin Plural is formed from the Latin Singular by changing the last Syllable.

(a).—Where the Singular termination -a is changed in the Plural into -æ:—

(b).—Where the singular termination -us is changed in the Plural into -i:—

(c).—Where the Singular termination -um is changed in the Plural into -a:—

(d).—Where the singular termination -is is changed in the Plural into -es:—

Words wherein the Plural is formed by inserting -e between the last two sounds of the singular, so that the former number always contains a syllable more than the latter:—

In all these words the c of the singular number is sounded as k, of the plural as s.

§ 160. The following is a list of the chief Greek substantives lately introduced, and preserving the Greek plural forms—

Words where the plural is formed from the original root by adding either -es or -a, but where the singular rejects the last letter of the original root.

§ 161. Miscellaneous elements.—Of miscellaneous elements we have two sorts; those that are incorporated in our language, and are currently understood (e.g., the Spanish word sherry, the Arabic word alkali, and the Persian word turban), and those that, even amongst the educated, are considered strangers. Of this latter kind (amongst many others) are the Oriental words hummum, kaftan, gul, &c.

Of the currently understood miscellaneous elements of the English language, the most important are from the French; some of which agree with those of the Latin of the fourth period, and the Greek in preserving the French plural forms—as beau, beaux, billets-doux.

Italian.—Some words of Italian origin do the same: as virtuoso, virtuosi.

Hebrew.—The Hebrew words, cherub and seraph do the same; the form cherub-im, and seraph-im, being not only plurals but Hebrew plurals.

Beyond the words derived from these five languages, none form their plurals other than after the English method, i.e., in -s: as waltzes, from the German word waltz.

§ 162. The extent to which a language, which like the English, at one and the same time requires names for many objects, comes in contact with the tongues of half the world, {106}and has, moreover, a great power of incorporating foreign elements, derives fresh words from varied sources, may be seen from the following incomplete notice of the languages which have, in different degrees, supplied it with new terms.

Arabic.—Admiral, alchemist, alchemy, alcohol, alcove, alembic, algebra, alkali, assassin, from a paper of Mr. Crawford, read at the British Association, 1849.

Hindu languages.—Calico, chintz, cowrie, curry, lac, muslin, toddy, &c.—Ditto.

Tungusian, or some similar Siberian language.—Mammoth, the bones of which are chiefly from the banks of the Lena.

§ 163. In § 157 a distinction is drawn between the direct and indirect, the latter leading to the ultimate origin of words.

Thus a word borrowed into the English from the French, might have been borrowed into the French from the Latin, into the Latin from the Greek, into the Greek from the Persian, &c., and so ad infinitum.

The investigation of this is a matter of literary curiosity rather than any important branch of philology.

The ultimate known origin of many common words sometimes goes back to a great date, and points to extinct languages—

§ 164. Again, a word from a given language may be introduced by more lines than one; or it may be introduced twice over; once at an earlier, and again at a later period. In such a case its form will, most probably, vary; and, what is more, its meaning as well. Words of this sort may be called di-morphic, their di-morphism, having originated in one of two reasons—a difference of channel, or a difference of date. Instances of the first are, syrup, sherbet, and shrub, all originally from the Arabic, srb; but introduced differently, viz., the first through the Latin, the second through the Persian, and the third through the Hindoo. Instances of the second are words like minster, introduced in the Anglo-Saxon, as contrasted with monastery, introduced during the Anglo-Norman period. By the proper application of these processes, we account for words so different in present form, yet so identical in origin, as priest and presbyter, episcopal and bishop, &c.

§ 165. Distinction.—The history of the languages that have been spoken in a particular country, is a different subject from the history of a particular language. The history of the languages that have been spoken in the United States of America, is the history of Indian languages. The history of the languages of the United States is the history of the Germanic language.

§ 166. Words of foreign simulating a vernacular origin.—These may occur in any mixed language whatever; they occur, however, oftener in the English than in any other.

Let a word be introduced from a foreign language—let it have some resemblance in sound to a real English one: lastly, let the meanings of the two words be not absolutely incompatible. We may then have a word of foreign origin taking the appearance of an English one. Such, amongst others, are beef-eater, from bœuffetier; sparrow-grass, asparagus; Shotover, Chateau vert;^[24] Jerusalem, Girasole;^[25] Spanish {108}beefeater, Spina befida; periwig, peruke; runagate, renegade; lutestring, lustrino;^[26] O yes, Oyez! ancient, ensign.^[27]

Dog-cheap.—This has nothing to do with dogs. The first syllable is god=good transposed, and the second the ch-p in chapman (=merchant) cheap, and East-cheap. In Sir J. Mandeville, we find god-kepe=good bargain.

Sky-larking.—Nothing to do with larks of any sort; still less the particular species, alauda arvensis. The word improperly spelt l-a-r-k, and banished to the slang regions of the English language, is the Anglo-Saxon lác=game, or sport; wherein the a is sounded as in father (not as in farther). Lek=game, in the present Scandinavian languages.

Zachary Macaulay=Zumalacarregui; Billy Ruffian=Bellerophon; Sir Roger Dowlass=Surajah Dowlah, although so limited to the common soldiers, and sailors who first used them, as to be exploded vulgarisms rather than integral parts of the language, are examples of the same tendency towards the irregular accommodation of misunderstood foreign terms.

Birdbolt.—An incorrect name for the gadus lota, or eel-pout, and a transformation of barbote.

Wormwood=weremuth, is an instance of a word from the same language, in an antiquated shape, being equally transformed with a word of really foreign origin.

§ 167. Sometimes the transformation of the name has engendered a change in the object to which it applies, or, at least, has evolved new ideas in connection with it. How easy for a person who used the words beef-eater, sparrow-grass, or Jerusalem, to believe that the officers designated by the former either eat or used to eat more beef than other people (or at least had an allowance of that viand); that the second word was the name for a grass, or herb of which sparrows were fond; and that Jerusalem artichokes came from Palestine.

What has just been supposed is sometimes a real {109}occurrence. To account for the name Shotover-hill, I have heard that Little John shot over it. Here the confusion in order to set itself right, breeds a fiction. Again, in chess, the piece now called the queen, was originally the elephant. This was in Persian, ferz. In French it became vierge, which, in time, came to be mistaken for a derivative, and virgo=the virgin, the lady, the queen.

§ 168. Sometimes, where the form of a word in respect to its sound is not affected, a false spirit of accommodation introduces an unetymological spelling; as frontispiece^[28] from frontispecium, sovereign, from sovrano, colleague from collega, lanthorn (old orthography) from lanterna.

The value of forms like these consists in their showing that language is affected by false etymologies as well as by true ones.

§ 169. In lambkin and lancet, the final syllables (-kin and -et) have the same power. They both express the idea of smallness or diminutiveness. These words are but two out of a multitude, the one (lamb) being of Saxon, the other (lance) of Norman origin. The same is the case with the superadded syllables: -kin is Saxon; -et Norman. Now to add a Saxon termination to a Norman word, or vice versâ, is to corrupt the English language.

§ 170. Introduction of new words—Hybridism.—Hybridism is a term derived from hybrid-a, a mongrel; a Latin word of Greek extraction.

The terminations -ize (as in criticize), -ism (as in criticism), -ic (as in comic), these, amongst many others, are Greek terminations. To add them to words of other than of Greek origin is to be guilty of hybridism.

The terminations -ble (as in penetrable), -bility (as in penetrability, -al (as in parental)—these, amongst many others, are Latin terminations. To add them to words of other than of Latin origin is to be guilty of hybridism.

Hybridism is the commonest fault that accompanies the introduction of new words. The hybrid additions to the English language are most numerous in works on science.

It must not, however, be concealed that several well established words are hybrid; and that, even in the writings of the classical Roman authors, there is hybridism between the Latin and the Greek.

The etymological view of every word of foreign origin is, not that it is put together in England, but that it is brought whole from the language to which it is vernacular. Now no derived word can be brought whole from a language unless, in that language, all its parts exist. The word penetrability is not derived from the English word penetrable, by the addition of -ty. It is the Latin word penetrabilitas imported.

In derived words all the parts must belong to one and the same language, or, changing the expression, every derived word must have a possible form in the language from which it is taken. Such is the rule against Hybridism.

§ 171. A true word sometimes takes the appearance of a hybrid without really being so. The -icle, in icicle, is apparently the same as the -icle in radicle. Now, as ice is Gothic, and -icle classical, hybridism is simulated. Icicle, however, is not a derivative but a compound; its parts being is and gicel, both Anglo-Saxon words.

§ 172. On Incompletion of the Radical.—Let there be in a given language a series of roots ending in -t, as sæmat. Let a euphonic influence eject the -t, as often as the word occurs in the nominative case. Let the nominative case be erroneously considered to represent the root, or radical, of the word. Let a derivative word be formed accordingly, i.e., on the notion that the nominative form and the radical form coincide. Such a derivative will exhibit only a part of the root; in other words, the radical will be incomplete.

Now all this is what actually takes place in words like hæmo-ptysis (spitting of blood), sema-phore (a sort of telegraph). The Greek imparisyllabics eject a part of the root in the nominative case; the radical forms being hæmat- and sæmat-, not hæm- and sæm-. {111}

Incompletion of the radical is one of the commonest causes of words being coined faultily. It must not, however, be concealed, that even in the classical writers, we have (in words like δίστομος) examples of incompletion of the radical.

§ 173. The preceding chapters have paved the way for a distinction between the historical analysis of a language, and the logical analysis of one.

Let the present language of England (for illustration's sake only) consist of 40,000 words. Of these let 30,000 be Anglo-Saxon, 5,000 Anglo-Norman, 100 Celtic, 10 Latin of the first, 20 Latin of the second, and 30 Latin of the third period, 50 Scandinavian, and the rest miscellaneous. In this case the language is considered according to the historical origin of the words that compose it, and the analysis (or, if the process be reversed, the synthesis) is an historical analysis.

But it is very evident that the English, or any other language, is capable of being contemplated in another view, and that the same number of words may be very differently classified. Instead of arranging them according to the languages whence they are derived, let them be disposed according to the meanings that they convey. Let it be said, for instance, that out of 40,000 words, 10,000 are the names of natural objects, that 1000 denote abstract ideas, that 1000 relate to warfare, 1000 to church matters, 500 to points of chivalry, 1000 to agriculture, and so on through the whole. In this case the analysis (or, if the process be reversed, the synthesis) is not historical but logical; the words being classed not according to their origin, but according to their meaning.

Now the logical and historical analysis of a language generally in some degree coincides, as may be seen by noticing the kind of words introduced from the Anglo-Norman, the Latin of the fourth period, and the Arabic.

CHAPTER II.

THE RELATION OF THE ENGLISH TO THE ANGLO-SAXON, AND THE STAGES OF THE ENGLISH LANGUAGE.

§ 174. The relation of the present English to the Anglo-Saxon is that of a modern language to an ancient one: the words modern and ancient being used in a defined and technical sense.

Let the word smiðum illustrate this. Smiðum, the dative plural of smið, is equivalent in meaning to the English to smiths, or to the Latin fabris. Smiðum however, is a single Anglo-Saxon word (a substantive, and nothing more); whilst its English equivalent is two words i.e., a substantive with the addition of a preposition). The letter s, in smiths shows that the word is plural. The -um, in smiðum, does this and something more. It is the sign of the dative case plural. The -um in smiðum, is the part of a word. The preposition to is a separate word with an independent existence. Smiðum is the radical syllable smið, plus the subordinate inflectional syllable -um, the sign of the dative case. To smiths is the substantive smiths, plus the preposition to, equivalent in power to the sign of a dative case, but different from it in form. As far, then, as the word just quoted is concerned, the Anglo-Saxon differs from the English thus. It expresses a given idea by a modification of the form of the root, whereas the modern English denotes the same idea by the addition of a preposition. The Saxon inflection is superseded by a combination of words.

The part that is played by the preposition with nouns, is played by the auxiliaries (have, be, &c.) with verbs.

The sentences in italics are mere variations of the same general statement. (1.) The earlier the stage of a given {113}language the greater the amount of its inflectional forms, and the later the stage of a given language, the smaller the amount of them. (2.) As languages become modern they substitute prepositions and auxiliary verbs for cases and tenses. (3.) The amount of inflection is in the inverse proportion to the amount of prepositions and auxiliary verbs. (4.) In the course of time languages drop their inflection and substitute in its stead circumlocutions by means of prepositions, &c. The reverse never takes place. (5.) Given two modes of expression, the one inflectional (smiðum), the other circumlocutional (to smiths), we can state that the first belongs to an early, the second to a late, stage of language.

The present chapter, then, showing the relation of the English to the Anglo-Saxon, shows something more. It exhibits the general relation of a modern to an ancient language. As the English is to the Anglo-Saxon, so are the Danish, Swedish, and Norwegian, to the old Norse; so also the Modern High German to the Mœso-Gothic; so the Modern Dutch of Holland to the Old Frisian; so, moreover, amongst the languages of a different stock, are the French, Italian, Spanish, Portuguese, Romanese and Wallachian to the Latin, and the Romaic to the Ancient Greek.

§ 175. Contrasted with the English, but contrasted with it only in those points where the ancient tongue is compared with the modern one, the Anglo-Saxon has the following differences.

Of Gender.—In Anglo-Saxon there are three genders, the masculine, the feminine, and the neuter. With adjectives each gender has its peculiar declension; with substantives there are also appropriate terminations, but only to a certain degree; e.g., of words ending in -a (nama, a name; cuma, a guest), it may be stated that they are always masculine; of words in -u (sunu, a son; gifu, a gift), that they are never neuter; in other words, that they are either mas. or fem.

The definite article varies with the gender of its substantive; þæt eage, the eye; se steorra, the star; seo tunge, the tongue. {114}

Of Number.—The plural form in -en (as in oxen), rare in English, was common in Anglo-Saxon. It was the regular termination of a whole declension; e.g., eágan, eyes; steorran, stars; tungan, tongues. Besides this, the Anglo-Saxons had forms in -u and -a, as ricu, kingdoms; gifa, gifts. The termination -s, current in the present English was confined to a single gender and to a single declension, as endas, ends; dagas, days; smiðas, smiths.

Of Case.—Of these the Saxons had, for their substantives, at least three; viz. the nominative, dative, genitive. With the pronouns and adjectives there was a true accusative form; and with a few especial words an ablative or instrumental one. Smið, a smith; smiðe, to a smith; smiðes, of a smith. Plural, smiðas, smiths; smiðum, to smiths; smiða, of smiths: he, he; hine, him; him, to him; his, his; se, the; þa, the; þy, with the; þam, to the; þæs, of the.

Of the dative in -um, the word whilom (at times, at whiles) is a still extant and an almost isolated specimen.

Of Declension.—In Anglo-Saxon it is necessary to determine the termination of a substantive. There is the weak, or simple declension for words ending in a vowel (as eage, steorra, tunga), and the strong, or complex declension for words ending in a consonant (smið, spræc, leáf). The letters i and u are dealt with as semivowels, semivowels being dealt with as consonants; so that words like sunu and gifu belong to the same declension as smið and sprǽc.

That the form of adjectives varies with their definitude or indefinitude, has been seen from § 93: definite adjectives following the inflection of the simple; indefinite ones that of the complex declension.

The Anglo-Saxon inflection of the participles present is remarkable. With the exception of the form for the genitive plural definite (which, instead of -ena, is -ra,) they follow the declension of the adjectives. From the masculine substantives formed from them, and denoting the agent, they may be distinguished by a difference of inflection. {115}

Pronouns Personal.—Of the pronominal inflection in Saxon, the character may be gathered from the chapter upon pronouns. At present, it may be stated that, like the Mœso-Gothic and the Icelandic, the Anglo-Saxon language possessed for the first two persons a dual number; inflected as follows:

Besides this, the demonstrative, possessive, and relative pronouns, as well as the numerals twa and þreo, had a fuller declension than they have at present.

Mood.—The subjunctive mood that in the present English (with the exception of the conjugation of the verb substantive) differs from the indicative only in the third person singular, was in Anglo-Saxon inflected as follows:

The Saxon infinitive ended in -an (lufian), and besides this there was a so-called gerundial form, to lufigenne. {116}

Tense.—In regard to tense, the Anglo-Saxon coincided with the English. The present language has two tenses, the present and the past; the Saxon had no more. This past tense the modern English forms either by addition (love, loved), or by change (fall, fell). So did the Anglo-Saxons.

Number and Person.—In the present English the termination -eth (moveth) is antiquated. In Anglo-Saxon it was the only form recognized. In English the plural number (indicative as well as subjunctive) has no distinguishing inflection. It was not so in Anglo-Saxon. There, although the persons were identical in form, the numbers were distinguished by the termination -að for the indicative, and -n for the subjunctive. (See above.) For certain forms in the second conjugation, see the remarks on the forms drunk and drank, in Part IV.

Such are the chief points in the declension of nouns and the conjugation of verbs that give a difference of character between the ancient Anglo-Saxon and the modern English: and it has already been stated that the difference between the New and the Old German, the Dutch and the Frisian, the Italian, &c., and the Latin, the Romaic and the Greek, &c., are precisely similar.

How far two languages pass with equal rapidity from their ancient to their modern, from their inflected to their uninflected state (in other words, how far all languages alter at the same rate), is a question that will be noticed elsewhere. At present, it is sufficient to say, that (just as we should expect à priori) languages do not alter at the same rate.

Akin to the last question is a second one: viz.: how far the rate of change in a given language can be accelerated by external circumstances. This second question bears immediately upon the history of the English language. The grammar of the current idiom compared with the grammar of the Anglo-Saxon is simplified. How far was this simplification of the grammar promoted by the Norman Conquest. The current views exaggerate the influence of the Norman Conquest and of French connexions. The remark of Mr. Price in his Preface to Warton, acceded to by Mr. Hallam in his Introduction to the Literature of Europe, is, that every one of the {117}other Low Germanic languages (affected by nothing corresponding to the Norman Conquest) displays the same simplification of grammar as the Anglo-Saxon (affected by the Norman Conquest) displays. Confirmatory of this remark, it may be added, that compared with the Icelandic, the Danish and Swedish do the same. Derogatory to it is the comparatively complex grammar of the new German, compared, not only with the Old High German, but with the Mœso-Gothic. An extract from Mr. Hallam shall close the present section and introduce the next.

§ 176. At a given period, then, the Anglo-Saxon of the standard, and (if the expression may be used) classical authors, such as Cædmon, Alfred, Ælfric, &c., had undergone such a change as to induce the scholars of the present age to denominate it, not Saxon, but Semi-Saxon. It had ceased to be genuine Saxon, but had not yet become English. In certain parts of the kingdom, where the mode of speech {118}changed more rapidly than elsewhere, the Semi-Saxon stage of our language came earlier. It was, as it were, precipitated.

The History of King Leir and his Daughters is found in two forms. Between these there is a difference either of dialect or of date, and possibly of both. Each, however, is Semi-Saxon. The extracts are made from Thorpe's Analecta Anglo-Saxonica, p. 143.

The Grave, a poetical fragment, the latter part of the Saxon Chronicle, a Homily for St. Edmund's Day (given in the Analecta), and above all the printed extracts of the poem of Layamon, are the more accessible specimens of the Semi-Saxon. The Ormulum, although in many points English rather than Saxon, retains the dual number of the Anglo-Saxon pronouns. However, lest too much stress be laid upon this circumstance, the epistolary character of the Ormulum must be borne in mind.

It is very evident that if, even in the present day, there were spoken in some remote district the language of Alfred and Ælfric, such a mode of speech would be called, not Modern English, but Anglo-Saxon. This teaches us that the stage of language is to be measured, not by its date, but by its structure. Hence, Saxon ends and Semi-Saxon begins, not at a given year, A.D., but at that time {119}(whenever it be) when certain grammatical inflections disappear, and certain characters of a more advanced stage are introduced.

1. The substitution of -an for -as, in the plural of substantives, munucan for munucas (monks); and, conversely, the substitution of -s for -n, as steorres for steorran (stars). The use of -s, as the sign of the plural, without respect to gender, or declension, may be one of those changes that the Norman Conquest forwarded; -s being the sign of the plural in Anglo-Norman.

2. The ejection or shortening of final vowels, þæt ylc for þæt ylce; sone for sunu; name for nama; dages for dagas.

4. The ejection of the -n of the infinitive mood, cumme for cuman (to come), nemne for nemnen (to name).

5. The ejection of -en in the participle passive, I-hote for gehaten (called, hight).

6. The gerundial termination -enne, superseded by the infinitive termination -en; as to lufian for to lufienne, or lufigenne.

7. The substitution of -en for -að in the persons plural of verbs; hi clepen (they call) for hi clypiað, &c.

The preponderance (not the occasional occurrence) of forms like those above constitute Semi-Saxon in contradistinction to standard Saxon, classical Saxon, or Anglo-Saxon proper.

§ 177. Old English Stage.—Further changes convert Semi-Saxon into Old English. Some, amongst others, are the following:—

1. The ejection of the dative plural termination -um, and the substitution of the preposition to and the plural sign -s; as to smiths for smiðum. Of the dative singular the -e is retained (ende, worde); but it is by no means certain that, although recognized in writing, it was recognized in pronunciation also.

2. The ejection of -es in the genitive singular whenever the {120}preposition of came before it; Godes love (God's love), but the love of God, and not the love of Godes.

3. The syllable -es as a sign of the genitive case extended to all genders and to all declensions; heart's for heortan; sun's for sunnan.

4. The same in respect to the plural number; sterres for steorran; sons for suna.

The preponderance of the forms above (and not their occasional occurrence) constitutes old English in contradistinction to Semi-Saxon.

The following extract from Henry's history (vol. viii. append. iv.) is the proclamation of Henry III. to the people of Huntingdonshire, A.D. 1258. It currently passes for the earliest specimen of English.

§ 178. The songs amongst the political verses printed by the Camden Society, the romance of Havelok the Dane, {121}William and the Werwolf, the Gestes of Alisaundre, King Horn, Ipomedon, and the King of Tars; and, amongst the longer works, Robert of Gloucester's Chronicle, and the poems of Robert of Bourn (Brunn), are (amongst others) Old English. Broadly speaking, the Old English may be said to begin with the reign of Henry III., and to end with that of Edward III.

1. A fuller inflection of the demonstrative pronoun, or definite article; þan, þenne, þære, þam;—in contradistinction to the Middle English.

3. The existence of a genitive plural in -r or -ra; heora, theirs; aller, of all;—ditto. This with substantives and adjectives is less common.

4. The substitution of heo for they, of heora for their, of hem for them;—in contradistinction to the later stages of English, and in contradistinction to old Lowland Scotch. (See Chapter III.)

5. A more frequent use of min and thin, for my and thy;—in contradistinction to middle and modern English.

6. The use of heo for she;—in contradistinction to middle and modern English and old Lowland Scotch.

7. The use of broader vowels; as in iclepud or iclepod (for icleped or yclept); geongost, youngest; ascode, asked; eldore, elder.

8. The use of the strong preterits (see the chapter on the tenses of verbs), where in the present English the weak form is found; wex, wop, dalf, for waxed, wept, delved.

9. The omission not only of the gerundial termination -enne, but also of the infinitive sign -en after to; to honte, to speke;—in contradistinction to Semi-Saxon.

10. The substitution of -en for -eþ or -eð in the first and second persons plural of verbs; we wollen, we will: heo schullen, they should;—ditto.

12. The substitution of ben and beeth, for synd and syndon=we, ye, they are;—in contradistinction to Semi-Saxon.

§ 179. The degree to which the Anglo-Saxon was actually influenced by the Anglo-Norman has been noticed. The degree wherein the two languages came in contact is, plainly, another consideration. The first is the question, How far one of two languages influenced the other? The second asks, How far one of two languages had the opportunity of influencing the other? Concerning the extent to which the Anglo-Norman was used, I retail the following statements and quotations.

That there was French in England before the battle of Hastings appears on the authority of Camden:—

§ 180. In Chaucer and Mandeville, and perhaps in all the writers of the reign of Edward III., we have a transition {123}from the Old to the Middle English. The last characteristic of a grammar different from that of the present English, is the plural form in -en; we tellen, ye tellen, they tellen. As this disappears, which it does in the reign of Queen Elizabeth (Spenser has it continually), the Middle English may be said to pass into the New or Modern English.

§ 181. The present tendencies of the English may be determined by observation; and as most of them will be noticed in the etymological part of this volume, the few here indicated must be looked upon as illustrations only.

1. The distinction between the subjunctive and indicative mood is likely to pass away. We verify this by the very general tendency to say if it is, and if he speaks, for if it be, and if he speak.

2. The distinction (as far as it goes) between the participle passive and the past tense is likely to pass away. We verify this by the tendency to say it is broke, and he is smote, for it is broken, and he is smitten.

3. Of the double forms, sung and sang, drank and drunk, &c. one only will be the permanent.

As stated above, these tendencies are a few out of a number, and have been adduced in order to indicate the subject rather than to exhaust it.

§ 182. What the present language of England would have been had the Norman Conquest never taken place, the analogy of Holland, Denmark, and of many other countries enables us to determine. It would have been much as it is at present. What it would have been had the Saxon conquest never taken place, is a question wherein there is far more speculation. Of France, of Italy, of Wallachia, and of the Spanish Peninsula, the analogies all point the same way. They indicate that the original Celtic would have been superseded by the Latin of the conquerors, and consequently that our language in its later stages would have been neither British nor Gaelic, but Roman. Upon these analogies, however, we may refine. Italy, was from the beginning, Roman; the Spanish Peninsula was invaded full early; no ocean divided Gaul from Rome; and the war against the ancestors of the Wallachians was a war of extermination.

CHAPTER III.

§ 183. The term Lowland is used to distinguish the Scotch of the South-east from the Scotch of the Highlands. The former is English in its immediate affinities, and Germanic in origin; the latter is nearly the same language with the Gaelic of Ireland, and is, consequently, Celtic.

The question as to whether the Lowland Scotch is a dialect of the English, or a separate and independent language, is a verbal rather than a real one.

Reasons for considering the Scotch and English as dialects of one and the same language lie in the fact of their being (except in the case of the more extreme forms of each) mutually intelligible.

Reasons for calling one a dialect of the other depend upon causes other than philological, e.g., political preponderance, literary development, and the like.

Reasons for treating the Scotch as a separate substantive language lie in the extent to which it has the qualities of a regular cultivated tongue, and a separate substantive literature—partially separate and substantive at the present time, wholly separate and substantive in the times anterior to the union of the crowns, and in the hands of Wyntoun, Blind Harry, Dunbar, and Lindsay.

§ 184. Reasons for making the philological distinction between the English and Scotch dialects exactly coincide with the geographical and political boundaries between the two kingdoms are not so easily given. It is not likely that the Tweed and Solway should divide modes of speech so accurately as they divide laws and customs; that broad and trenchant lines of demarcation should separate the Scotch {125}from the English exactly along the line of the Border; and that there should be no Scotch elements in Northumberland, and no Northumbrian ones in Scotland. Neither is such the case. Hence, in speaking of the Lowland Scotch, it means the language in its typical rather than in its transitional forms; indeed, it means the literary Lowland Scotch which, under the first five Stuarts, was as truly an independent language as compared with the English, as Swedish is to Danish, Portuguese to Spanish, or vice versâ.

§ 185. This limitation leaves us fully sufficient room for the notice of the question as to its origin; a notice all the more necessary from the fact of its having created controversy.

What is the primâ facie view of the relations between the English of England, and the mutually intelligible language (Scotch or English, as we choose to call it) of Scotland? One of three:—

1. That it originated in England, and spread in the way of extension and diffusion northwards, and so reached Scotland.

2. That it originated in Scotland, and spread in the way of extension and diffusion southwards, and so reached England.

In any of these cases it is Angle, or Saxon, or Anglo-Saxon, even as English is Angle, or Saxon, or Anglo-Saxon.

§ 186. A view, however, different from these, and one disconnecting the Lowland Scotch from the English and Anglo-Saxon equally, is what may be called the Pict doctrine. Herein it is maintained that the Lowland Scotch is derived from the Pict, and that the Picts were of Gothic origin. The reasoning upon these matters is to be found in the Dissertation upon the Origin of the Scottish Language prefixed to Jamieson's Etymological Dictionary: two extracts from which explain the view which the author undertakes to combat:—

a. "It is an opinion which, after many others, has been pretty generally received, and, perhaps, almost taken for granted, that the language spoken in the Lowlands of {126}Scotland is merely a corrupt dialect of the English, or at least of the Anglo-Saxon."

b. "It has generally been supposed that the Saxon language was introduced into Scotland in the reign of Malcolm Canmore by his good queen and her retinue; or partly by means of the intercourse which prevailed between the inhabitants of Scotland and those of Cumberland, Northumberland, Westmoreland, and Durham, which were held by the Kings of Scotland as fiefs of the crown of England. An English writer, not less distinguished for his amiable disposition and candour than for the cultivation of his mind, has objected to this hypothesis with great force of argument."

§ 187. Now, as against any such notion as that involved in the preceding extracts, the reasoning of the learned author of the Scottish Dictionary may, perhaps, be valid. No such view, however, is held, at the present moment, by any competent judge; and it is doubtful whether, in the extreme way in which it is put forward by the opponent of it, it was ever maintained at all.

Be this, however, as it may, the theory which is opposed to it rests upon the following positions—

From whence it follows that—assuming what is true concerning the Orkneys is true concerning the Lowland Scotch—the Lowland Scotch was Pict, Scandinavian, Gothic, and (as such) more or less Belgic.

For the non-Gothic character of the Picts see the researches of Mr. Garnett, as given in § 139, as well as a paper—believed to be from the same author—in the Quarterly Review for 1834. {127}

§ 188. That what is true concerning the Orkneys (viz. that they were Scandinavian) is not true for the south and eastern parts of Scotland, is to be collected from the peculiar distribution of the Scottish Gaelic; which indicates a distinction between the Scandinavian of the north of Scotland and the Scandinavian of the east of England. The Lowland Scotch recedes as we go northward. Notwithstanding this, it is not the extreme north that is most Gaelic. In Caithness the geographical names are Norse. Sutherland, the most northern county of Scotland, takes its name from being south; that is, of Norway. The Orkneys and Shetland are in name, manners, and language, Norse or Scandinavian. The Hebrides are Gaelic mixed with Scandinavian. The Isle of Man is the same. The word Sodor (in Sodor and Man) is Norse, with the same meaning as it has in Sutherland. All this indicates a more preponderating, and an earlier infusion of Norse along the coast of Scotland, than that which took place under the Danes upon the coasts of England, in the days of Alfred and under the reign of Canute. The first may, moreover, have this additional peculiarity, viz. of being Norwegian rather than Danish. Hence I infer that the Scandinavians settled in the northern parts of Scotland at an early period, but that it was a late period when they ravaged the southern ones; so that, though the language of Orkney may be Norse, that of the Lothians may be Saxon.

To verify these views we want not a general dictionary of the Scottish language taken altogether, but a series of local glossaries, or at any rate a vocabulary, 1st, of the northern; 2ndly, of the southern Scottish.

Between the English and Lowland Scotch we must account for the likeness as well as the difference. The Scandinavian theory accounts for the difference only.

§ 189. Of the following specimens of the Lowland Scotch, the first is from The Bruce, a poem written by Barbour, Archdeacon of Aberdeen, between the years 1360 and 1375; the second from Wyntoun; the third from Blind Harry's poem, Wallace, 1460; and the fourth from Gawin Douglas's translation of the Æneid, A.D. 1513. {128}

§ 190. In the way of orthography, the most characteristic difference between the English and Scotch is the use, on the part of the latter, of qu for wh; as quhen, quhare, quhat, for when, where, what. The substitution of sch for sh (as scho for she), and of z for the Old English ȝ (as zour for ȝeowr, your), is as much northern English as Scotch. {131}

In pronunciation, the substitution of d for ð (if not a point of spelling), as in fader for father; of a for o, as báith for both; of s for sh, as sall for shall; and the use of the guttural sound of ch, as in loch, nocht, are the same.

The ejection of the n before t, or an allied sound, and the lengthening of the preceding vowel, by way of compensation, as in begouth for beginneth, seems truly Scotch. It is the same change that in Greek turns the radical syllable ὀδοντ into ὀδούς.

The formation of the plural of verbs in -s, rather than in -th (the Anglo-Saxon form), is Northern English as well as Scotch:—Scotch, slepys, lovys; Northern English, slepis, lovis; Old English, slepen, loven; Anglo-Saxon slepiað, lufiað.

The formation of the plural number of the genitive case by the addition of the syllable -is (blastis, birdis, bloomis), instead of the letter -s (blasts, birds, blooms), carries with it a metrical advantage, inasmuch as it gives a greater number of double rhymes.

The same may be said of the participial forms, affrayit, assurit, for affrayd, assured.

Concerning the comparative rate of change in the two languages no general assertion can be made. In the Scotch words sterand, slepand, &c., for steering, sleeping, the form is antiquated, and Anglo-Saxon rather than English. It is not so, however, with the words thai (they), thaim (them), thair (their), compared with the contemporary words in English, heo, hem, heora. In these it is the Scottish that is least, and the English that is most Anglo-Saxon.

CHAPTER IV.

§ 191. The languages mentioned in the present chapter claim their place on one ground only,—they have been the subject of controversy. The notice of them will be brief. The current texts upon which the controversies have turned will be quoted; whilst the opinion of the present writer is left to be collected from the title of the chapter.

The Belgæ.—By some these are considered a Germanic rather than a Celtic tribe; the view being supported by the following extracts from Cæsar:—"Gallia est omnis divisa in tres partes; quarum unam incolunt Belgæ, aliam Aquitani, tertiam, qui ipsorum lingua Celtæ, nostra Galli, appellantur. Hi omnes lingua, institutis, legibus inter se differunt. Gallos—a Belgis Matrona et Sequana dividit."—B. G. i. "Belgæ ab extremis Galliæ finibus oriuntur."—B. G. ii. "Quum ab his quæreret, quæ civitates, quantæque in armis essent, et quid in bello possent, sic reperiebat: plerosque Belgas esse ortos a Germanis, Rhenumque antiquitùs transductos, propter loci fertilitatem ibi consedisse; Gallosque, qui ea loca incolerent, expulisse; solosque esse qui patrum nostrorum memoria, omni Gallia vexata Teutones Cimbrosque intra fines suos ingredi prohibuerunt."—B. G. ii. 4. "Britanniæ pars interior ab iis incolitur quos natos in insulâ ipsâ memoriâ proditum dicunt: maritima pars ab iis, qui prædæ ac belli inferendi causa ex Belgio transierant."—B. G. v. 12.

§ 192. The possibly Germanic origin of the Belgæ, and the Belgic element of the British population, are matters which bear upon the question indicated in § 10, or that of the Germanic influences anterior to A.D. 449. {133}

They have a still more important bearing, the historian over and above identifying the Belgæ with the Germans, affirms that what applies to the Belgæ applies to the Picts also.

Now this is one of the arguments in favour of the doctrine exhibited (and objected to) in pp. 124-127, and the extent of questions upon which it bears, may be collected from the following quotation:—"A variety of other considerations might be mentioned, which, although they do not singly amount to proof, yet merit attention, as viewed in connexion with what has been already stated.

"As so great a part of the eastern coast of what is now called England was so early peopled by the Belgæ, it is hardly conceivable that neither so enterprising a people, nor any of their kindred tribes, should ever think of extending their descents a little farther eastward. For that the Belgæ and the inhabitants of the countries bordering on the Baltic, had a common origin, there seems to be little reason to doubt. The Dutch assert that their progenitors were Scandinavians, who, about a century before the common era, left Jutland and the neighbouring territories, in quest of new habitations.^[29] The Saxons must be viewed as a branch from the same stock; for they also proceeded from modern Jutland and its vicinity. Now, there is nothing repugnant to reason in supposing that some of these tribes should pass over directly to the coast of Scotland opposite to them, even before the Christian era. For Mr. Whitaker admits that the Saxons, whom he strangely makes a Gaulic people, in the second century applied themselves to navigation, and soon became formidable to the Romans.^[30] Before they could become formidable to so powerful a people, they must have been at least so well acquainted with navigation as to account it no great enterprise to cross from the shores of the Baltic over to Scotland, especially if they took the islands of Shetland and Orkney in their way.

"As we have seen that, according to Ptolemy, there were, in his time, different tribes of Belgæ, settled on the northern {134}extremity of our country: the most natural idea undoubtedly is, that they came directly from the Continent. For had these Belgæ crossed the English Channel, according to the common progress of barbarous nations, it is scarcely supposable that this island would have been settled to its utmost extremity so early as the age of Agricola.

"There is every reason to believe, that the Belgic tribes in Caledonia, described by Ptolemy, were Picts. For as the Belgæ, Picts, and Saxons seem to have had a common origin, it is not worth while to differ about names. These frequently arise from causes so trivial, that their origin becomes totally inscrutable to succeeding ages. The Angles, although only one tribe, have accidentally given their name to the country which they invaded, and to all the descendants of the Saxons and Belgæ, who were by far more numerous.

"It is universally admitted, that there is a certain national character, of an external kind, which distinguishes one people from another. This is often so strong that those who have travelled through various countries, or have accurately marked the diversities of this character, will scarcely be deceived even as to a straggling individual. Tacitus long ago remarked the striking resemblance between the Germans and Caledonians. Every stranger, at this day, observes the great difference and complexion between the Highlanders and Lowlanders. No intelligent person in England is in danger of confounding the Welsh with the posterity of the Saxons. Now, if the Lowland Scots be not a Gothic race, but in fact the descendants of the ancient British, they must be supposed to retain some national resemblance of the Welsh. But will any impartial observer venture to assert, that in feature, complexion, or form, there is any such similarity as to induce the slightest apprehension that they have been originally the same people?"^[31]

It is doubtful, however, whether Cæsar meant to say more than that over above certain differences which distinguished the Belgæ from the other inhabitants of the common country Gallia, there was an intermixture of Germans.

The import of a possibly Germanic origin for the Belgæ gives us the import of a possibly Germanic origin for—

§ 193. The Caledonians.—A speculative sentence of Tacitus indicates the chance of the Caledonians being Germanic:—"Britanniam qui mortales initio coluerint, indigenæ an advecti, ut inter barbaros, parum compertum. Habitus corporum varii: atque ex eo argumenta: namque rutilæ Caledoniam habitantium comæ, magni artus, Germanicam originem adseverant."—Agricola, xi.

The continuation of the passage quoted in § 193 has induced the notion that there have been in Britain Spanish, Iberic, or Basque tribes:—"Silurum colorati vultus, et torti plerumque crines, et posita contra Hispania, Iberos veteres trajecisse, easque sedes occupâsse fidem faciunt."—Agricola, xi.

As this, although an opinion connected with the history of the languages of Great Britain, is not an opinion connected with the history of the English language, it is a question for the Celtic, rather than the Gothic, philologist. The same applies to the points noticed in §§ 136-138. Nevertheless they are necessary for the purposes of minute philological analysis.

§ 194. As early as the year A.D. 1676, an opinion was advanced by^[32] Aylett Sammes, in a work entitled Britannia Antiqua Illustrata, that the first colonisers of Ireland were the merchants of Tyre and Sidon. In confirmation of this opinion the existence of several Eastern customs in Ireland was adduced by subsequent antiquarians. Further marks of an Eastern origin of the Irish were soon found in the Gaelic dialect of that country. Finally, the matter (in the eyes at least of the national writers) was satisfactorily settled by the famous discovery, attributed to General Vallancey, of the true meaning of the Carthaginian lines in Plautus.

In the Little Carthaginian (Pœnulus) of the Latin comic writer Plautus, a portion of the dialogue is carried on in the language of Carthage.

That the Punic language of Carthage should closely {136}resemble that of the mother-city Tyre, which was Phœnician; and that the Phœnician of Tyre should be allied to the language of Palestine and Syria, was soon remarked by the classical commentators of the time. Joseph Scaliger asserted that the Punic of the Pœnulus differed but little from pure Hebrew—"Ab Hebraismi puritate parum abesse."

Emendated and interpreted by Bochart, the first ten lines of a speech in Act v. s. 1. stand thus:—

Six lines following these were determined to be Liby-Phœnician, or the language of the native Africans in the neighbourhood of Carthage, mixed with Punic. These, it was stated, had the same meaning with the ten lines in Carthaginian.

The following lines of Plautus have, by all commentators, {137}been viewed in the same light, viz. as the Latin version of the speech of the Carthaginian.

Guided by the metrical paraphrase of the original author, Bochart laid before the scholars of his time a Latin version, of which the following is an English translation:—

To professed classics and to professed orientalists, the version of Bochart has, on the whole, appeared satisfactory. Divisions of opinion there have been, it is true, even amongst those who received it; but merely upon matters of detail. Some have held that the Punic is Syriac rather than Hebraic, whilst others have called in to its interpretation the Arabic, {138}the Maltese, or the Chaldee; all (be it observed) languages akin to the Hebrew. Those who look further than this for their affinities, Gesenius^[33] dismisses in the following cavalier and cursory manner:—"Ne eorum somnia memorem, qui e Vasconum et Hiberniæ linguis huic causæ succurri posse opinati sunt; de quibus copiosius referre piget."

The remark of Gesenius concerning the pretended affinities between the Punic and Hibernian arose from the discovery attributed to General Vallancey; viz. that the speech in Plautus was Irish Gaelic, and consequently that the Irish was Carthaginian, and vice versâ. The word attributed is used because the true originator of the hypothesis was not Vallancey, but O'Neachtan.

From the quotations already given, the general reader may see that both the text and the translation of Plautus are least violated in the reading and rendering of Bochart, a reading and rendering which no Gothic or Semitic scholar has ever set aside.

§ 195. The hypothesis of an aboriginal Finnic population in Britain and elsewhere.—A Celtic population of Britain preceded the Germanic. Are there any reasons for believing that any older population preceded the Celtic?

The reasoning upon this point is preeminently that of the Scandinavian (i.e. Danish, Swedish, and Norwegian) school of philology and ethnology.

Arndt, I believe, was the first who argued that if the so-called Indo-European nations were as closely connected with each other as they are generally considered, their separation from the common stock must have been subsequent to the occupation of Europe by some portion or other of the human species—in other words, that this earlier population must have been spread over those areas of which the Indo-Europeans took possession only at a later period.

That the divisions of such an earlier population were, at least, as closely connected with each other as the different members of the so-called Indo-European class, was a reasonable opinion. It was even reasonable to suppose that they were more closely connected; since the date of their diffusion must have been nearer the time of the original dispersion of mankind.

If so, all Europe (the British Isles included) might have had as its aborigines a family older than the oldest members of the Indo-European stock; a family of which every member may now be extinct, or a family of which remains may still survive.

1. Parts beyond the limits of the area occupied by the so-called Indo-Europeans.

2. Parts within the limits of the so-called Indo-Europeans; but so fortified by nature as to have been the stronghold of a retiring population.

a. The countries beyond the Indo-Europeans of the Scandinavian and Slavonic areas, i.e. the countries of the Laplanders and Finnlanders.

b. The countries beyond the Indo-Europeans of the Iranian stock, i.e. the Dekkan, or the country of those natives of India (whatever they may be) whose languages are not derived from the Sanscrit.

a. The Basque districts of the Pyrenees, where the language represents that of the aborigines of Spain anterior to the conquest of the Roman.

b. The Albanians.—Such the doctrine of the continuity of an ante-Indo-European population, from Cape Comorin to Lapland, and from Lapland to the Pyrenees. There is some philological evidence of this: whether there is enough is another matter.

This view, which on its philological side has been taken up by Rask, Kayser, and the chief Scandinavian scholars, and which, whether right or wrong, is the idea of a bold and comprehensive mind, as well as a powerful instrument of criticism in the way of a provisional theory, has also been adopted on its physiological side by the chief Scandinavian anatomists and palæontologists—Retzius, Eschricht, Niilson, and others. Skulls differing in shape from the Celtic skulls of Gaul, and from the Gothic skulls of Germany and Scandinavia, have been found in considerable numbers; and generally in burial-places of an apparently greater antiquity than those which contain typical Celtic, or typical Gothic crania. Hence there is some anatomical as well as philological evidence: whether there is enough is another question.

PART III.

CHAPTER I.

§ 196. To two points connected with the subject of the following Chapter, the attention of the reader is requested.

The syllables ka and ga are similar syllables. The vowel is in each the same, and the consonant is but slightly different. Now the words ka and ga are more allied to each other than the words ka and ba, ka and ta, &c., because the consonantal sounds of k and g are more allied than the consonantal sounds of k and b, k and t.

Comparing the syllables ga and ka, we see the affinity between the sounds, and we see it at the first glance. It lies on the surface, and strikes the ear at once.

It is, however, very evident that ways might be devised, or might arise from accident, of concealing the likeness between the two sounds, or, at any rate, of making it less palpable. One of such ways would be a faulty mode of spelling. If instead of ga we wrote gha the following would be the effect: the syllable would appear less simple than it really was; it would look as if it consisted of three parts instead of two, and consequently its affinity to ka would seem less than it really was. It is perfectly true that a little consideration would tell us that, as long as the sound remained the same, the relation {142}of the two syllables remained the same; and that, if the contrary appeared to be the case, the ear was misled by the eye. Still a little consideration would be required. Now in the English language we have, amongst others, the following modes of spelling that have a tendency to mislead:—

The sounds of ph and of f, in Philip and fillip, differ to the eye, but to the ear are identical. Here a difference is simulated.

The sounds of th in thin, and of th in thine, differ to the ear, but to the eye seem the same. Here a difference is concealed.

These last sounds appear to the eye to be double or compound. This is not the case; they are simple single sounds, and not the sounds of t followed by h, as the spelling leads us to imagine.

II. Besides improper modes of spelling, there is another way of concealing the true nature of sounds. If I say that ka and ga are allied, the alliance is manifest; since I compare the actual sounds. If I say ka and gee are allied, the alliance is concealed; since I compare, not the actual sounds, but only the names of the letters that express those sounds. Now in the English language we have, amongst others, the following names of letters that have a tendency to mislead:—

The sounds fa and va are allied. The names eff and vee conceal this alliance.

The sounds sa and za are allied. The names ess and zed conceal the alliance.

In comparing sounds it is advisable to have nothing to do either with letters or names of letters. Compare the sounds themselves.

In many cases it is sufficient, in comparing consonants, to compare syllables that contain those consonants; e.g., to determine the relations of p, b, f, v, we say pa, ba, fa, va; or for those of s and z, we say sa, za. Here we compare syllables, each consonant being followed by a vowel. At times this is insufficient. We are often obliged to isolate the consonant from its vowel, and bring our organs to utter (or half utter) imperfect sounds of p', b', t', d'. In doing this we isolate the consonant. {143}

§ 197. Let any of the vowels (for instance, the a in father) be sounded. The lips, the tongue, and the parts within the throat remain in the same position: and as long as these remain in the same position the sound is that of the vowel under consideration. Let, however, a change take place in the position of the organs of sound; let, for instance, the lips be closed, or the tongue be applied to the front part of the mouth: in that case the vowel sound is cut short. It undergoes a change. It terminates in a sound that is different, according to the state of those organs whereof the position has been changed. If, on the vowel in question, the lips be closed, there then arises an imperfect sound of b or p. If, on the other hand, the tongue be applied to the front teeth, or to the fore part of the palate, the sound is one (more or less imperfect) of t or d. This fact illustrates the difference between the vowels and the consonants. It may be verified by pronouncing the a in fate, ee in feet, oo in book, o in note, &c.

It is a further condition in the formation of a vowel sound, that the passage of the breath be uninterrupted. In the sound of the l' in lo (isolated from its vowel) the sound is as continuous as it is with the a in fate. Between, however, the consonant l and the vowel a there is this difference: with a, the passage of the breath is uninterrupted; with l, the tongue is applied to the palate, breaking or arresting the passage of the breath.

§ 198. The primary division of our articulate sounds is into vowels and consonants. The latter are again divided into liquids (l, m, n, r) and mutes (p, b, f, v, t, d, g, s, z, &c.) Definitions for the different sorts of articulate sounds have still to be laid down. In place of these, we have general assertions concerning the properties and qualities of the respective classes. Concerning the consonants as a class, we may predicate one thing concerning the liquids, and concerning the mutes, another. What the nature of these assertions is, will be seen after the explanation of certain terms.

§ 199. Sharp and flat.—Take the sounds of p, f, t, k, s; isolate them from their vowels, and pronounce them. The sound is the sound of a whisper. {144}

Let b, v, d, g, z, be similarly treated. The sound is no whisper, but one at the natural tone of our voice.

Now p, f, t, k, s (with some others that will be brought forward anon) are sharp, whilst b, v, &c. are flat. Instead of sharp, some say hard, and instead of flat, some say soft. The Sanskrit terms sonant and surd are, in a scientific point of view, the least exceptionable. They have, however, the disadvantage of being pedantic. The tenues of the classics (as far as they go) are sharp, the mediæ flat.

Continuous and explosive.—Isolate the sounds of b, p, t, d, k, g. Pronounce them. You have no power of prolonging the sounds, or of resting upon them. They escape with the breath, and they escape at once.

It is not so with f, v, sh, zh. Here the breath is transmitted by degrees, and the sound can be drawn out and prolonged for an indefinite space of time. Now b, p, t, &c. are explosive f, v, &c. continuous.

§ 200. Concerning the vowels, we may predicate a) that they are all continuous, b) that they are all flat.

Concerning the liquids, we may predicate a) that they are all continuous, b) that they are all flat.

Concerning the mutes, we may predicate a) that one half of them is flat, and the other half sharp, and b) that some are continuous, and that others are explosive.

CHAPTER II.

§ 202.—The attention of the reader is now directed to the following foreign vowel sounds.

1. é fermé, of the French.—This is a sound allied to, but different from, the a in fate, and the ee in feet. It is intermediate to the two.

2. u of the French, ü of the Germans, y of the Danes.—This sound is intermediate to the ee in feet, and the oo in book.

3. o chiuso, of the Italians.—Intermediate to the o in note, and the oo in book.

For these sounds we have the following sequences: a in fate, é fermé, ee in feet, ü in übel (German), oo in book, o chiuso, o in note. And this is the true order of alliance among the vowels; a in fate, and o in note, being the extremes; the other sounds being transitional or intermediate. As the English orthography is at once singular and faulty, it exhibits the relationship but imperfectly.

§ 203. The system of the mutes.—Preliminary to the consideration of the system of the mutes, let it be observed:—

1. that the th in thin is a simple single sound, different from the th in thine, and that it may be expressed by the sign þ.

2. That the th in thine is a simple single sound, different from the th in thin, and that it may be expressed by the sign ð.

3. That the sh in shine is a simple single sound, and that it may be expressed by the sign σ (Greek σῖγμα).

4. That the z in azure, glazier (French j), is a simple single sound, and that it may be expressed by the sign ζ (Greek ζῆτα). {146}

5. That in the Laplandic, and possibly in many other languages, there are two peculiar sounds, different from any in English, German, and French, &c., and that they may respectively be expressed by the sign κ and the sign γ (Greek κάππα and γάμμα).

With these preliminary notices we may exhibit the system of the sixteen mutes; having previously determined the meaning of two fresh terms, and bearing in mind what was said concerning the words sharp and flat, continuous and explosive.

Lene and aspirate.—From the sound of p in pat, the sound of f in fat differs in a certain degree. This difference is not owing to a difference in their sharpness or flatness. Each is sharp. Neither is it owing to a difference in their continuity or explosiveness; although, at the first glance, such might appear to be the case. F is continuous, whilst p is explosive. S, however, is continuous, and s, in respect to the difference under consideration, is classed not with f the continuous sound but with p the explosive one. I am unable to account for the difference between p and f. It exists: it is visible. It has been expressed by a term. P is called lene, f is called aspirate.

Hence p, b, t, d, k, g, s, z, are lene; f, v, þ, ð, κ, γ, σ, ζ, are aspirate. Also p, f, t, þ, k, κ, s, σ, are sharp, whilst b, v, d, ð, g, γ, z, ζ, are flat; so that there is a double series of relationship capable of being expressed as follows:—

I am not familiar enough with the early grammarians to know when the terms lene and aspirate were first used. They were the Latin equivalents to the Greek words ψίλον (psilon) and δάσυ (dasy) respectively. The Greek terms are preferable. They convey no determinate idea, whereas the Latin terms convey a false one. The origin of the word aspirate I imagine to be as follows. The Latin language, wanting both the sound of the Greek theta, and the sign to express it (θ) rendered it by th. This orthography engenders the false notion that θ differed from τ by the addition of the aspirate h. To guard against similar false notions, I rarely hereafter use the word aspirate without qualifying it by the addition of the adjective so-called.

All the so-called aspirates are continuous; and, with the exception of s and z, all the lenes are explosive.

I believe that in the fact of each mute appearing in a fourfold form (i.e. sharp, or flat, lene, or (so-called) aspirate), lies the essential character of the mutes as opposed to the liquids.

Y and w.—These sounds, respectively intermediate to γ and i (the ee in feet), and to υ and u (oo in book), form a transition from the vowels to the consonants.

§ 204. It has been seen that the sixteen mutes are reducible to four series. Of these series, p, t, k, s, may respectively be taken as the types. Of the liquids it may be predicated as follows:—

1. That m is allied to the series p.—The combination inp has a tendency to become imp.

2. That n is allied to the series t.—The combination imt has a tendency to become either impt, or int.

3. That l is allied to the series k.—The evidence of this lies deep in comparative philology.

4. That r is allied to the series s.—The evidence of this is of the same nature with that of the preceding assertion.

The series p and k have this peculiarity.—They are connected with the vowels through w and u (oo), and through y and i (ee) respectively.

§ 205. The French word roi and the English words oil, {148}house, are specimens of a fresh class of articulations; viz., of compound vowel sounds or diphthongs. The diphthong oi is the vowel o modified, plus the semivowel y (not the vowel i) modified. The diphthongal sound in roi is the vowel o modified, plus the semivowel w (not the vowel u or oo) modified. In roi the semivowel element precedes, in oil it follows. In roi it is the semivowel allied to series p; in oil it is the semivowel allied to series k. The nature of the modification that the component parts of a diphthong undergo has yet to be determined; although it is certain there is one. If it were not so, the articulations would be double, not compound.

1. Diphthongs with the semivowel w, a) preceding, as in the French word roi, b) following, as in the English word new.

2. Diphthongs with the semivowel y, a) preceding, as is common in the languages of the Lithuanic and Slavonic stocks, b) following, as in the word oil.

The diphthongs in English are four; ow as in house, ew as in new, oi as in oil, i as in bite, fight.

§ 206. Chest, jest.—Here we have compound consonantal sounds. The ch in chest is t + sh (σ), the j in jest is d + zh (ζ). I believe that in these combinations one or both the elements, viz., t and sh, d and zh, are modified; but I am unable to state the exact nature of this modification.

§ 207. Ng.—The sound of the ng in sing, king, throng, when at the end of a word, or of singer, ringing, &c. in the middle of a word, is not the natural sound of the combination n and g, each letter retaining its natural power and sound; but a simple single sound, of which the combination ng is a conventional mode of expressing.

§ 208. Other terms, chiefly relating to the vowels, have still to be explained. The é of the French has been called fermé, or close (Italian, chiuso). Its opposite, the a in fate, is open.

Compared with a in fate, and the o in note, a in father, {149}and the aw in bawl, are broad, the vowels of note and fate being slender.

§ 209. In fat, the vowel is, according to common parlance, short; in fate, it is long. Here we have the introduction of two fresh terms. For the words long and short, I would fain substitute independent and dependent. If from the word fate I separate the final consonantal sound, the syllable, fa remains. In this syllable the a has precisely the sound that it had before. It remains unaltered. The removal of the consonant has in nowise modified its sound or power. It is not so with the vowel in the word fat. If from this I remove the consonant following, and so leave the a at the end of the syllable, instead of in the middle, I must do one of two things: I must sound it either as the a in fate, or else as the a in father. Its (so-called) short sound it cannot retain, unless it be supported by a consonant following. For this reason it is dependent. The same is the case with all the so-called short sounds, viz., the e in bed, i in fit, u in bull, o in not, u in but.

1. That the words independent and dependent correspond with the terms perfect and imperfect of the Hebrew grammarians.

2. That the Hebrew grammars give us the truest notions respecting these particular properties of vowels.

The following sentences are copied from Lee's Hebrew Grammar, Art. 33, 34:—"By perfect vowels is meant, vowels which, being preceded by a consonant" (or without being so preceded), "will constitute a complete syllable, as בָּ‎ bā. By imperfect vowels is meant those vowels which are not generally" (never) "found to constitute syllables without either the addition of a consonant or of an accent. Such syllables, therefore, must be either like בּדַ‎ bad, or בֲּ‎ bā, i.e., followed by a consonant, or accompanied by an accent." For further remarks on this subject, see the chapter on accent.

§ 210. Before i, e, and y of the English alphabet, and before ü and ö German, the letters c and g have the tendency to assume the sound and power of s or z, of sh or zh, of ch or j; {150}in other words, of becoming either s or some sound allied to s. Compared with a, o, and u (as in gat, got, gun), which are full, i, e, y, are small vowels.

It not every vowel that is susceptible of every modification. I (ee) and u (oo) are incapable of becoming broad. E in bed (as I have convinced myself), although both broad and slender, is incapable of becoming independent. For the u in but, and for the ö of certain foreign languages, I have no satisfactory systematic position.

From these, the semivowels w and y make a transition to the consonants v and the so-called aspirate of g (γ, not being in English), respectively.

§ 213. Concerning the vowel system I venture no assertion. The consonantal system I conceive to have been exhibited above in its whole fulness. The number of mutes, specifically distinct, I consider to be sixteen and no more: the number of liquids, four. What then are the powers of the numerous letters in alphabets like those of Arabia and Armenia? What {151}is the German ch, and Irish gh? Varieties of one or other of the sounds exhibited above, and not articulations specifically distinct.

§ 214. There is a difference between a connexion in phonetics and a connexion in grammar.—Phonetics is a word expressive of the subject-matter of the present chapter. The present chapter determines (amongst other things) the systematic relation of articulate sounds. The word phônæticos (φωνήτιχος) signifies appertaining to articulate sounds. It is evident that between sounds like b and v, s and z, there is a connexion in phonetics. Now in the grammar of languages there is often a change, or a permutation of letters: e.g., in the words tooth, teeth, the vowel, in price, prize, the consonant, is changed. Here there is a connexion in grammar.

That the letters most closely allied in phonetics should be most frequently interchanged in grammar, is what, on à priori grounds, we most naturally are led to expect. And that such is often the case, the study of languages tells us. That, however, it is always so, would be a hasty and an erroneous assertion. The Greek language changes p into f. Here the connexion in phonetics and the connexion in language closely coincide. The Welsh language changes p into m. Here the connexion in phonetics and the connexion in language do not closely coincide.

CHAPTER III.

§ 215. Certain combinations of articulate sounds are incapable of being pronounced. The following rule is one that, in the forthcoming pages, will frequently be referred to. Two (or more) mutes, of different degrees of sharpness and flatness, are incapable of coming together in the same syllable. For instance, b, v, d, g, z, &c. being flat, and p, f, t, k, s, &c. being sharp, such combinations as abt, avt, apd, afd, agt, akd, atz, ads, &c., are unpronounceable. Spelt, indeed, they may be; but attempts at pronunciation end in a change of the combination. In this case either the flat letter is changed to its sharp equivalent (b to p, d to t, &c.) or vice versâ (p to b, t to d). The combinations abt, and agt, to be pronounced, must become either apt or abd, or else akt or agd.

For determining which of the two letters shall be changed, in other words, whether it shall be the first that accommodates itself to the second, or the second that accommodates itself to the first, there are no general rules. This is settled by the particular habit of the language in consideration.

The word mutes in the second sentence of this section must be dwelt on. It is only with the mutes that there is an impossibility of pronouncing the heterogeneous combinations above mentioned. The liquids and the vowels are flat; but the liquids and vowels, although flat, may be followed by a sharp consonant. If this were not the case, the combinations ap, at, alp, alt, &c. would be unpronounceable.

§ 216. Unstable combinations.—That certain sounds in combination with others have a tendency to undergo changes, may be collected from the observation of our own language, as we find it spoken by those around us, or by ourselves. The ew in new is a sample of what may be called an unsteady or unstable combination. There is a natural tendency to change it either into oo (noo) or yoo (nyoo); perhaps also into yew (nyew).

§ 217. Effect of the semivowel y on certain letters when they precede it.—Taken by itself the semivowel y, followed by a vowel (ya, yee, yo, you, &c.), forms a stable combination. Not so, however, if it be preceded by a consonant, of the series t, k, or s, as tya, tyo; dya, dyo; kya, kyo; sya, syo. There then arises an unstable combination. Sya and syo we pronounce as sha and sho; tya and tyo we pronounce as cha and ja (i.e. tsh, dzh.). This we may verify from our pronunciation of words like sure, picture, verdure (shoor, pictshoor, verdzhoor), having previously remarked that the u in those words is not sounded as oo but as yoo. The effect of the semivowel y, taken with instability of the combination ew, accounts for the tendency to pronounce dew as if written jew.

§ 218. The evolution of new sounds.—To an English ear the sound of the German ch falls strange. To an English organ it is at first difficult to pronounce. The same is the case with the German vowels ö and ü and with the French sounds u, eu, &c.

To a German, however, and a Frenchman, the sound of the English th (either in thin or thine) is equally a matter of difficulty.

The reason of this lies in the fact of the respective sounds being absent in the German, French, and English languages; since sounds are easy or hard to pronounce just in proportion as we have been familiarised with them.

There is no instance of a new sound being introduced at once into a language. Where they originate at all, they are evolved, not imported. {154}

§ 219. Evolution of sounds.—Let there be a language where there is no such a sound as that of z, but where there is the sound of s. The sound of z may be evolved under (amongst others) the following conditions. 1. Let there be a number of words ending in the flat mutes; as slab, stag, stud, &c. 2. Let a certain form (the plural number or the genitive case) be formed by the addition of is or es; as slabis, stages, studes, &c. 3. Let the tendency that words have to contract eject the intermediate vowel, e or i, so that the s of the inflexion (a sharp mute) and the b, d, g, &c. of the original word (flat mutes) be brought into juxta-position, slabs, studs, stags. There is then an incompatible termination, and one of two changes must take place; either b, d, or g must become p, t, or k (slaps, staks, stuts); or s must become z (stagz, studz, slabz). In this latter case z is evolved. Again,

Let there be a language wherein there are no such sounds as sh, ch (tsh), or j (dzh); but where there are the sounds of s, t, d, and y.

Let a change affect the unstable combinations sy, ty, dy. From this will arise the evolved sounds of sh, ch, and j.

The phenomena of evolution help to determine the pronunciation of dead languages.

§ 220. On the value of a sufficient system of sounds.—In certain imaginable cases, a language may be materially affected by the paucity of its elementary articulate sounds.

In a given language let there be the absence of the sound z, the other conditions being those noted in the case of the words stag, slab, stud, &c. Let the intermediate vowel be ejected. Then, instead of the s being changed into an evolved z, let the other alternative take place; so that the words become staks, slaps, stuts. In this latter case we have an alteration of the original word, brought about by the insufficiency of the system of articulate sounds.

§ 221. Double consonants rare.—It cannot be too clearly understood that in words like pitted, stabbing, massy, &c. there is no real reduplication of the sounds of t, b, and s, respectively. Between the words pitted (as with the small-pox) and pitied (as being an object of pity) there is a difference in {155}spelling only. In speech the words are identical. The reduplication of the consonant is in English, and the generality of languages, a conventional mode of expressing upon paper the shortness (dependence) of the vowel that precedes.

§ 222. Real reduplications of consonants, i.e., reduplications of their sound, are, in all languages, extremely rare. I am fully aware of certain statements made respecting the Laplandic and Finlandic languages, viz., that doubled consonants are, in them, of common occurrence. Notwithstanding this, I have an impression that it is generally under one condition that true reduplication takes place. In compound and derived words, where the original root ends, and the superadded affix begins with the same letter, there is a reduplication of the sound, and not otherwise. In the word soulless, the l is doubled to the ear as well as to the eye; and it is a false pronunciation to call it souless (soless). In the "Deformed Transformed" it is made to rhyme with no less, improperly.

In the following words, all of which are compounds, we have true specimens of the doubled consonant.

It must not, however, be concealed, that, in the mouths even of correct speakers, one of the doubled sounds is often dropped.

§ 223. True aspirates rare.—The criticism applied to words like pitted, &c., applies also to words like Philip, thin, thine, &c. There is therein no sound of h. How the so-called aspirates differ from their corresponding lenes has not yet been determined. That it is not by the addition of h is evident. Ph and th are conventional modes of spelling simple single sounds, which might better be expressed by simple single signs. {156}

In our own language the true aspirates, like the true duplications, are found only in compound words; and there they are often slurred in the pronunciation.

Now in certain languages the true aspirates are of common occurrence, i.e., sounds like the t in nuthook, the ph in haphazard, &c., are as frequent as the sounds of p, b, s, &c. In the spelling of these sounds by means of the English we are hampered by the circumstance of th and ph being already used in a different sense.

CHAPTER IV.

§ 224. 1. Let there be two syllables, of which the one ends in m, and the other begins with r, as we have in the syllables num- and -rus of the Latin word numerus.

2. Let an ejection of the intervening letters bring these two syllables into immediate contact, numrus. The m and r form an unstable combination. To remedy this there is a tendency (mark, not an absolute necessity) to insert an intervening sound.

In English, the form which the Latin word numerus takes is number; in Spanish, nombre. The b makes no part of the original word, but has been inserted for the sake of euphony; or, to speak more properly, by a euphonic process. The word euphony is derived from εὖ (well), and φώνη (fônæ, a voice). The province of euphony has not been very accurately determined.

§ 225. In the word number, nombre, the letter inserted was b; and for b being the particular letter employed, there is a reason derived from the system of articulate sounds.

1. That the letter inserted should be a consonant is evident. The vowel e (in numerus) had been previously ejected.

2. That it should be a mute is evident. A liquid would have given the unstable or unpronounceable combinations mnr, mlr, mrr, mmr.

3. That it should be a consonant, either of series b or of series s, was natural; it being series b and series s with which m and r are respectively connected.

4. That it should be a consonant of series b, rather than one of series s, we collect from the fact that msr (numsrus) or mzr (numzrus) give inharmonious, and, consequently, unstable combinations. {158}

5. That of the b series, it should be b or v (flat) rather than p or f (sharp), we infer from the fact of m and r both being flat.

6. Of v and b, the latter alone gives a stable combination, so that we have the Spanish form nombre, and not nomvre.

In this we have an illustration of the use of attending to the nature and connections of articulate sounds in general.

§ 226. The affinity of m for the series b, of n for the series t, gives occasion to further euphonic changes. The combinations mt, md, mþ, mð, are unstable. The syllables emt, emd, are liable to one of two modifications. Either p or b will be inserted, and so make them empt (as in tempt), embd (as in Embden), or else the m will become n, forming the syllable ent, end, enþ, enð.

Similar tendencies, in a certain degree, affect the combinations enp, enb. They are liable to become emp, or emb. Any one may see that the word enperor embarrasses the utterance.

§ 227. The combination tupt is stable, so also is the combination tuft. But the combination tupth is unstable: since the p is lene, the þ is a (so-called) aspirate. Hence arises a process of accommodation by which the word becomes either tupt or tufth (tufþ).

In respect to the unstable combination tupth, we may observe this, viz. that the ways of altering it are two. Either the first letter may be accommodated to the second, tufþ, or the second may be accommodated to the first, tupt. Which of these two changes shall take place is determined by the particular habit of the language. In Greek we add to the radical syllable τυπ-, the inflectional syllable -θην. The first letter, π, is accommodated to the second, θ, and the word becomes τυφθην (tyfþæn), as in ἐτύφθην (etyfþæn). In English we add to the radical syllable stag, the inflectional syllable s. Here the second letter is accommodated to the first, and the resulting word is not staks, but stagz.

§ 228. The Irish Gaelic, above most other languages, illustrates a euphonic principle that modifies the vowels of a word. The vowels a, o, u, are full, whilst i, e, y, are small. Now if to a syllable containing a small vowel, as buil, there be added {159}a syllable containing a broad one, as -am, a change takes place. Either the first syllable is accommodated to the second, or the second to the first; so that the vowels respectively contained in them are either both full or both small. Hence arises, in respect to the word quoted, either the form bualam, or else the form builim.

§ 229. In the words give and gave we have a change of tense expressed by a change of vowel. In the words price and prize a change of meaning is expressed by a change of consonant. In clothe and clad there is a change both of a vowel and of a consonant. In the words to use and a use there is a similar change, although it is not expressed by the spelling. To the ear the verb to use ends in z, although not to the eye. The following are instances of the permutation of letters.

In have and had we have the ejection of a sound; in work and wrought, the transposition of one. Important changes are undergone by the sounds k, g, and the allied ones nk, ng, y, as will be seen in the chapter on verbs.

It must be noticed that the list above is far from being an exhaustive one. The expression too of the changes undergone has been rendered difficult on account of the imperfection of our orthography. The whole section has been written in illustration of the meaning of the word permutation, rather than for any specific object in grammar.

§ 230. In all the words above the change of sound has been brought about by the grammatical inflection of the word wherein it occurs. This is the case with the words life and live, and with all the rest. With the German word leben, compared with the corresponding word live, in English, the change is similar. It is brought about, however, not by a grammatical inflection, but by a difference of time, and by a difference of place. This indicates the distinction between the permutation of letters and the transition of letters. In dealing with permutations, we compare different parts of speech; in dealing with transitions, we compare different languages, or different stages of a single language.

CHAPTER V.

§ 231. In respect to the formation of syllables, I am aware of no more than one point that requires any especial consideration.

In certain words, of more than one syllable, it is difficult to say to which syllable an intervening consonant belongs. For instance, does the v in river, and the v in fever, belong to the first or the second syllable? Are the words to be divided thus, ri-ver, fe-ver? or thus, riv-er, fev-er?

In the first place, the case is capable of being viewed in two points of view—an etymological and a phonetic one.

That the c and r in become, berhymed, &c. belong to the second syllable, we determine at once by taking the words to pieces; whereby we get the words come and rhymed in an isolated independent form. But this fact, although it settles the point in etymology, leaves it as it was in phonetics; since it in nowise follows, that, because the c in the simple word come is exclusively attached to the letter that follows it, it is, in the compound word become, exclusively attached to it also.

To the following point of structure in the consonantal sounds the reader's attention is particularly directed.

1. Let the vowel a (as in fate) be sounded.—2. Let it be followed by the consonant p, so as to form the syllable āp. To form the sound of p, it will be found that the lips close on the sound of a, and arrest it. Now, if the lips be left to themselves they will not remain closed on the sound, but will open again, in a slight degree indeed, but in a degree sufficient to cause a kind of vibration, or, at any rate, to allow an {162}escape of the remainder of the current of breath by which the sound was originally formed. To re-open in a slight degree is the natural tendency of the lips in the case exhibited above.

Now, by an effort, let this tendency to re-open be counteracted. Let the remaining current of breath be cut short. We have, then, only this, viz., so much of the syllable āp as can be formed by the closure of the lips. All that portion of it that is caused by their re-opening is deficient. The resulting sound seems truncated, cut short, or incomplete. It is the sound of p, minus the remnant of breath. All of the sound p that is now left is formed, not by the escape of the breath, but by the arrest of it.

The p in āp is a final sound. With initial sounds the case is different. Let the lips be closed, and let an attempt be made to form the syllable pa by suddenly opening them. The sound appears incomplete; but its incompleteness is at the beginning of the sound, and not at the end of it. In the natural course of things there would have been a current of breath preceding, and this current would have given a vibration, now wanting. All the sound that is formed here is formed, not by the arrest of breath, but by the escape of it.

I feel that this account of the mechanism of the apparently simple sound p, labours under all the difficulties that attend the description of a sound; and for this reason I again request the reader to satisfy himself either of its truth or its inaccuracy, before he proceeds to the conclusions that will be drawn from it.

The account, however, being recognised, we have in the current natural sound of p two elements:—

1. That formed by the current of air and the closure of the lips, as in āp. This may be called the sound of breath arrested.

2. That formed by the current of air and the opening of the lips, as in pā. This may be called the sound of breath escaping.

Now what may be said of p may be said of all the other consonants, the words tongue, teeth, &c. being used instead of lips, according to the case. {163}

Let the sound of breath arrested be expressed by π, and that of breath escaping be expressed by ϖ, the two together form the current natural sound p (π+ϖ=p).

Thus āp (as quoted above) is p - ϖ, or π; whilst pa (sounded similarly) is p - π, or ϖ.

In the formation of syllables, I consider that the sound of breath arrested belongs to the first, and the sound of breath escaping to the second syllable; that each sound being expressed by a separate sign, the word happy is divided thus, haπ-ϖy; and that such is the case with all consonants between two syllables. The whole consonant belongs neither to one syllable nor the other. Half of it belongs to each. The reduplication of the p in happy, the t in pitted, &c, is a mere point of spelling, of which more will be said in the chapter on orthography.

CHAPTER VI.

§ 232. The dependent vowels, as the a in fat, i in fit, u in but, o in not, have this character; viz. they are all uttered with rapidity, and pass quickly in the enunciation, the voice not resting on them. This rapidity of utterance becomes more evident when we contrast with them the prolonged sounds of the a in fate, ee in feet, oo in book, o in note; wherein the utterance is retarded, and wherein the voice rests, delays, or is prolonged. The f and t of fate are separated by a longer interval than the f and t of fat; and the same is the case with fit, feet, &c.

Let the n and the t of not be each as 1, the o also being as 1: then each letter, consonant or vowel, shall constitute ⅓ of the whole word.

Let, however, the n and t of note be each as 1, the o being as 2. Then, instead of each consonant constituting ⅓ of the whole word, it shall constitute but ¼.

Upon the comparative extent to which the voice is prolonged, the division of vowels and syllables into long and short has been established: the o in note being long, the o in not being short. And the longness or shortness of a vowel or syllable is said to be its quantity.

§ 233. The division of vowels into long and short coincides nearly with the division of them into independent and dependent. Mark the word vowels, and mark the word nearly. In the length and shortness of vowels there are degrees. This is especially the case with the broad vowels. The a in father is capable of being pronounced either very quickly, or very slowly. It may be attend most rapidly and yet preserve its broad character, i.e., become neither the a in fat, nor the a in fate. {165}

Subject to the views laid down in the next section, the vowel ee in seeing is long, and it is certainly independent. Whether the syllable see- be long is another question.

Clear notions upon these matters are necessary for determining the structure of the English and classical metres.

§ 234. The qualified manner in which it was stated that the vowel in the word seeing was long, and the attention directed to the word vowels in the preceding section, arose from a distinction, that is now about to be drawn, between the length of vowels and the length of syllables.

The independent vowel in the syllable see- is long; and long it remains, whether it stand as it is, or be followed by a consonant, as in seen, or by a vowel, as in see-ing.

The dependent vowel in the word sit is short. If followed by a vowel it becomes unpronounceable, except as the ea in seat or the i in sight. By a consonant, however, it may be followed, and still retain its dependent character and also its shortness. Such is the power it has in the word quoted, sit. Followed by a second consonant, it still retains its shortness, e.g., sits. Whatever the comparative length of the syllables, see and seen, sit and sits, may be, the length of their respective vowels is the same.

Now, if we determine the character of the syllable by the character of the vowel, all syllables are short wherein there is a short vowel, and all are long wherein there is a long one. Measured by the quantity of the vowel the word sits is short, and the syllable see- in seeing is long.

But it is well known that this view is not the view commonly taken of the syllables see (in seeing) and sits. It is well known, that, in the eyes of a classical scholar, the see (in seeing) is short, and that in the word sits the i is long. The classic differs from the Englishman thus,—He measures his {166}quantity, not by the length of the vowel but, by the length of the syllable taken altogether. The perception of this distinction enables us to comprehend the following statements.

I. That vowels long by nature may appear to become short by position, and vice versâ.

II. That, by a laxity of language, the vowel may be said to have changed its quantity, whilst it is the syllable alone that has been altered.

III. That, if one person measures his quantities by the vowels, and another by the syllables, what is short to the one, shall be long to the other, and vice versâ. The same is the case with nations.

IV. That one of the most essential differences between the English and the classical languages is that the quantities (as far as they go) of the first are measured by the vowel, those of the latter by the syllable. To a Roman the word monument consists of two short syllables and one long one; to an Englishman it contains three short syllables.

These remarks are appreciated when we consider the comparative characters of the classical and the English prosody.

CHAPTER VII.

§ 235. In the word tyrant there is an emphasis, or stress, upon the first syllable. In the word presume there is an emphasis, or stress, on the second syllable. This emphasis, or stress, is called Accent. The circumstance of a syllable bearing an accent is sometimes expressed by a mark (´); in which case the word is said to be accentuated, i.e., to have the accent signified in writing.

Words accented on the last syllable—Brigáde, preténce, harpoón, reliéve, detér, assúme, besoúght, beréft, befóre, abroád, abóde, abstrúse, intermíx, superádd, cavaliér.

Words accented on the last syllable but one—An´chor, ar´gue, hásten, fáther, fóxes, smíting, húsband, márket, vápour, bárefoot, archángel, bespátter, disáble, terrífic.

Words accented on the last syllable but two—Régular, an´tidote, for´tify, suscéptible, incontrovértible.

Words accented on the last syllable but three (rare)—Réceptacle, régulating, tálkativeness, ábsolutely, lúminary, inévitable, &c.

A great number of words are distinguished by the accent alone. The following list is from Nares' Orthoepy, a work to which the reader is referred.

That class of words that by a change of accent are converted from nouns into verbs (súrvey, survéy, cóntrast, contrást, &c.) will be noticed more at large in the Chapter on Derivation.

§ 236. In words like thínking, fóxes, lon´ger, len´gthen, &c. we have two parts; first the original word, the root, or the radical part, as think, fox, long, length, &c.; and next, the inflectional, or the subordinate part, -ing, -es, -er, -en, &c.

To assert as a universal rule that the accent is always on the root, and never on the subordinate part of a word, is too much. Although in the English language such an assertion (with one exception) is found true; by the French and other languages it is invalidated.

In words like len´g-then-ing, we have a second inflectional or subordinate syllable; and the accent remains in its original place, absolutely, but not relatively. It is all the farther from the end of the word. Besides indicating the propriety of determining the place of the accent by counting from the end, rather than the beginning of a word, this circumstance indicates something else.

Imagine the English participles to be declined, and to possess cases, formed by the addition of fresh syllables. In this case the word len´gthening would become a quadri-syllable. But to throw the accent to the fourth syllable from the end is inconvenient. Hence a necessity of removing it from the radical, and placing it on an inflectional syllable.

The German word lében (to live) illustrates the foregoing sentence. Léb- is the root, léb-end=living, from whence lebéndig=lively (with the accent on an inflectional syllable), although this last word might without inconvenience have been accented on the first syllable; that being only the third from the end.

Confusion between the radical and inflectional syllables of a word, arising from the situation of the accent, may work the deterioration of a language.

§ 237. In týrant and presúme, we deal with single words; and in each word we determine which syllable is accented. {169}Contrasted with the sort of accent that follows, this may be called a verbal accent.

the pronoun us is strongly brought forward. An especial stress or emphasis is laid upon it, denoting that there are other beings to whom it might not appear, &c. This is collected from the context. Here there is a logical accent. "When one word in a sentence is distinguished by a stress, as more important than the rest, we may say that it is emphatical, or that an emphasis is laid upon it. When one syllable in a word is distinguished by a stress, and more audible than the rest, we say that it is accented, or that an accent is put upon it. Accent, therefore, is to syllables what emphasis is to sentences; it distinguishes one from the crowd, and brings it forward to observation."—(Nares' Orthoepy, Part II. Chap. I.)

§ 238. Accent plays an important part in determining the nature of certain compound words—For this, see the Chapter on Composition.

It also plays an important part in determining the nature of the English metres—See Prosody.

Thirdly (the subject of the present section), it plays an important part in all systems of orthography.

The quotation from Professor Lee's Hebrew Grammar, in p. 149, is referred to; and a particular attention to a somewhat difficult subject is requisite.

The second syllable in the word monument is what a classical scholar would call short. The vowel is short, and the syllable taken altogether is short. Herein it agrees with the first syllable mon-. It differs, however, from the syllable mon- in being destitute of an accent, mónument. With the third syllable -ment, it agrees in the eyes of an Englishman, but differs in the eyes of a scholar. The vowels u and e are equally short, and, as the Englishman measures by the vowel {170}the syllables -u and -ment are both short. Not so, however, with the scholar. He measures by the syllable and determines that the e, although naturally a short vowel, is made long by position. However, in being each destitute of an accent the syllables -u and -ment agree. Be it remarked a second time that the accent in mónument lies on the first syllable.

If, however, the syllable -nu take an accent; that is, if the place of the accent be removed from the first to the second syllable, the vowel u still being kept short, we have a word which we spell thus, monumment. Now the u in monumment is not only short, but dependent. It is upon this effect of an accent that the quotation from Lee's Hebrew Grammar, p. 149, especially bears.

And now two questions arise:—1. How is it that the accent has the effect of rendering such a syllable as the u in monumment dependent? 2. Why do we in spelling such a syllable double the consonant?

An accent falling upon a syllable must, of necessity, do one of two things: it must affect the vowel, or it must affect the consonant. If it affect the vowel, the vowel becomes the predominant part of the syllable, as in mónooment; but, if it affect the consonant, the consonant becomes the predominant part of the syllable, as monum´ment.

In words like monumment the consonant is, strictly speaking, as single as it is in monument, or monooment. Its absolute sound is the same. Not so its relative sound. This is exaggerated by two circumstances:—1, The comparative shortness of the vowel u; 2, the fact of the accent falling on it. The increased relative importance of the letter m in the word monumment is mistaken for a reduplication of the sound. This is the reason why in most languages the shortness of a vowel is expressed by the doubling of the consonant following; this doubling being no true reduplication of the sound, but a mere orthographical conventionality.

§ 239. Accent and quantity, as may have been collected from pp. 164-167, do not coincide. Nothing shows this more {171}clearly than words like the adjective augúst, and the substantive Aúgust (the month), where the quantity remains the same, although the accent is different. The following quotation from Mr. Guest's English Rhythms is made for the sake of four things:—

1. Of showing that the generality of writers have the credit of confusing accent with quantity—

3. Of indicating the propriety of the expressions in italics—It is not stated that the consonant c is doubled, but that it is added to the first syllable. The difference lies, not in its reduplication, but in its distribution.

4. Of taking a slight exception—A syllable (accented or unaccented) must be either independent or dependent; if the latter, then in most immediate contact with the consonant that follows.

CHAPTER VIII.

§ 240. The present chapter is one, not upon the details of the pronunciation of the English language, but upon the principles of orthoepy. For the details of pronunciation the reader is referred to Nares' Orthoepy, and to the common pronouncing dictionaries, with the preliminary recommendation to use them with caution. Orthoepy, a word derived from the Greek orthon (upright), and epos (a word), signifies the right utterance of words. Orthoepy differs from orthography by determining how words are spoken, whereas orthography decides how they are spelt. The one is a question of speech, the other a question of spelling. Orthography presupposes orthoepy.

§ 241. Of pronunciation there are two kinds, the colloquial and the rhetorical. In common conversation we pronounce the i in wind, like the i in bit; in rehearsing, or in declamation, however, we pronounce it like the i in bite; that is, we give it a diphthongal sound. In reading the Scriptures we say blesséd; in current speech we say blest. It is the same with many words occurring in poetry.

§ 242. Errors in pronunciation are capable of being classified. In the first place, they may be arranged according to their situation. The man who pronounces the verb to survéy, as if it was súrvey (that is, with the accent on the wrong syllable), errs in respect to the accentuation of the word; the situation, or seat of his error, being the accent. To say orātor instead of orător is to err in respect to the quantity of the word, the seat of the error being in the quantity; and to pronounce the a in father, as it is pronounced in Yorkshire, or the s in sound, as it is pronounced in Devonshire (that is, as z), is to err in {173}the matter of the articulate sounds. To mispronounce a word because it is misspelt^[34] is only indirectly an error of orthoepy. It is an error, not so much of orthoepy, as of orthography; and to give a wrong inflection to a word is not bad pronunciation but bad grammar. For practical purposes, however, many words that are really points of grammar and of orthography, may be dealt with as points of orthoepy.

That the preceding classification is natural I am induced to believe by the following circumstances. Errors in the way of articulation generally arise from a source different from those of accent and of quantity. Errors in accent and quantity are generally referable to insufficient grammatical or etymological knowledge, whilst the errors of articulation betray a provincial dialect.

The misdivision of syllables, an orthoepical error of a fourth kind, has in the English, and perhaps in other languages, given rise to a peculiar class of words. There have been those who have written a nambassador for an ambassador, misdividing the syllables, and misdistributing the sound of the letter n. The double form (a and an) of the English indefinite article, encourages this misdivision. Now, in certain words an error of this kind has had a permanent influence. The English word nag is, in Danish, ög; the n, in English, having originally belonged to the indefinite an, which preceded it. The words, instead of being divided thus, an ag, were divided thus, a nag, and the fault became perpetuated. That the Danish is the true form we collect, firstly, from the ease with which the English form is accounted for, and, secondly, from the old Saxon form ehu, Latin equus. In adder we have the process reversed. The true form is nadder, old English; natter, German. Here the n is taken from the substantive and added to the article. In newt and eft we have each form. The list of words of this sort can be increased.

§ 243. In the second place, faults of pronunciation may be arranged according to their cause.

1. The fault of incompetent enunciation.—A person who says sick for thick, or elebben for eleven, does so, not because he knows no better, but because he cannot enounce the right sounds of th and v. He is incompetent to it. His error is not one of ignorance. It is an acoustic or a phonetic defect. As such it differs from—

2. The fault of erroneous enunciation.—This is the error of a person who talks of jocholate instead of chocolate. It is not that he cannot pronounce rightly, but that he mistakes the nature of the sound required. Still more the person who calls a hedge a nedge, and an edge a hedge.

§ 244. Incompetent enunciation, and erroneous enunciation are, however, only the proximate and immediate causes of bad orthoepy. Amongst the remote causes (the immediate causes of erroneous enunciation) are the following.

I. Undefined notions as to the language to which a word belongs.—The flower called anemone is variously pronounced. Those who know Greek say anemōne, speaking as if the word was written anemohny. The mass say, anemŏne, speaking as if the word was written anemmony. Now, the doubt here is as to the language of the word. If it be Greek, it is anemōne.

And if it be English, it is (on the score of analogy) as undoubtedly anémmony. The pronunciation of the word in point is determined when we have determined the language of it.

II. Mistakes as to fact, the language of a word being determined.—To know the word anemōne to be Greek, and to use it as a Greek word, but to call it anemŏny, is not to be undecided as to a matter of language, but to be ignorant as to a matter of quantity.

III. Neglect of analogy.—Each and all of the following words, orator, theatre, senator, &c. are in the Latin language, from whence they are derived, accented on the second syllable; as orátor, theátre, senátor. In English, on the contrary, they are accented on the first; as órator, théatre, {175}sénator. The same is the case with many other words similarly derived. They similarly suffer a change of accent. So many words do this, that it is the rule in English for words to throw their accent from the second syllable (counting from the end of the word) to the third. It was on the strength of this rule,—in other words, on the analogies of orator, &c., that the English pronunciation of the Greek word ἀνεμώνη was stated to be anémmone. Now, to take a word derived from the Latin, and to look to its original quantity only, without consulting the analogies of other words similarly derived, is to be neglectful of the analogies of our own language, and attentive to the quantities of a foreign one.

These, amongst others, the immediate causes of erroneous enunciation, have been adduced not for the sake of exhausting, but for the sake of illustrating the subject.

§ 245. In matters of orthoepy it is the usual custom to appeal to one of the following standards.

I. The authority of scholars.—This is of value up to a certain point only. The fittest person for determining the classical pronunciation of a word like anemone is the classical scholar; but the mere classical scholar is far from being the fittest person to determine the analogies that such a word follows in English.

II. The usage of educated bodies, such as the bar, the pulpit, the senate, &c.—These are recommended by two circumstances: 1. The chance that each member of them is sufficiently a scholar in foreign tongues to determine the original pronunciation of derived words, and sufficiently a critic in his own language to be aware of the analogies that are in operation. 2. The quantity of imitators that, irrespective of the worth of his pronunciation, each individual can carry with him. On this latter ground the stage is a sort of standard.

The objection to the authority of educated bodies is its impracticability. It is only the usage of the component individuals that can be determined. Of these many may carry with them the dialects of their provinces, so that, although good standards on points of accent and quantity, they are bad ones upon points of articulation. {176}

III. The authority of societies constituted with the express purpose of taking cognizance of the language of the country.—These, although recognized in Italy and other parts of the Continent, have only been proposed in Great Britain. Their inefficacy arises from the inutility of attempting to fix that which, like language, is essentially fluctuating.

IV. The authority of the written language.—The value of this may be collected from the chapter on orthography.

V. These, amongst others, the standards that have been appealed to, are adduced not for the sake of exhausting the subject, but to show the unsatisfactory nature of authority in matters of speech.

§ 246. For a person, on a point of pronunciation, to trust to his own judgment, he must be capable, with every word that he doubts about, of discussing three questions:—

I. The abstract or theoretical propriety of a certain pronunciation.—To determine this he must have a sufficient knowledge of foreign tongues and a sufficient knowledge of English analogies. He must also have some test by which he can determine to what language an equivocal word belongs. Of tests for this purpose, one, amongst others, is the following:—Let it be asked whether the word lens (in Optics) is English or Latin; whether it is to be considered as a naturalised word or a strange one. The following fact will give an answer. There is of the word lens a plural number, and this plural number is the English form lenses, and not the Latin form lentes. The existence of an English inflection proves that the word to which it belongs is English, although its absence does not prove the contrary. That the word anemone is English (and consequently pronounced anemŏne) we know from the plural form, which is not anemonæ, but anemones.

II. The preference of one pronunciation over another on the score of utility.—The word ascetic, for certain orthographical reasons, notwithstanding its origin from the Greek word askeó, is called assetic. From similar reasons there is a tendency to call the word sceptic, septic. Theoretical propriety (and, be it observed, the analogy of ascetic has not been overlooked) is in {177}favour of the word being sounded skeptic. The tendency of language, however, is the other way. Now, the tendency of language and the theoretical propriety being equal, there is an advantage (a point of utility) in saying skeptic, which turns the scale. By sounding the k we distinguish the word skeptic from septic. By this the language gains a point in perspicuity, so that we can talk of the anti-skeptic writings of Bishop Warburton and of the anti-septic properties of charcoal.

III. The tendencies of language.—From p. 153, we see that the combination ew is an unstable combination, that it has a tendency to become yoo, and that the y in yoo has a tendency to change a d preceding into j; in other words, we see the reason why, by many persons, dew is pronounced jew.

It is generally an easier matter to say how a word will be sounded a hundred years hence, than to determine its present pronunciation. Theoretical propriety is in favour of dew, so also is the view in the way of utility. Notwithstanding this, posterity will say jew, for the tendencies of language are paramount to all other influences.

We may now judge of the relative value of the three lines of criticism exhibited above. Other things being equal, the language should have the advantage of the doubt, and the utility of a given pronunciation should prevail over its theoretical propriety. Where, however, the tendencies are overwhelming, we can only choose whether, in doubtful words, we shall speak like our ancestors, or like our posterity.^[35]

CHAPTER IX.

§ 247. Orthoepy determines the correct pronunciation of words, and deals with a language as it is spoken; orthography determines the correct spelling of words, and deals with a language as it is written. The term is derived from the Greek words orthos (upright), and graphé, or grafæ (writing). Orthography is less essential to language than orthoepy; since all languages are spoken, whilst but a few languages are written. Orthography presupposes orthoepy. Orthography addresses itself to the eye, orthoepy to the ear. Orthoepy deals with the articulate sounds that constitute syllables and words; orthography treats of the signs by which such articulate sounds are expressed in writing. A letter is the sign of an articulate (and, in the case of h, of an inarticulate) sound.

A full and perfect system of orthography consists in two things:—1. The possession of a sufficient and consistent alphabet. 2. The right application of such an alphabet. This position may be illustrated more fully.

§ 248. First, in respect to a full and perfect alphabet. Let there be in a certain language, simple single articulate sounds, to the number of forty, whilst the simple single signs, or letters, expressive of them, amount to no more than thirty. In this case the alphabet is insufficient. It is not full enough: since ten of the simple single articulate sounds have no corresponding signs whereby they may be expressed. In our own language, the sounds (amongst others) of th in thin, and of th in thine, are simple and single, whilst there is no sign equally simple and single to spell them with.

An alphabet, however, may be sufficient, and yet imperfect. It may err on the score of inconsistency. Let there be in a {179}given language two simple single sounds, for instance, the p in pate, and the f in fate. Let these sounds stand in a given relation to each other. Let a given sign, for instance, פ‎ (as is actually the case in Hebrew), stand for the p in pate; and let a second sign be required for the f in fate. Concerning the nature of this latter sign, two views may be taken. One framer of the alphabet, perceiving that the two sounds are mere modifications of each other, may argue that no new sign (or letter) is at all necessary, but that the sound of f in fate may be expressed by a mere modification of the sign (or letter) פ‎, and may be written thus פּ‎, or thus פ‎´ or פ‎', &c.; upon the principle that, like sounds should be expressed by like signs. The other framer of the alphabet, contemplating the difference between the two sounds, rather than the likeness, may propose, not a mere modification of the sign פ‎, but a letter altogether new, such as f, or φ, &c., upon the principle that sounds of a given degree of dissimilitude should be expressed by signs of a different degree of dissimilitude.

Hitherto the expression of the sounds in point is a matter of convenience only. No question has been raised as to its consistency or inconsistency. This begins under conditions like the following:—Let there be in the language in point the sounds of the t in tin, and of the th in thin; which (it may be remembered) are precisely in the same relation to each other as the p in pate and the f in fate. Let each of these sounds have a sign (or letter) expressive of it. Upon the nature of these signs, or letters, will depend the nature of the sign or letter required for the f in fate. If the letter expressing the th in thin be a mere modification of the letter expressing the t in tin, then must the letter expressive of the f in fate be a mere modification of the letter expressing the p in pate, and vice versâ. If this be not the case, the alphabet is inconsistent.

In the English alphabet we have (amongst others) the following inconsistency:—The sound of the f in fate, in a certain relation to the sound of the p in pate, is expressed by a totally distinct sign; whereas, the sound of the th in thin (similarly related to the t in tin) is expressed by no new sign, but by a mere modification of t; viz., th. {180}

A third element in the faultiness of an alphabet is the fault of erroneous representation. The best illustration of this we get from the Hebrew alphabet, where the sounds of ת‎ and ט‎, mere varieties of each other, are represented by distinct and dissimilar signs, whilst ת‎ and תּ‎, sounds specifically distinct, are expressed by a mere modification of the same sign, or letter.

§ 249. The right application of an alphabet.—An alphabet may be both sufficient and consistent, accurate in its representation of the alliances between articulate sounds, and in nowise redundant; and yet, withal, it may be so wrongly applied as to be defective. Of defect in the use or application of the letters of an alphabet, the three main causes are the following:—

1. Unsteadiness in the power of letters.—Of this there are two kinds. In the first, there is one sound with two (or more) ways of expressing it. Such is the sound of the letter f in English. In words of Anglo-Saxon origin it is spelt with a single simple sign, as in fill; whilst in Greek words it is denoted by a combination, as in Philip. The reverse of this takes place with the letter g; here a single sign has a double power; in gibbet it is sounded as j, and in gibberish as g in got.

2. The aim at secondary objects.—The natural aim of orthography, of spelling, or of writing (for the three terms mean the same thing), is to express the sounds of a language. Syllables and words it takes as they meet the ear, it translates them by appropriate signs, and so paints them, as it were, to the eye. That this is the natural and primary object is self-evident; but beyond this natural and primary object there is, with the orthographical systems of most languages, a secondary one, viz. the attempt to combine with the representation of the sound of a given word the representation of its history and origin.

The sound of the c, in city, is the sound that we naturally spell with the letter s, and if the expression of this sound was the only object of our orthographists, the word would be spelt accordingly (sity). The following facts, however, traverse {181}this simple view of the matter. The word is a derived word; it is transplanted into our own language from the Latin, where it is spelt with a c (civitas); and to change this c into s conceals the origin and history of the word. For this reason the c is retained, although, as far as the mere expression of sounds (the primary object in orthography) is concerned, the letter is a superfluity. In cases like the one adduced the orthography is bent to a secondary end, and is traversed by the etymology.

3. Obsoleteness.—It is very evident that modes of spelling which at one time may have been correct, may, by a change of pronunciation, become incorrect; so that orthography becomes obsolete whenever there takes place a change of speech without a correspondent change of spelling.

§ 250. Difference between the change of a sound and the original false expression of a sound.—The letter u is a simple single sign. The sound of ow, in town, is a diphthongal, or a double, sound. Now, in Anglo-Saxon, the modern word town is spelt tún. In this case one of two things must have taken place: either the word must have changed its sound, or the Anglo-Saxons must have expressed it falsely and improperly.

§ 251. From the foregoing sections we arrive at the theory of a full and perfect alphabet and orthography, of which a few (amongst many others) of the chief conditions are as follow:—

1. That for every simple single sound, incapable of being represented by a combination of letters, there be a simple single sign.

2. That sounds within a determined degree of likeness be represented by signs within a determined degree of likeness; whilst sounds beyond a certain degree of likeness be represented by distinct and different signs, and that uniformly.

5. That the primary aim of orthography be to express the sounds of words, and not their histories. {182}

With these principles in our mind we may measure the imperfections of our own and of other alphabets.

§ 252. Previous to considering the sufficiency or insufficiency of the English alphabet, it is necessary to enumerate the elementary articulate sounds of the language. The enumeration of these is, strictly speaking, a point, not of orthography, but of orthoepy. It is, however, so intimately connected with the former that the present chapter seems its proper place. The vowels belonging to the English language are the twelve following:—

The consonantal sounds are, 1. the two semivowels; 2. the four liquids; 3. fourteen out of the sixteen mutes; 4. ch in chest, and j in jest, compound sibilants; 5. ng, as in king; 6. the aspirate h. In all, twenty-four.

Some writers would add to these the additional sound of the é fermé of the French; believing that the vowel in words like their and vein has a different sound from the vowel in words like there and vain. For my own part I cannot detect such a difference either in my own speech or that of my neighbours; although I am far from denying that in certain dialects of our language such may have been the case. The following is an extract from the Danish grammar for Englishmen, by Professor Rask, whose eye, in the matter in question, seems to have misled his ear: "The é fermé, or close é, is very frequent in Danish, but scarcely perceptible in English; unless in such words as, their, vein, veil, which appear to sound a little different from there, vain, vale."

The vowels being twelve, the diphthongs four, and the consonantal sounds twenty-four, we have altogether as many as forty sounds, some being so closely allied to each other as to be mere modifications, and others being combinations rather than simple sounds; all, however, agreeing in requiring to be expressed by letters or by combinations of letters, and to be distinguished from each other.

Now, although every sound specifically distinct should be expressed by a distinct sign, it does not follow that mere modifications or varieties (especially if they be within certain limits) should be so expressed. In the Greek language sounds as like as the o in not and the o in note are expressed by signs as unlike as ο and ω; that is, by the letters omicron and omega respectively; and so it is with ε and η. All that can be said in this case is, that it is the character of the Greek alphabet to represent a difference which the English neglects.

With respect to the diphthongs it is incorrect, uncommon, and inconvenient to represent them by simple single signs, rather than by combinations. In the English language the sounds {184}of ou, ew, and oi, are properly spelt with two letters. Not so, however, of i in bite.

The compound sibilants may also be expressed not by single signs, but by the combinations tsh and dzh; although, for certain reasons, such a mode of spelling is inconvenient. With these views we may appreciate,

A. In respect to the vowels.—Notwithstanding the fact that the sounds of the a in father, fate, and fat, and of the o and the aw in note, not, and bawl, are modifications of a and o respectively, we have still six vowel sounds specifically distinct, for which (y being a consonant rather than a vowel) we have but five signs. The u in duck, specifically distinct from the u in bull, has no specifically distinct sign to represent it.

B. In respect to the consonants.—The th in thin, the th in thine, the sh in shine, the z in azure, and the ng in king, five sounds specifically distinct, and five sounds perfectly simple require corresponding signs, which they have not.

II. Its inconsistency.—The f in fan, and the v in van sounds in a certain degree of relationship to p and b, are expressed by signs as unlike as f is unlike p, and as v is unlike b. The sound of the th in thin, the th in thine, the sh in shine, similarly related to t, d, and s, are expressed by signs as like t, d, and s, respectively, as th and sh.

The compound sibilant sound of j in jest is spelt with the single sign j, whilst the compound sibilant sound in chest is spelt with the combination ch.

III. Erroneousness.—The sound of the ee in feet is considered the long (independent) sound of the e in bed; whereas it is the long (independent) sound of the i in pit.

The i in bite is considered as the long (independent) sound of the i in pit; whereas it is a diphthongal sound.

The u in duck is looked upon as a modification of the u in bull; whereas it is a specifically distinct sound.

The ou in house and the oi in oil are looked upon as the compounds of o and i and of o and u respectively; whereas the latter element of them is not i and u, but y and w.

The th in thin and the th in thine are dealt with as one {185}and the same sound; whereas they are sounds specifically distinct.

The ch in chest is dealt with as a modification of c (either with the power of k or of s); whereas its elements are t and sh.

IV. Redundancy.—As far as the representation of sounds is concerned the letter c is superfluous. In words like citizen it may be replaced by s; in words like cat by k. In ch, as in chest, it has no proper place. In ch, as in mechanical, it may be replaced by k.

The diphthongal forms æ and œ, as in Æneas and Crœsus, except in the way of etymology, are superfluous and redundant.

V. Unsteadiness.—Here we have (amongst many other examples), 1. The consonant c with the double power of s and k; 2. g with its sound in gun and also with its sound in gin; 3. x with its sounds in Alexander, apoplexy, Xenophon.

In the foregoing examples a single sign has a double power; in the words Philip and filip, &c., a single sound has a double sign.

In respect to the degree wherein the English orthography is made subservient to etymology, it is sufficient to repeat the statement that the c, æ, and œ are retained in the alphabet for etymological purposes only.

The defects noticed in the preceding sections are absolute defects, and would exist, as they do at present, were there no language in the world except the English. This is not the case with those that are now about to be noticed; for them, indeed, the word defect is somewhat too strong a term. They may more properly be termed inconveniences.

Compared with the languages of the rest of the world the use of many letters in the English alphabet is singular. The letter i (when long or independent) is, with the exception of England, generally sounded as ee. With Englishmen it has a diphthongal power. The inconvenience of this is the necessity that it imposes upon us, in studying foreign languages, of {186}unlearning the sound which we give it in our own, and of learning the sound which it bears in the language studied. So it is (amongst many others) with the letter j. In English this has the sound of dzh, in French of zh, and in German of y. From singularity in the use of letters arises inconvenience in the study of foreign tongues.

In using j as dzh there is a second objection. It is not only inconvenient, but it is theoretically incorrect. The letter j was originally a modification of the vowel i. The Germans, who used it as the semivowel y, have perverted it from its original power less than the English have done, who sound it dzh.

I). Its convenience or inconvenience in respect to learning foreign tongues.—The sound given to the a in fate is singular. Other nations sound it as a in father.

The sound given to the e, long (or independent), is singular. Other nations sound it either as a in fate, or as é fermé.

The sound given to the i in bite is singular. Other nations sound it as ee in feet.

The sound given to the oo in fool is singular. Other nations sound it as the o in note, or as the ó chiuso.

The sound given to the u in duck is singular. Other nations sound it as the u in bull.

The sound given to the ou in house is singular. Other nations, more correctly, represent it by au or aw.

The sound given to the w in wet is somewhat singular, but is also correct and convenient. With many nations it is not found at all, whilst with those where it occurs it has the sound (there or thereabouts) of v.

The sound given to y is somewhat singular. In Danish it has a vowel power. In German the semivowel sound is spelt with j.

The sound given to z is not the sound which it has in German and Italian; but its power in English is convenient and correct.

The sound given to ch in chest is singular. In other languages it has generally a guttural sound; in French that of {187}sh. The English usage is more correct than the French, but less correct than the German.

II.) The historical propriety of certain letters.—The use of i with a diphthongal power is not only singular and inconvenient, but also historically incorrect. The Greek iota, from whence it originates, has the sound of i and ee, as in pit and feet.

The y, sounded as in yet, is historically incorrect. It grew out of the Greek υ, a vowel, and no semivowel. The Danes still use it as such, that is, with the power of the German ü.

The use of c for k in words derived from the Greek, as mechanical, ascetic, &c., is historically incorrect. The form c is the representative of γ and σ and not of the Greek kappa.

In remodelling alphabets the question of historical propriety should be recognized. Other reasons for the use of a particular letter in a particular sense being equal, the historical propriety should decide the question. The above examples are illustrative, not exhaustive.

§ 253. On certain conventional modes of spelling.—In the Greek language the sounds of o in not and of o in note (although allied) are expressed by the unlike signs or letters ο and ω, respectively. In most other languages the difference between the sounds is considered too slight to require for its expression signs so distinct and dissimilar. In some languages the difference is neglected altogether. In many, however, it is expressed, and that by some modification of the original letter.

Let the sign (ˉ) denote that the vowel over which it stands is long, or independent, whilst the sign (˘) indicates shortness, or dependence. In such a case, instead of writing not and nωt, like the Greeks, we may write nŏt and nōt, the sign serving for a fresh letter. Herein the expression of the nature of the sound is natural, because the natural use of (ˉ) and (˘) is to express length or shortness, dependence or independence. Now, supposing the broad sound of o {188}to be already represented, it is very evident that, of the other two sounds of o, the one must be long (independent), and the other short (dependent); and as it is only necessary to express one of these conditions, we may, if we choose, use the sign (ˉ) alone; its presence denoting length, and its absence shortness (independence or dependence).

As signs of this kind, one mark is as good as another; and instead of (ˉ) we may, if we choose, substitute such a mark as (´) (and write nót=nōt=nωt=nōte); provided only that the sign (´) expresses no other condition or affection of a sound. This use of the mark (´), viz. as a sign that the vowel over which it is placed is long (independent), is common in many languages. But is this use of (´) natural? For a reason that the reader has anticipated, it is not natural, but conventional. It is used elsewhere not as the sign of quantity, but as the sign of accent; consequently, being placed over a letter, and being interpreted according to its natural meaning, it gives the idea, not that the syllable is long, but that it is emphatic or accented. Its use as a sign of quantity is an orthographical expedient, or a conventional mode of spelling.

The English language abounds in orthographical expedients; the mode of expressing the quantity of the vowels being particularly numerous. To begin with these:

The reduplication of a vowel where there is but one syllable (as in feet, cool), is an orthographical expedient. It merely means that the syllable is long (or independent).

The juxta-position of two different vowels, where there is but one syllable (as in plain, moan), is an orthographical expedient. It generally means the same as the reduplication of a vowel, i.e., that the syllable is long (independent).

The addition of the e mute, as in plane, whale (whatever may have been its origin), is, at present, but an orthographical expedient. It denotes the lengthening of the syllable.

The reduplication of the consonant after a vowel, as in spotted, torrent, is in most cases but an orthographical expedient. It merely denotes that the preceding vowel is short (dependent). {189}

The use of ph for f in Philip, is an orthographical expedient, founded upon etymological reasons.

The use of th for the simple sound of the first consonant in thin and thine, is an orthographical expedient. The combination must be dealt with as a single letter.

X, however, and q are not orthographical expedients. They are orthographical compendiums.

The above instances have been adduced as illustrations only. Further details will be found hereafter. For many of them we can give a reason (for instance, for the reduplication of a consonant to express the shortness of the preceding vowel), and of many of them we can give an historical account (see Chapter X.).

§ 254. The mischief of orthographical expedients is this:—When a sign, or letter, is used in a conventional, it precludes us from using it (at least without further explanation) in its natural sense: e.g., the double o in mood constitutes but one syllable. If in a foreign language we had, immediately succeeding each other, first the syllable mo, and next the syllable od, we should have to spell it mo-od, or möod or mo-ỏd, &c. Again, it is only by our knowledge of the language that the th in nuthook, is not pronounced like the th in burthen. In the languages of India the true sound of t + h is common. This, however, we cannot spell naturally because the combination th conveys to us another notion. Hence such combinations as thh, or t‛, &c., in writing Hindoo words.

A second mischief of orthographical conventionalities, is the wrong notions that they engender, the eye misleading the ear. That th is really t + h, no one would have believed had it not been for the spelling.

§ 255. The present section is the partial application of the preceding observations. It is a running commentary upon the orthographical part of Dr. Johnson's Grammar. Presuming a knowledge of the detail of the English orthography, it attempts an explanation of some of its leading characters. Many of these it possesses in common with other tongues. Several are peculiar to itself. {190}

"A, sounded as aw, or as a modification of o."—A, as in father, and o, as in note (as may be seen in p. 150), form the extremities of the vowel system. Notwithstanding this, the two sounds often interchange. The orthographical systems of most languages bear witness to this. In French the au in autel has the sound of o; in Danish aa=o (baade being pronounced bohde); in Swedish å has the same power. In Old English the forms hond, strond, &c., occur, instead of hand, strand, &c. In Anglo-Saxon, brád, stán, &c., correspond to the English forms broad, stone. I am not able to say whether a changes oftenest to o, or o to a. The form hond is older than the form hand. In the word salt, however, the a was pronounced as the a in fat before it was pronounced (as at present) like the o in not. If this were not the case it would never have been spelt with an a. In the words launch and haunch, by some called lanch, hanch, and by others lawnch, hawnch, we find a present tendency to interchange these sounds.

The change from a to o takes place most especially before the liquid l, wall, call, fall. When the liquid l is followed by another consonant, it (viz. l) is generally sunk in pronunciation, falcon, salmon, &c., pronounced faucon, sammon, or saumon. The reason of this lies in the following fact, viz., that syllables wherein there are, at the same time, two final consonants and a long vowel, have a tendency to become shortened by one of two processes, viz., either by ejecting one of the consonants, or by shortening the vowel. That the l in falcon is affected not by the change of a to o, but by the change of a short vowel to a long, or of a slender one to a broad one, is shown in the tendency which the common people have to say hode for hold, as well as by the Scotch form gowd for gold. This fact bears upon the difficult problem in the Greek (and in other languages), viz., whether the lengthening of the vowel in words like ὀδοὺς (compared with ὀδόντος), is the cause or the effect of the rejection of the consonant.

"E is long, as in scene; or short, as in cellar."'—Johnson. It has been stated before that the (so-called) long sound of e is non-existent, and the e in scene, is the (so-called) long sound of the i in pit. {191}

The power of e mute in words like cane, bane, tune, robe, pope, fire, cure, tube, has already been noticed. It serves to denote the length of the preceding vowel. For this purpose it is retained; but it was not for this purpose that it was invented. Originally it expressed a sound, and it is only by a change of language that it has come, as it were by accident, to be an orthographical expedient.

Let a word consist of two syllables. Let the latter end in a vowel. Let there be between the vowel of the first and the vowel of the second syllable, one consonant and no more, e. g., namæ. Let the consonant belong to the root of the word; and let the first syllable of the word be the essential and the radical part of it. Let this same syllable (as the essential and radical part of it) have an accent. The chances are that, under such circumstances, the vowel of the first syllable will be long (independent), just as the chances are that a vowel followed by two consonants will be short. Let a change in language affect the final vowel, so that a word which was originally pronounced nama, should become, first, namë, and afterwards nām, naim, or næm; the vowel being sounded as the a in fate. Let the final e, although lost in pronunciation, be retained in the spelling. The chances are that, the above conditions being given, such an e (final and mute) shall, whenever it occurs, occur at the end of a long syllable. The next process is for a succeeding generation to mistake a coincidence for a sign, and to imagine that an e mute expresses the length of syllable.

I consider this to be the key to the use of the e mute in all words where it is preceded by one consonant only.

From the circumstance that the French and the English are the only nations wherein the e mute is part and parcel of the orthography, it has been hastily imagined that the employment of it is to be attributed to the Norman Conquest. The truth, however, is, that we find it equally in words of Saxon and of Norman origin.

The fact that, in certain words, an e mute is preceded by {192}two consonants and by a short vowel, does not militate against the view given above.

"I has a sound, long, as in fine, and short, as in fin. That is eminently observable in i, which may be likewise remarked in other letters, that the short sound is not the long sound contracted, but a sound wholly different."—Johnson. This extract has been made in order to add the authority of Johnson to the statement so often repeated already; viz., that the i in bite is not the long sound of the i in bit.

As a vowel, y is wholly superfluous. It is a current remark that more words end in y (fortify, pretty) than in any other letter. This is true only in respect to their spelling. As a matter of speech, the y final has always the sound either of the ee in feet, or of the i in bite. Such is the case with the words fortify and pretty, quoted above. For some reason or other, the vowel e is never, in English, written at the end of words, unless when it is mute; whilst i is never written at all. Instead of cri, we write cry, &c. This is a peculiarity of our orthography, for which I have no satisfactory reason. It may be, that with words ending in e, y is written for the sake of showing that the vowel is not mute, but sounded. Again, the adjectives ending in y as any, and the adverbs in ly, as manly, in the older stages of our language ended, not in y, but in ig (manlig, ænig); so that the present y, in such words, may be less the equivalent of i than the compendium of ig. I venture this indication with no particular confidence.

The b in debtor, subtile, doubt, agrees with the b in lamb, limb, dumb, thumb, womb, in being mute. It differs, however, in another respect. The words debtor, subtle, doubt, are of classical, the words lamb, limb, dumb, &c., are of Saxon, origin. In debtor, &c., the b was, undoubtedly, at one time, pronounced, since it belonged to a different syllable; debitor, subtilis, dubito, being the original forms. I am far from being certain that with the other words, lamb, &c., this was the case. With them the b belonged (if it belonged to the word at all) to the same syllable as the m. I think, {193}however, that instead of this being the case, the b, in speech, never made a part of the word at all; that it belongs now, and that it always belonged, to the written language only; and that it was inserted in the spelling upon what may be called the principle of imitation. For a further illustration of this, see the remarks on the word could.

"Ch has a sound which is analysed into tsh, as church, chin, crutch. C might be omitted in the language without loss, since one of its sounds might be supplied by s, and the other by k, but that it preserves to the eye the etymology of words, as face from facies, captive from captivus"—Johnson.

Before a, o, u (that is, before a full vowel), c is sounded as k; before e, i, and y (that is, before a small vowel), it has the power of s. This change of sound according to the nature of the vowel following, is so far from being the peculiarity of the English, that it is common in all languages; except that sometimes c, instead of becoming s, becomes ts, tsh, ksh, in other words, some other sibilant; but always a sibilant. A reference to p. 153 will explain this change. At a certain time, k (written c, as is the case in Latin) becomes changed by the vowel following into ksh, and from thence into s, ts, or tsh. That the syllables cit, cyt, cet, were at one time pronounced kit, kyt, ket, we believe: 1. from the circumstance that if it were not so, they would have been spelt with an s; 2. from the comparison of the Greek and Latin languages, where the words cete, circus, cystis, Latin, are κητὴ, κίρκος, κύστις, Greek.

In the words mechanical, choler, &c., derived from the Greek, it must not be imagined that the c represents the Greek kappa or κ. The combination c + h is to be dealt with as a single letter. Thus it was that the Romans, who had in their language neither the sound of χ, nor the sign κ, rendered the Greek chi (χ), just as by th they rendered θ, and by ph, φ.

The faulty representation of the Greek χ has given rise to a faulty representation of the Greek κ, as in ascetic, from ἀσκήτικος.

"C, according to the English orthography, never ends a {194}word; therefore we write stick, block, which were originally sticke, blocke. In such words c is now mute."—Johnson. Just as there was a prejudice against i or e ending a word there seems to have been one in the case of c. In the word Frederick there are three modes of spelling: 1. Frederic; 2. Frederik; 3. Frederick. Of these three it is the last only that seems, to an Englishman, natural. The form Frederic seems exceptionable, because the last letter is c, whilst Frederik is objected to because k comes in immediate contact with the short vowel.

Now the reason against c ending a word seems this. From what has been remarked above, c seems, in and of itself, to have no power at all. Whether it shall be sounded as k or as s seems undetermined, except by the nature of the vowel following. If the vowel following be small, c=s, if full, c=k. But c followed by nothing is equivocal and ambiguous. Now c final is c followed by nothing; and therefore c equivocal, ambiguous, indefinite, undetermined. This is the reason why c is never final. Let there be such words as sticke and blocke. Let the k be taken away. The words remain stice, bloce. The k being taken away, there is a danger of calling them stise, blose.

A verbal exception being taken, the statement of Dr. Johnson, that in words like stick and block the c is mute, is objectionable. The mute letter is not so much the c as the k.

"G at the end of a word is always hard, as ring, sing."—Johnson. A verbal exception may be taken here. Ng, is not a combination of the sounds of n+g, but the representation of a simple single sound; so that, as in the case of th and sh, the two letters must be dealt with as a single one.

"G before n is mute, as gnash, sign, foreign."—Johnson. The three words quoted above are not in the same predicament. In words like gnash the g has been silently dropped on the score of euphony (see remarks on k); in sign and foreign the g has not been dropped, but changed. It has taken the allied sound of the semivowel y, and so, with the preceding vowel, constitutes a diphthong. {195}

Before a, o, u (full vowels), g has the sound, as in gay, go, gun: before e, i, y, that of gem, giant.

At the end of a word (that is, followed by nothing at all), or followed by a consonant, it has the same sound that it has before a, o, u—agog, grand. This shows that such is its natural sound. In hedge and oblige the e mute serves to show that the g is to be pronounced as j.

Let there be the word rŏg. Let the vowel be lengthened. Let this lengthening be expressed by the addition of e mute, roge. There is now a risk of the word being called roje. This is avoided by inserting u, as in prorogue. Why, however, is it that the u runs no chance of being pronounced, and the word of being sounded prorogwé? The reason for this lies in three facts. 1. The affinities between the sounds of ga and ka. 2. The fact that qu is merely kw. 3. The fact that in qu, followed by another vowel, as in quoit (pronounced koyt), antique, &c., the u is altogether omitted in pronunciation. In other words, the analogy of qu is extended to gu.

For the varied sounds of gh in plough, tough, enough (enow), through, we must remember that the original sound of gh was a hard guttural, as is at present the case in Scotland, and between g, h, f, v, w, there are frequent interchanges.

"H is a note of aspiration."—It is under the notion that th, ph, sh, as in thin, thine, Philip, shine, are aspirated sounds, that h is admitted in the spelling. As has been repeatedly stated, th, ph, sh are to be treated as single signs or letters.

"J, consonant, sounds uniformly like the soft g (i.e., as in gem), and is, therefore, a letter useless, except in etymology, as ejaculation, jester, jocund, juice."—Johnson. It may be added that it never occurs in words of Saxon origin, and that in the single word Allelujah it has the sound of y, as in the German.

K never comes before a, o, u, or before a consonant. It is used before e, i, y, where c would, according to the English analogy, be liable to be sounded as s; as in kept, king, skirt. These words, if written cept, cing, scirt, would run the risk of being sounded sept, sing, sirt. Broadly speaking, k is never {196}used except where c would be inconvenient. The reason of this lies in the fact of there being no such letter as k in the Latin language. Hence arose in the eyes of the etymologist the propriety of retaining, in all words derived from the Latin (crown, concave, concupiscence, &c.), the letter c, to the exclusion of k. Besides this, the Anglo-Saxon alphabet, being taken from the Roman, excluded k, so that c was written even before the small vowels, a, e, i, y; as cyning, or cining, a king. C then supplants k upon etymological grounds only. In the languages derived from the Latin this dislike to the use of k leads to several orthographical inconveniences. As the tendency of c, before e, i, y, to be sounded as s (or as a sound allied to s), is the same in those languages as in others; and as in those languages, as in others, there frequently occur such sounds as kit, ket, kin, &c., a difficulty arises as to the spelling. If spelt cit, cet, &c., there is the risk of their being sounded sit, set. To remedy this, an h is interposed—chit, chet, &c. This, however, only substitutes one difficulty for another, since ch is, in all probability, already used with a different sound, e.g., that of sh, as in French, or that of k guttural, as in German. The Spanish orthography is thus hampered. Unwilling to spell the word chimera (pronounced kimera) with a k; unable to spell it with either c or ch, it writes the word quimæra. This distaste for k is an orthographical prejudice. Even in the way of etymology it is but partially advantageous, since in the other Gothic languages, where the alphabet is less rigidly Latin, the words that in English are spelt with a c, are there written with k,—kam, German; komme, Danish; skrapa, Swedish;=came, come, scrape.

"Skeptic, for so it should be written, not sceptic."—Johnson. Quoted for the sake of adding authority to the statement made in p. 193, viz., that the Greek kappa is to be represented not by c, but by k.

"K is never doubled, but c is used before it to shorten the vowel by a double consonant, as cŏckle, pĭckle."—Johnson. {197}This is referable to the statement that k is never used where c is admissible.

"K is used before n, knell, knot, but totally loses its sound."—Johnson. This, however, is not the ease in the allied languages; in German and Danish, in words like knecht, knive, the k is sounded. This teaches us that such was once the case in English. Hence we learn that in the words knife, knight (and also in gnaw, gnash), we have an antiquated or obsolete orthography.

For the ejection of the sound of l in calf, salmon, falcon, &c. see under a. For the l in could, see that word.

"N is sometimes mute after m, as damn, condemn, hymn."—Johnson. In all these words the n originally belonged to a succeeding syllable, dam-no, condem-no, hym-nus.

Q, accurately speaking, is neither a letter, nor an abbreviation. It is always followed by u, as queen, quilt, and the two letters qu must be looked upon as a single sign, equivalent to (but scarcely an abbreviation) of kw. Q is not=k alone. The combination qu, is never sounded koo. Neither is kw. If it were so, there would be in the word queen (currently speaking) three sounds of u, viz., two belonging to q (=kw), and one belonging to u itself. W being considered as=2 u: q=k + ½ w. This view of q bears upon the theory of words like prorogue, &c.

The reader is referred to p. 152. There he is told that, when a word ends in a flat consonant, b, v, d, g, the plural termination is not the sound of s, but that of z (stagz, dogz); although s be the letter written. Such also is the case with words ending in the vowels or the liquids (peaz, beanz, hillz, not peace, beance, hillce). This fact influences our orthography. The majority of words ending in s are found to be plural numbers, or else (what is the same thing in respect to form) either genitive cases, or verbs of the third person singular; whilst in the majority of these the s is sounded as z. Hence, the inference from analogy that s single, at the end of words, is sounded as z. Now this fact hampers the orthography of those words wherein s final retains its natural sound, as since, once, mass, mace; for let these be {198}written sins, ons, mas, the chances are that they will be pronounced sinz, onz, maz. To remedy this, the s may be doubled, as in mass. This, however, can be done in a few cases only. It cannot be done conveniently where the vowel is long, the effect of a double consonant being to denote that the preceding vowel is short. Neither can it be done conveniently after a consonant, such combinations as sinss, &c., being unsightly. This throws the grammarian upon the use of c, which, as stated above, has, in certain situations, the power of s. To write, however, simply sinc, or onc, would induce the risk of the words being sounded sink, onk. To obviate this, e is added, which has the double effect of not requiring to be sounded (being mute), and of showing that the c has the sound of s (being small).

"It is the peculiar quality of s that it may be sounded before all consonants, except x and z, in which s is comprised, x being only ks, and z only a hard [flat] or gross s. This s is therefore termed by grammarians suæ potestatis litera, the reason of which the learned Dr. Clarke erroneously supposed to be, that in some words it might be doubled at pleasure."—Johnson. A reference to the current Greek Grammars will indicate another reason for σ being called suæ potestatis litera. It will there be seen that, whilst π, β, φ—κ, γ, χ—τ, δ, θ—are grouped together, as tenues, mediæ, and aspiratæ, and as inter se cognatæ, σ stands by itself; ζ its media (flat sound) being treated as a double letter, and sh, its so-called aspirate, being non-existent in the Greek language.

"Th has two sounds; the one soft [flat], as thus, whether; the other hard [sharp], as thing, think. The sound is soft [flat] in all words between two vowels, as father, whether; and between r and a vowel, as burthen."—Johnson. The reason of the latter statement lies in the fact of both the vowels and r being flat (see p. 152), and so exerting a flattening influence upon the sounds in contact with them.

In the substantives breath and cloth, the th is sharp (i.e., as th in thin); in the verbs breathe and clothe, the th is flat (i.e., {199}as th in thine).—A great number of substantives may be made verbs by changing the sound of their final consonant. However, with the words breathe and clothe, a second change has taken place, viz., the vowel has been lengthened. Now of these two changes, viz., the lengthening of the vowel, and the flattening of the consonant, which is the one represented by the e mute, in clothe and breathe, as compared with cloth and breath? I imagine the former. Hence an exception is taken to the following statement of Dr. Johnson:—"When it (th) is softened [flattened] at the end of a word, an e silent must be added, as breath, breathe, cloth, clothe."

The sounds of the s in sure, of the t in picture (when pronounced pictshure), and of the z in azure and glazier, are explained in p. 153.

The present chapter is intended not to exhaust the list, but to illustrate the character of those orthographical expedients which insufficient alphabets, changes in language, and the influences of etymology engender both in the English and in other tongues.

CHAPTER X.

§ 256. The preceding chapter has exhibited the theory of a full and perfect alphabet; it has shown how far the English alphabet falls short of such a standard; and, above all, it has exhibited the various conventional modes of spelling which the insufficiency of alphabets, combined with other causes, has engendered. The present chapter gives a history of our alphabet, whereby many of its defects are accounted for. These defects, it may be said, once for all, the English alphabet shares with those of the rest of the world; although, with the doubtful exception of the French, it possesses them in a higher degree than any.

With few, if any, exceptions, all the modes of writing in the world originate, directly or indirectly, from the Phœnician, Hebrew, or Semitic alphabet. This is easily accounted for when we call to mind,—1. The fact that the Greek, the Latin, and the Arabian alphabets, are all founded upon this; and, 2. The great influence of the nations speaking those three languages. The present sketch, however, is given only for the sake of accounting for defects.

§ 257. Phœnician, Hebrew, or Semitic Period.—At a certain period the alphabet of Palestine, Phœnicia, and the neighbouring languages of the Semitic tribes, consisted of twenty-two separate and distinct letters. For these see the Hebrew Grammars and the Phœnicia of Gesenius.

The chances are, that, let a language possess as few elementary articulate sounds as possible, an alphabet of only twenty-two letters will be insufficient. Now, in the particular case of the languages in point, the number of elementary sounds, as we infer from the present Arabic, was above the average. {201}It may safely be asserted, that the original Semitic alphabet was insufficient for even the Semitic languages.

It was, moreover, inconsistent: since sounds as like as those of teth and tau (mere variations of each other) were expressed by signs as unlike as ט‎ and ת‎; whilst sounds as unlike as those of beth with a point, and beth without a point (b and v), were expressed (if expressed at all) by signs as like as ב‎ and בּ‎.

In this state it was imported into Greece. Now, as it rarely happens that any two languages have precisely the same elementary articulate sounds, so it rarely happens that an alphabet can be transplanted from one tongue to another, and be found, at once, to coincide.

The Greeks had, in all probability, sounds which were wanting in Palestine and Phœnicia. In Palestine and Phœnicia it is certain that there were sounds wanting in Greece.

Of the twenty-two Phœnician letters the Greeks took but twenty-one. The eighteenth letter, tsadi, ץ‎, was never imported into Europe.

§ 258. Greek Period.—Compared with the Semitic, the Old Greek alphabet ran thus:—

Such the order and form of the Greek and Hebrew letters. Here it may be remarked, that, of each alphabet, it is only the modern forms that are compared; the likeness in the shape of the letters may be seen by comparing them in their {202}older stages. Of these the exhibition, in a work like the present, is inconvenient. They may, however, be studied in the work already referred to in the Phœnicia of Gesenius. The names of the letters are as follows:—

§ 259. The Asiatic alphabet of Phœnicia and Palestine is now adapted to the European language of Greece. The first change took place in the manner of writing. The Orientals wrote from right to left; the Greeks from left to right. Besides this, the following principles, applicable whenever the alphabet of one language is transferred to another, were recognised:—

1. Letters for which there was no use were left behind. This was the case, as seen above, with the eighteenth letter, tsadi.

2. Letters expressive of sounds for which there was no precise equivalent in Greek, were used with other powers. This was the case with letters 5, 8, 16, and probably with some others.

3. Letters of which the original sound, in the course of time, became changed, were allowed, as it were, to drop out of the alphabet. This was the case with 6 and 19.

4. For such simple single elementary articulate sounds as there was no sign or letter representant, new signs, or letters, were invented. This principle gave to the Greek alphabet the new signs φ, χ, υ, ω.

5. The new signs were not mere modifications of the older {203}ones (as was the case with פּ‎, פ‎, בּ‎, ב‎, &c. in Hebrew), but new, distinct, and independent letters.

In all this there was an improvement. The faults of the newer Greek alphabet consisted in the admission of the compendium ψ=ps, and the retention of the fifteenth letter (samech, xi), with the power of ks, it being also a compendium.

§ 260. The Italian or old Latin period.—That it was either from the original Phœnician, or from the old Greek, that the Italian alphabets were imported, we learn from the existence in them of the letters f and q, corresponding respectively to the sixth and nineteenth letters; these having, in the second stage of the Greek alphabet, been ejected.

§ 261. The first alphabet imported into Italy was the Etruscan. In this the β, δ, and ο were ejected, their sounds (as it is stated) not being found in the Etruscan language. Be it observed, that the sounds both of β and δ are flat. Just as in the Devonshire dialect the flat sounds (z, v, &c.) have the preponderance, so, in the Etruscan, does there seem to have been a preponderating quantity of the sharp sounds. This prepares us for a change, the effects whereof exist in almost all the alphabets of Europe. In Greek and Hebrew the third letter (gimel, gamma) had the power of the flat mute g, as in gun. In the Etruscan it had the power of k. In this use of the third letter the Romans followed the Etruscans: but, as they had also in their language the sound of g (as in gun), they used, up to the Second Punic War, the third letter (viz. c), to denote both sounds. In the Duillian column we have Macestratos, Carthacinienses.^[36] Afterwards, however, the separate sign (or letter) g was invented, being originally a mere modification of c. The place of g in the alphabet is involved in the history of z.

§ 262. The Roman alphabet had a double origin. For the first two centuries after the foundation of the city the alphabet used was the Etruscan, derived directly from the Greek, and from the old Greek. This accounts for the presence of f and q.

Afterwards, however, the Romans modified their alphabet by the alphabet of the Italian Greeks; these Italian Greeks using the late Greek alphabet. This accounts for the presence of v, originating in the Greek ypsilon.

In accommodating the Greek alphabet to their own language, the Latins recognised the following principles:—

I. The ejection of such letters as were not wanted. Thus it was that the seventh letter (zayn, zæta) was thrown out of the alphabet, and the new letter, g, put in its place. Subsequently, z was restored for the sake of spelling Greek words, but was placed at the end of the alphabet. Thus also it was, that thæta, kappa (c being equivalent to k), and the fifteenth letter, were ejected, while ψ and χ were never admitted. In after-times the fifteenth letter (now xi) was restored, for the same reason that z was restored, and, like z, was placed at the end of the alphabet.

II. The use of the imported letters with a new power. Hence the sixth letter took the sound, not of v or w, but of f; and the eighth of h.

Beyond this the Romans made but slight alterations. In ejecting kappa, thæta and chi, they did mischief. The same in changing the power of c. The representation of φ by ph, and of θ by th was highly erroneous. The retention of x and q was unnecessary. V and j, two letters whereby the alphabet was really enriched, were mere modifications of u and i respectively. Y also seems a modification of v.

Neither the Latin, Greek, nor Hebrew orthographies were much warped to etymological purposes.

It should be observed, that in the Latin the letters have no longer any names (like beth, bæta), except such as are derived from their powers (be, ce).

It may now be seen that with a language containing such sounds as the th in thin and thine, and the ch in the German auch, it is to their advantage to derive their alphabet from the Greek; whilst, with a language containing such sounds as h and v, it is to their advantage to derive it from the Latin.

It may also be seen, that, without due alterations and {205}additions, the alphabet of one country will not serve as the alphabet of another.

§ 263. The Mœso-Gothic alphabet.—In the third century the classical alphabets were applied to a Gothic language. I use the word alphabets because the Mœso-Gothic letters borrowed from both the Latin and the Greek. Their form and order may be seen in Hickes' Thesaurus and in Lye's Grammar. With the Greek they agree in the following particulars.

1. In the sound of the third letter being not that of κ (c), but of the g in gun.

3. In expressing the simple single sound of th by a simple single sign. This sign, however, has neither the shape nor alphabetical position of the Greek thæta.

With the Latin they agree, 1. in possessing letters equivalent to f, g, h, q, y.

The Mœso-Gothic alphabet seems to have been formed on eclectic principles, and on principles sufficiently bold. Neither was its application traversed by etymological views. I cannot trace its influence, except, perhaps, in the case of the Anglo-Saxon letters þ and ƿ, upon any other alphabet; nor does it seem to have been acted upon by any earlier Gothic alphabet.

§ 264. The Anglo-Saxon alphabet.—What sort of an alphabet the Gothic languages possess we know: what sort of alphabet they require, we can determine. For the following sounds (amongst others) current in the Gothic, either one or both of the classical languages are deficient in corresponding signs.

4. The flat sound of the same, or the probable sound of the h in þurh, leoht, &c., Anglo-Saxon.—A sign neither in Greek nor Latin. {206}

7. The ch in chest.—A sign neither in Greek nor Latin, unless we suppose that at the time when the Anglo-Saxon alphabet was formed, the Latin c in words like civitas had the power, which it has in the present Italian, of ch.

8. The j in jest.—A sign neither in Greek nor Latin, unless we admit the same supposition in respect to g, that has been indicated in respect to c.

9. The sound of the kj; in the Norwegian kjenner; viz., that (thereabouts) of ksh.—A sign neither in Latin nor Greek.

11. The sound of the German ü, Danish y.—No sign in Latin; probably one in Greek, viz., υ.

12. Signs for distinguishing the long and short vowels, as ε and η, ο and ω.—Wanting in Latin, but existing in Greek.

In all these points the classical alphabets (one or both) were deficient. To make up for their insufficiency one of two things was necessary, either to coin new letters, or to use conventional combinations of the old.

In the Anglo-Saxon alphabet (derived from the Latin) we have the following features:—

2. The absence of the letter j, either with the power of y, as in German, of zh, as in French, or of dzh, as in English.

9. The presence of the simple single signs þ and ð, for the th in thin, and the th in thine.

Of the Anglo-Saxon alphabet we may safely say that it was insufficient. The points wherein the Latin alphabet was {207}improved in its adaptation to the Gothic tongues, are, 1. the admission of þ and ð; 2. the evolution of w out of u. Upon this latter circumstance, and on k and z, I make the following extract from the Latin Dedication of Otfrid's Krist:—"Hujus enim linguæ barbaries, ut est inculta et indisciplinabilis, atque insueta capi regulari freno grammaticæ artis, sic etiam in multis dictis scriptu est difficilis propter literarum aut congeriem, aut incognitam sonoritatem. Nam interdum tria u u u ut puto quærit in sono; priores duo consonantes, ut mihi videtur, tertium vocali sono manente," And, further, in respect to other orthographical difficulties:—"Interdum vero nec a, nec e, nec i, nec u, vocalium sonos præcanere potui, ibi y Grecum mihi videbatur ascribi. Et etiam hoc elementum lingua hæc horrescit interdum; nulli se characteri aliquotiens in quodam sono nisi difficile jungens. K et z sæpius hæc lingua extra usum Latinitatis utitur; quæ grammatici inter litteras dicunt esse superfluas. Ob stridorem autem dentium interdum ut puto in hac lingua z utuntur, k autem propter faucium sonoritatem."

§ 265. The Anglo-Norman Period.—Between the Latin alphabet, as applied to the Anglo-Saxon, and the Latin alphabet, as applied to the Norman-French, there are certain points of difference. In the first place, the sound-system of the languages (like the French) derived from the Latin, bore a greater resemblance to that of the Romans, than was to be found amongst the Gothic tongues. Secondly, the alphabets of the languages in point were more exclusively Latin. In the present French, Italian, Spanish, and Portuguese, there is an exclusion of the k. This is not the case with the Anglo-Norman. Like the Latins, the Anglo-Normans considered that the sound of the Greek θ was represented by th: not, however, having this sound in their language, there was no corresponding sign in their alphabet. The greatest mischief done by the Norman influence was the ejection from the English alphabet of þ and ð. In other respects the alphabet was improved. The letters z, k, j, were either imported or more currently recognised. The letter y took a semi-vowel power, having been previously represented by e; {208}itself having the power of i. The mode of spelling the compound sibilant with ch was evolved. My notions concerning this mode of spelling are as follows:—At a given period the sound of ce in ceaster, originally that of ke, had become, first, that of ksh, and, secondly, that of tsh; still it was spelt ce, the e, in the eyes of the Anglo-Saxons, having the power of y. In the eyes also of the Anglo-Saxons the compound sound of ksh, or tsh, would differ from that of k by the addition of y: this, it may be said, was the Anglo-Saxon view of the matter. The Anglo-Norman view was different. Modified by the part that, in the combination th, was played by the aspirate h, it was conceived by the Anglo-Normans, that ksh, or tsh, differed from k, not by the addition of y (expressed by e), but by that of h. Hence the combination ch as sounded in chest. The same was the case with sh. This latter statement is a point in the history, not so much of an alphabet, as of an orthography.

The preceding sketch, as has been said more than once before, has been given with one view only, viz., that of accounting for defective modes of spelling. The history of almost all alphabets is the same. Originally either insufficient, erroneous, or inconsistent, they are transplanted from one language to a different, due alterations and additions rarely being made.

§ 266. The reduplication of the consonant following, to express the shortness (dependence) of the preceding vowel, is as old as the classical languages: terra, θάλασσα. The following extract from the Ormulum (written in the thirteenth century) is the fullest recognition of the practice that I have met with. The extract is from Thorpe's Analecta Anglo-Saxonica.

Concerning the various other orthographical expedients, such as the reduplication of the vowel to express its length (mood), &c., I can give no satisfactory detailed history. The influence of the Anglo-Norman, a language derived from the Latin, established, in its fullest force, the recognition of the etymological principle.

§ 267. "I cannot trace the influence of the Mœso-Gothic alphabet, except, perhaps, in the case of the Anglo-Saxon letters þ and ƿ, upon any other alphabet; nor does it seem to have been itself acted upon by any earlier Gothic alphabet." (See p. 205.) The reason for the remark in Italics was as follows: In the Icelandic language the word run signifies a letter, and the word runa a furrow, or line. It has also some secondary meanings, which it is unnecessary to give in detail. Upon a vast number of inscriptions, some upon rocks, some upon stones of a defined shape, we find an alphabet different (at least, apparently so) from that of the Greeks, Latins, and Hebrews, and also unlike that of any modern nation. In this alphabet there is a marked deficiency of curved or rounded lines, and an exclusive preponderance of straight ones. As it was engraved rather than written, this is what we naturally expect. These letters are called Runes, and the alphabet which they constitute is called the Runic alphabet. Sometimes, by an extension of meaning, the Old Norse language, wherein they most frequently occur, is called the Runic language. This is as incorrect as to call a language an alphabetic language. To say, however, the Runic stage of a language is neither inaccurate nor inconvenient. The Runic alphabet, whether borrowed or invented by the early Goths, is of greater antiquity {210}than either the oldest Teutonic or the Mœso-Gothic alphabets. The forms, names, and order of the letters may be seen in Hickes' Thesaurus, in Olai Wormii Literatura Runica, in Rask's Icelandic Grammar, and in W. Grimm's Deutsche Runer.

The original number of the Runic letters is sixteen; expressing the sounds of f, u, þ, o, r, k, h, n, a, i, s, t, b, l, m, y. To these are added four spurious Runes, denoting c, x, æ, ö, and eight pointed Runes after the fashion of the pointed letters in Hebrew. In all this we see the influence of the imported alphabet upon the original Runes, rather than that of the original Runes upon the imported alphabet. It should, however, be remarked, that in the Runic alphabet the sound of th in thin is expressed by a simple sign, and that by a sign not unlike the Anglo-Saxon þ.

§ 268. The Order of the Alphabet.—In the history of our alphabet, we have had the history of the changes in the arrangement, as well as of the changes in the number and power of its letters. The following question now presents itself: viz., Is there in the order of the letters any natural arrangement, or is the original as well as the present succession of letters arbitrary and accidental? In the year 1835 I conceived, that in the order of the Hebrew alphabet I had discovered a very artificial arrangement. I also imagined that this artificial arrangement had been detected by no one besides myself. Two years afterwards a friend^[38] stated to me that he had made a similar observation, and in 1839 appeared, in Mr. Donaldson's New Cratylus, the quotation with which the present section will be concluded. The three views in the main coincide; and, as each has been formed independently (Mr. Donaldson's being the first recorded), they give the satisfactory result of three separate investigations coinciding in a theory essentially the same. The order of the Hebrew alphabet is as follows:—

Let beth, vaw, and pe (b, v, p) constitute a series called series P. Let gimel, kheth, and koph (g, kh, k`) constitute a series called series K. Let daleth, teth, and tau (d, t`, t) constitute a series called series T. Let aleph, he, and ayn constitute a series called the vowel series. Let the first four letters be taken in their order.

Herein the consonant of series B comes next to the letter of the vowel series; that of series K follows; and, in the last place, comes the letter of series D. After this the order changes: daleth being followed by he of the vowel series.

In this second sequence the relative positions of v, kh, and t` are the same in respect to each other, and the same in respect to the vowel series. The sequence itself is broken by the letter zayn, but it is remarkable that the principle of the sequence is the same. Series P follows the vowel, and series T is farthest from it. After this the system becomes but fragmentary. Still, even now, pe, of series P, follows ayn; tau, of {212}series D, is farthest from it; and koph, of series K, is intermediate. I am satisfied that we have in the Hebrew alphabet, and in all alphabets derived from it (consequently in the English), if not a system, the rudiments of a system, and that the system is of the sort indicated above; in other words, that the order of the alphabet is a circulating order.

In Mr. Donaldson's hands this view is not only a fact, but an instrument of criticism:—"The fact is, in our opinion, the original Semitic alphabet contained only sixteen letters. This appears from the organic arrangement of their characters. The remaining sixteen letters appear in the following order:—aleph, beth, gimel, daleth, he, vaw, kheth, teth, lamed, mem, nun, samech, ayn, pe, koph, tau. If we examine this order more minutely, we shall see that it is not arbitrary or accidental, but strictly organic, according to the Semitic articulation. We have four classes, each consisting of four letters: the first and second classes consist each of three mutes, preceded by a breathing; the third of the three liquids and the sibilant, which, perhaps, closed the oldest alphabet of all; and the fourth contains the three supernumerary mutes, preceded by a breathing. We place the characters first vertically:—

In the horizontal arrangement we shall, for the sake of greater simplicity, omit the liquids and the sibilant, and then we have {213}

In this we see, that, while the horizontal lines give us the arrangement of the mutes according to the breathings, the vertical columns exhibit them arranged according to the organ by which they are produced. Such a classification is obviously artificial."

§ 269. Parallel and equivalent orthographies.—Let there be in two given languages the sound of k, as in kin. Let each of these languages represent it by the same letter, k. In this case, the two orthographies are identical. Let, however, one nation represent it by k, and another by c. In this case the orthographies are not identical, but parallel. The same is the case with combinations. Let one nation (say the Anglo-Saxon) represent the sound of y (in ye) by e, whilst another nation (the Norse) represents it by j. What the Anglo-Saxon spells ceaster, the Northman spells kjaster; and what the Northman spells kjære, the Anglo-Saxon spells ceære. Let the sound of this ce and kj undergo a change, and become ksh; kjære and ceære, being pronounced kshære. The view of the Northman and Anglo-Saxon will be the same; each will consider that the compound sound differs from the simple one by the addition of the sound of y; that sound being expressed in one nation by e, and in the other by j. In this case the two expressions of the compound sound are parallel, its elements being considered the same, although the signs by which those elements are expressed are different.

Let, however, a different view of the compound sound be taken. Let it be thought that the sound of ksh differs from that of k, not by the addition of the sound of y, but by that of h; and so let it be spelt kh or ch. In this case the orthographies kh and kj (or ce) are not parallel, but equivalent. They express the same sound, but they do not denote the same elements. The same sound is, very possibly, expressed by the Anglo-Saxon ce, the Norwegian kj, and the English ch. In this case ce and kj are parallel, ce and ch equivalent, orthographies.

PART IV.

CHAPTER I.

§ 270. The word etymology, derived from the Greek, in the current language of scholars and grammarians, has a double meaning. At times it is used in a wide, and at times in a restricted, sense. What follows is an exhibition of the province or department of etymology.

If in the English language we take such a word as fathers, we are enabled to divide it into two parts; in other words, to reduce it into two elements. By comparing it with the word father, we see that the s is neither part nor parcel of the original word. The word fathers is a word capable of being analysed; father being the original primitive word, and s the secondary superadded termination. From the word father, the word fathers is derived, or (changing the expression) deduced, or descended. What has been said of the word fathers may also be said of fatherly, fatherlike, fatherless, &c. Now, from the word father, all these words (fathers, fatherly, fatherlike and fatherless) differ in form, and (not, however, necessarily) in meaning. To become such a word as fathers, &c., the word father is changed. Of changes of this sort, it is the province of etymology to take cognizance.

Compared with the form fathers, the word father is the older form of the two. The word father is a word current in this the nineteenth century. The same word was current in {215}the first century, although under a different form, and in a different language. Thus, in the Latin language, the form was pater; and earlier still, there is the Sanskrit form pitr. Now, just as the word father, compared with fathers, is original and primitive, so is pater, compared with father, original and primitive. The difference is, that in respect to father and fathers, the change that takes place, takes place within the same language, whilst the change that takes place between pater and father takes place within different languages. Of changes of this latter kind it is the province of etymology to take cognizance.

In its widest signification, etymology takes cognizance of the changes of the form of words. However, as the etymology that compares the forms fathers and father is different from the etymology that compares father and pater, we have, of etymology, two sorts: one dealing with the changes of form that words undergo in one and the same language (father, fathers), the other dealing with the changes that words undergo in passing from one language to another (pater, father).

The first of these sorts may be called etymology in the limited sense of the word, or the etymology of the grammarian. In this case it is opposed to orthoepy, orthography, syntax, and the other parts of grammar. This is the etymology of the ensuing pages.

The second may be called etymology in the wide sense of the word, historical etymology, or comparative etymology.

It must be again repeated that the two sorts of etymology agree in one point, viz., in taking cognizance of the changes of form that words undergo. Whether the change arise from grammatical reasons, as father, fathers, or from a change of language taking place in the lapse of time, as pater, father, is a matter of indifference.

In the Latin pater, and in the English father, we have one of two things, either two words descended or derived from each other, or two words descended or derived from a common original source.

In fatherlike we have a compound word capable of being analysed into the two primitive words, 1. father; 2. like.

With these preliminaries we may appreciate (or criticise) Dr. Johnson's explanation of the word etymology.

"Etymology, n. s. (etymologia, Lat.) ἔτυμος (etymos) true, and λόγος (logos) a word.

"1. The descent or derivation of a word from its original; the deduction of formations from the radical word; the analysis of compounds into primitives.

CHAPTER II.

§ 271. The nature of gender is best exhibited by reference to those languages wherein the distinction of gender is most conspicuous. Such a language, amongst others, is the Latin.

How far is there such a thing as gender in the English language? This depends upon the meaning that we attach to the word gender.

In the Latin language, where there are confessedly genders, we have the words taurus, meaning a bull, and vacca, meaning a cow. Here the natural distinction of sex is expressed by wholly different words. With this we have corresponding modes of expression in English: e.g.,

The mode, however, of expressing different sexes by wholly different words is not a matter of gender. The words boy and girl bear no etymological relation to each other; neither being derived from the other, nor in any way connected with it.

§ 272. Neither are words like cock-sparrow, man-servant, he-goat, &c., as compared with hen-sparrow, maid-servant, she-goat, &c., specimens of gender. Here a difference of sex is indicated by the addition of a fresh term, from which is formed a compound word.

§ 273. In the Latin words genitrix=a mother, and genitor=a father, we have a nearer approach to gender. Here the difference of sex is expressed by a difference of termination; {218}the words genitor and genitrix being in a true etymological relation, i. e., either derived from each other, or from some common source. With this we have, in English corresponding modes of expression: e. g.,

This, however, in strict grammatical language, is an approach to gender rather than gender itself. Its difference from true grammatical gender is as follows:—

The syllables in italics are the signs of the cases and numbers. Now these signs are the same in each word, the difference of meaning (or sex) not affecting them.

§ 274. Contrast, however, with the words genitor and genitrix the words domina=a mistress, and dominus=a master.

Here the letters in italics, or the signs of the cases and numbers, are different, the difference being brought about by the difference of gender. Now it is very evident that, if genitrix be a specimen of gender, domina is something more.

As terms, to be useful, must be limited, it may be laid down, as a sort of definition, that there is no gender where there is no affection of the declension: consequently, that, although we have, in English, words corresponding to genitrix and genitor, we have no true genders until we find words corresponding to dominus and domina.

§ 275. The second element in the notion of gender, although I will not venture to call it an essential one, is the following:—In the words domina and dominus, mistress and master, there is a natural distinction of sex; the one being masculine, or male, the other feminine, or female. In the words sword and lance there is no natural distinction of sex. Notwithstanding this, the word hasta, in Latin, is as much a feminine gender as domina, whilst gladius=a sword is, like dominus, a masculine noun. From this we see that, in languages wherein there are true genders, a fictitious or conventional sex is attributed even to inanimate objects. Sex is a natural distinction, gender a grammatical one.

§ 276. "Although we have, in English, words corresponding to genitrix and genitor, we have no true genders until we find words corresponding to dominus and domina."—The sentence was intentionally worded with caution. Words like dominus and domina, that is, words where the declension is affected by the sex, are to be found.

The pronoun him, from the Anglo-Saxon and English he, as compared with the pronoun her, from the Anglo-Saxon heò, is affected in its declension by the difference of sex, and is a true, though fragmentary, specimen of gender: for be it observed, that as both words are in the same case and number, the difference in form must be referred to a difference of sex expressed by gender. The same is the case with the form his as compared with her.

The pronoun it (originally hit), as compared with he, is a specimen of gender. {220}

The relative what, as compared with the masculine who, is a specimen of gender.

The forms it (for hit) and he are as much genders as hic and hæc, and the forms hic and hæc are as much genders as dominus and domina.

§ 277. The formation of the neuter gender by the addition of -t, in words like wha-t, i-t, and tha-t, occurs in other Indo-European languages. The -t in tha-t is the -d in istu-d, Latin, and the -t in ta-t, Sanskrit. Except, however, in the Gothic tongues, the inflection -t is confined to the pronouns. In the Gothic this is not the case. Throughout all those languages where there is a neuter form for adjectives at all, that form is either -t, or a sound derived from it:—Mœso-Gothic, blind-ata; Old High German, plint-ez; Icelandic, blind-t; German, blind-es=blind, cæc-um.—See Bopp's Comparative Grammar, Eastwick and Wilson's translation, p. 171.

§ 278. Just as there are in English fragments of a gender modifying the declension, so are there, also, fragments of the second element of gender; viz., the attribution of sex to objects naturally destitute of it. The sun in his glory, the moon in her wane, are examples of this. A sailor calls his ship she. A husbandman, according to Mr. Cobbett, does the same with his plough and working implements:—"In speaking of a ship we say she and her. And you know that our country-folks in Hampshire call almost everything he or she. It is curious to observe that country labourers give the feminine appellation to those things only which are more closely identified with themselves, and by the qualities or conditions of which their own efforts, and their character as workmen, are affected. The mower calls his scythe a she, the ploughman calls his plough a she; but a prong, or a shovel, or a harrow, which passes promiscuously from hand to hand, and which is appropriated to no particular labourer, is called a he."—English Grammar, Letter V.

Now, although Mr. Cobbett's statements may account for a sailor calling his ship she, they will not account for the custom of giving to the sun a masculine, and to the moon a {221}feminine, pronoun, as is done in the expressions quoted at the head of this section; still less will it account for the circumstance of the Germans reversing the gender, and making the sun feminine, and the moon masculine.

Let there be a period in the history of a nation wherein the sun and moon are dealt with, not as inanimate masses of matter, but as animated divinities. Let there, in other words, be a period in the history of a nation wherein dead things are personified, and wherein there is a mythology. Let an object like the sun be deemed a male, and an object like the moon a female, deity.

The Germans say the sun in her glory; the moon in his wane. This difference between the usage of the two languages, like so many others, is explained by the influence of the classical languages upon the English.—"Mundilfori had two children; a son, Mâni (Moon), and a daughter, Sôl (Sun)."—Such is an extract (taken second-hand from Grimm, vol. iii. p. 349) out of an Icelandic mythological work, viz., the prose Edda. In the classical languages, however, Phœbus and Sol are masculine, and Luna and Diana feminine. Hence it is that, although in Anglo-Saxon and Old-Saxon the sun is feminine, it is in English masculine.

Philosophy, charity, &c., or the names of abstract qualities personified, take a conventional sex, and are feminine from their being feminine in Latin.

As in these words there is no change of form, the consideration of them is a point of rhetoric, rather than of etymology.

Upon phrases like Cock Robin, Robin Redbreast, Jenny Wren, expressive of sex, much information may be collected from Grimm's Deutsche Grammatik, vol. iii. p. 359.

§ 279. The remainder of this chapter is devoted to miscellaneous remarks upon the true and apparent genders of the English language.

1. With the false genders like baron, baroness, it is a general rule that the feminine form is derived from the masculine, and not the masculine from the feminine; as peer, peeress. The words widower, gander, and drake are exceptions. For {222}the word wizard, from witch, see the section on augmentative forms.

2. The termination -ess, in which so large a portion of our feminine substantives terminate, is not of Saxon but of classical origin, being derived from the termination -ix, genitrix.

3. The words shepherdess, huntress, and hostess are faulty; the radical part of the word being Germanic, and the secondary part classical: indeed, in strict English grammar, the termination -ess has no place at all. It is a classic, not a Gothic, element.

4. The termination -inn, so current in German, as the equivalent to -ess, and as a feminine affix (freund=a friend; freundinn=a female friend), is found only in one or two words in English.

Carlin means an old woman: Icelandic, kerling; Sw., käring; Dan. kælling. Root, carl.

Bruin=the bear, may be either a female form, as in Old High German përo=a he-bear, pirinn=a she-bear, or it may be the Norse form björn=a bear, male or female.

Words like margravine and landgravine prove nothing, being scarcely naturalised.

5. The termination -str, as in webster, songster, and baxter, was originally a feminine affix. Thus, in Anglo-Saxon,

The same is the case in the present Dutch of Holland: e.g., spookster=a female fortune-teller; bakster=a {223}baking-woman; waschster=a washerwoman. (Grimm, Deutsche Grammatik, iii. p. 339.) The word spinster still retains its original feminine force.

6. The words songstress and seamstress, besides being, as far as concerns the intermixture of languages, in the predicament of shepherdess, have, moreover, a double feminine termination; 1st. -str, of Germanic, 2nd. -ess, of classical, origin.

7. In the word heroine we have a Greek termination, just as -ix is a Latin, and -inn a German one. It must not, however, be considered as derived from hero, by any process of the English language, but be dealt with as a separate importation from the Greek language.

8. The form deaconess is not wholly unexceptionable; since the termination -ess is of Latin, the root deacon of Greek origin: this Greek origin being rendered all the more conspicuous by the spelling, deacon (from diaconos), as compared with the Latin decanus.

9. The circumstance of prince ending in the sound of s, works a change in the accent of the word. As s is the final letter, it is necessary, in forming the plural number, and the genitive case, to add, not the simple letter s, as in peers, priests, &c., but the syllable -es. This makes the plural number and genitive case the same as the feminine form. Hence the feminine form is accented princéss, while peéress, príestess, &c., carry the accent on the first syllable. Princéss is remarkable as being the only word in English where the accent lies on the subordinate syllable.

10. It is uncertain whether kit, as compared with cat, be a feminine form or a diminutive form; in other words, whether it mean a female cat or a young cat.—See the Chapter on the Diminutives.

11. Goose, gander.—One peculiarity in this pair of words has already been indicated. In the older forms of the word goose, such as χὴν, Greek; anser, Latin; gans, German, as well as in the derived form gander, we have the proofs that, originally, there belonged to the word the sound of the letter n. In the forms ὀδοὺς, ὀδόντος, Greek; dens, dentis, Latin; zahn, {224}German; tooth, English, we find the analogy that accounts for the ejection of the n, and the lengthening of the vowel preceding. With respect, however, to the d in gander, it is not easy to say whether it is inserted in one word or omitted in the other. Neither can we give the precise power of the -er. The following forms (taken from Grimm, iii. p. 341) occur in the different Gothic dialects. Gans, fem.; ganazzo, masc., Old High German—gôs, f.; gandra, m., Anglo-Saxon—gâs, Icelandic, f.; gaas, Danish, f.; gassi, Icelandic, m.; gasse, Danish, m.—ganser, ganserer, gansart, gänserich, gander, masculine forms in different New German dialects.

12. Observe, the form gänserich has a masculine termination. The word täuberich, in provincial New German, has the same form and the same power. It denotes a male dove; taube, in German, signifying a dove. In gänserich and täuberich, we find preserved the termination -rich (or -rik), with a masculine power. Of this termination we have a remnant, in English, preserved in the curious word drake. To duck the word drake has no etymological relation whatsoever. It is derived from a word with which it has but one letter in common; viz. the Latin anas=a duck. Of this the root is anat-, as seen in the genitive case anatis. In Old High German we find the form anetrekho=a drake; in provincial New High German there is enterich and äntrecht, from whence come the English and Low German form drake. (Grimm, Deutsche Grammatik, iii. p. 341.)

13. Peacock, peahen, bridegroom.—In these compounds, it is not the words pea and bride that are rendered masculine or feminine by the addition of cock, hen, and groom, but it is the words cock, hen, and groom that are modified by prefixing pea and bride. For an appreciation of this distinction, see the Chapter on Composition.

CHAPTER III.

§ 280. In the Greek language the word patær signifies a father, speaking of one, whilst patere signifies two fathers, speaking of a pair, and thirdly, pateres signifies fathers, speaking of any number beyond two. The three words, patær, patere, and pateres, are said to be in different numbers, the difference of meaning being expressed by a difference of form. These numbers have names. The number that speaks of one is the singular, the number that speaks of two is the dual (from the Latin word duo=two), and the number that speaks of more than two is the plural.

All languages have numbers, but all languages have not them to the same extent. The Hebrew has a dual, but it is restricted to nouns only (in Greek being extended to verbs). It has, moreover, this peculiarity; it applies, for the most part, only to things which are naturally double, as the two eyes, the two hands, &c. The Latin has no dual number at all, except the natural dual in the words ambo and duo.

§ 281. The question presents itself,—to what extent have we numbers in English? Like the Greek, Hebrew, and Latin, we have a singular and a plural. Like the Latin, and unlike the Greek and Hebrew, we have no dual.

§ Different from the question, to what degree have we numbers? is the question,—over what extent of our language have we numbers? This distinction has already been foreshadowed or indicated. The Greeks, who said typtô=I beat, typteton=ye two beat, typtomen=we beat, had a dual number for their verbs as well as their nouns; while the Hebrew dual was limited to the nouns only. In the Greek, then, the dual {226}number is spread over a greater extent of the language than in the Hebrew.

There is no dual in the present English. It has been seen, however, that in the Anglo-Saxon there was a dual. But the Anglo-Saxon dual, being restricted to the personal pronouns (wit=we two; git=ye two), was not co-extensive with the Greek dual.

In the present Danish and Swedish there is no dual. In the Old Norse and in the present Icelandic a dual number is to be found.

From this we learn that the dual number is one of those inflections that languages drop as they become modern.

The numbers, then, in the present English are two, the singular and the plural. Over what extent of language have we a plural? The Latins say, bonus pater=a good father; boni patres=good fathers. In the Latin, the adjective bonus changes its form with the change of number of the substantive that it accompanies. In English it is only the substantive that is changed. Hence we see that in the Latin language the numbers were extended to adjectives, whereas in English they are confined to the substantives and pronouns. Compared with the Anglo-Saxon, the present English is in the same relation as it is with the Latin. In the Anglo-Saxon there were plural forms for the adjectives.

For the forms selves and others, see the Syntax. For the present, it is sufficient to foreshadow a remark which will be made on the word self, viz. that whether it be a pronoun, a substantive, or an adjective, is a disputed point.

Words like wheat, pitch, gold, &c., where the idea is naturally singular; words like bellows, scissors, lungs, &c., where the idea is naturally plural; and words like deer, sheep, where the same form serves for the singular and plural, inasmuch as there takes place no change of form, are not under the province of etymology.

§ 282. The current rule is, that the plural number is formed from the singular by adding s, as father, fathers. {227}However, if the reader will revert to the Section upon the sharp and flat Mutes, where it is stated that mutes of different degrees of sharpness and flatness cannot come together in the same syllable, he will find occasion to take to the current rule a verbal exception. The letter added to the word father, making it fathers, is s to the eye only. To the ear it is z. The word sounds fatherz. If the s retained its sound, the spelling would be fatherce. In stags, lads, &c., the sound is stagz, ladz. The rule, then, for the formation of the English plurals, rigorously expressed, is as follows.—The plural is formed from the singular, by adding to words ending in a vowel, a liquid or flat mute, the flat lene sibilant (z); and to words ending in a sharp mute, the sharp lene sibilant (s): e.g. (the sound of the word being expressed), pea, peaz; tree, treez; day, dayz; hill, hillz; hen, henz; gig, gigz; trap, traps; pit, pits; stack, stacks. Upon the formation of the English plural some further remarks are necessary.

I. In the case of words ending in b, v, d, the th in thine=ð, or g, a change either of the final flat consonant, or of the sharp s affixed, was not a matter of choice, but of necessity; the combinations abs, avs, ads, aðs, ags, being unpronounceable. See the Section on the Law of Accommodation.

II. Whether the first of the two mutes should be accommodated to the second (aps, afs, ats, aþs, asks), or the second to the first (abz, avz, aðz, agz), is determined by the habit of the particular language in question; and, with a few apparent exceptions (mark the word apparent), it is the rule of the English language to accommodate the second sound to the first, and not vice versâ.

III. Such combinations as peas, trees, hills, hens, &c. (the s preserving its original power, and being sounded as if written peace, treece, hillce, hence), being pronounceable, the change from s to z, in words so ending, is not a matter determined by the necessity of the case, but by the habit of the English language.

IV. Although the vast majority of our plurals ends, not in s, but in z, the original addition was not z, but s. This we {228}infer from three facts: 1. From the spelling; 2. from the fact of the sound of z being either rare or non-existent in Anglo-Saxon; 3. from the sufficiency of the causes to bring about the change.

It may now be seen that some slight variations in the form of our plurals are either mere points of orthography, or else capable of being explained on very simple euphonic principles.

§ 283. Boxes, churches, judges, lashes, kisses, blazes, princes.—Here there is the addition, not of the mere letter s, but of the syllable -es. As s cannot be immediately added to s, the intervention of a vowel becomes necessary; and that all the words whose plural is formed in -es really end either in the sounds of s, or in the allied sounds of z, sh, or zh, may be seen by analysis; since x=ks, ch=tsh, and j or ge=dzh, whilst ce, in prince, is a mere point of orthography for s.

Monarchs, heresiarchs.—Here the ch equals not tsh, but k, so that there is no need of being told that they do not follow the analogy of church, &c.

Cargoes, echoes.—From cargo and echo, with the addition of e; an orthographical expedient for the sake of denoting the length of the vowel o.

Beauty, beauties; key, keys.—Like the word cargoes, &c., these forms are points, not of etymology, but of orthography.

§ 284. "A few apparent exceptions."—These words are taken from Observation II. in the present section. The apparent exceptions to the rule there laid down are the words loaf, wife, and a few others, whose plural is not sounded loafs, wifs (loafce, wifce), but loavz, wivz (written loaves, wives). Here it seems as if z had been added to the singular; and, contrary to rule, the final letter of the original word been accommodated to the z, instead of the z being accommodated to the final syllable of the word, and so becoming s. It is, however, very probable that instead of the plural form being changed, it is the singular that has been modified. In the Anglo-Saxon the f at the end of words (as in the present Swedish) had the power of v. In the allied language the words in point are spelt with the flat mute, as weib, laub, kalb, halb, stab, {229}German. The same is the case with leaf, leaves; calf, calves; half, halves; staff, staves; beef, beeves: this last word being Anglo-Norman.

Pence.—The peculiarity of this word consists in having a flat liquid followed by the sharp sibilant s (spelt ce), contrary to the rule given above. In the first place, it is a contracted form from pennies; in the second place, its sense is collective rather than plural; in the third place, the use of the sharp sibilant lene distinguishes it from lens, sounded lenz. That its sense is collective rather than plural (a distinction to which the reader's attention is directed), we learn from the word sixpence, which, compared with sixpences, is no plural, but a singular form.

Dice.—In respect to its form, peculiar for the reason that pence is peculiar. We find the sound of s after a vowel, where that of z is expected. This distinguishes dice for play, from dies (diez) for coining. Dice, perhaps, like pence, is collective rather than plural.

In geese, lice, and mice, we have, apparently, the same phenomenon as in dice, viz., a sharp sibilant (s) where a flat one (z) is expected. The s, however, in these words is not the sign of the plural, but the last letter of the original word.

Alms.—This is no true plural form. The s belongs to the original word, Anglo-Saxon, ælmesse; Greek, ἐλεημοσύνη; just as the s in goose does. How far the word, although a true singular in its form, may have a collective signification, and require its verb to be plural, is a point not of etymology, but of syntax. The same is the case with the word riches, from the French richesse. In riches the last syllable being sounded as ez, increases its liability to pass for a plural.

News, means, pains.—These, the reverse of alms and riches, are true plural forms. How far, in sense, they are singular is a point not of etymology, but of syntax.

Mathematics, metaphysics, politics, ethics, optics, physics.—The following is an exhibition of my hypothesis respecting these words, to which I invite the reader's criticism. All the words in point are of Greek origin, and all are derived from a Greek adjective. Each is the name of some department of {230}study, of some art, or of some science. As the words are Greek, so also are the sciences which they denote, either of Greek origin, or else such as flourished in Greece. Let the arts and sciences of Greece be expressed, in Greek, rather by a substantive and an adjective combined, than by a simple substantive; for instance, let it be the habit of the language to say the musical art, rather than music. Let the Greek for art be a word in the feminine gender; e.g., τέχνη (tekhnæ), so that the musical art be ἡ μουσίκη τέχνη (hæ mousikæ tekhnæ). Let, in the progress of language (as was actually the case in Greece), the article and substantive be omitted, so that, for the musical art, or for music, there stand only the feminine adjective, μουσίκη. Let there be, upon a given art or science, a series of books, or treatises; the Greek for book, or treatise, being a neuter substantive, βίβλιον (biblion). Let the substantive meaning treatise be, in the course of language, omitted, so that whilst the science of physics is called φυσίκη (fysikæ), physic, from ἡ φυσίκη τέχνη, a series of treatises (or even chapters) upon the science shall be called φύσικα (fysika) or physics. Now all this was what happened in Greece. The science was denoted by a feminine adjective singular, as φυσίκη (fysicæ), and the treatises upon it, by the neuter adjective plural, as φύσικα (fysica). The treatises of Aristotle are generally so named. To apply this, I conceive, that in the middle ages a science of Greek origin might have its name drawn from two sources, viz., from the name of the art or science, or from the name of the books wherein it was treated. In the first case it had a singular form, as physic, logic; in the second place a plural form, as mathematics, metaphysics, optics.

In what number these words, having a collective sense, require their verbs to be, is a point of syntax.

In the first place it is a double plural: the -en being the -en in oxen, whilst the simpler form child-er occurs in the old English, and in certain provincial dialects.

In Icelandic, no plural termination is commoner than {231}that in -r; as geisl-ar=flashes, tung-ur=tongues, &c. Nevertheless, it is not the Icelandic that explains the plural form in question.

Besides the word childer, we collect from the other Gothic tongue the following forms in -r.—

and others, the peculiarity of which is the fact of their all being of the neuter gender. The particular Gothic dialect wherein they occur most frequently is the Dutch of Holland.

Now, the theory respecting the form so propounded by Grimm (D. G. iii. p. 270) is as follows:—

2. Which was, originally, no sign of a plural number, but merely a neuter derivative affix, common to the singular as well as to the plural number.

3. In this form it appears in the Mœso-Gothic: ag-is=fear (whence ague=shivering), hat-is=hate, rigv-is=smoke (reek). In none of these words is the -s radical, and in none is it limited to the singular number.

To these views Bopp adds, that the termination in question is the Sanskrit -as, a neuter affix; as in têj-as=splendour, strength, from tij=to sharpen.—V. G. pp. 141-259, Eastwick's and Wilson's translation.

To these doctrines of Grimm and Bopp, it should be added, that the reason why a singular derivational affix should become the sign of the plural number, lies, most probably, in the collective nature of the words in which it occurs: Husir=a collection of houses, eigir=a collection of eggs, eggery or eyry. For further observations on the power of -r, and for reasons for believing it to be the same as in the words Jew-r-y, yeoman-r-y, see a paper of Mr. Guest's, Philol. Trans., May 26, 1843. There we find the remarkable form lamb-r-en, from Wicliffe, Joh. xxi. Lamb-r-en : lamb :: child-r-en : child. {232}

§ 286. The form in -en.—In the Anglo-Saxon no termination of the plural number is more common than -n: tungan, tongues; steorran, stars. Of this termination we have evident remains in the words oxen, hosen, shoon, eyne, words more or less antiquated. This, perhaps, is no true plural. In welk-in=the clouds, the original singular form is lost.

§ 287. Men, feet, teeth, mice, lice, geese.—In these we have some of the oldest words in the language. If these were, to a certainty, true plurals, we should have an appearance somewhat corresponding to the weak and strong tenses of verbs; viz., one series of plurals formed by a change of the vowel, and another by the addition of the sibilant. The word kye, used in Scotland for cows, is of the same class. The list in Anglo-Saxon of words of this kind is different from that of the present English.

§ 288. Brethren.—Here there are two changes. 1. The alteration of the vowel. 2. The addition of -en. Mr. Guest quotes the forms brethre and brothre from the Old English. The sense is collective rather than plural.

Peasen=pulse.—As children is a double form of one sort (r + en), so is peasen a double form of another (s + en); pea, pea-s, pea-s-en. Wallis speaks to the singular power of the form in -s:—"Dicunt nonnulli a pease, pluraliter peasen; at melius, singulariter a pea, pluraliter pease:"—P. 77. He might have added, that, theoretically, pease was the proper singular form; as shown by the Latin pis-um.

If this were a plural form, it would be a very anomalous one. The -en, however, is no more a sign of the plural than is the -es in rich-es (richesse). The proper form is in -ain or -eyn.

Chickens.—A third variety of the double inflection (en + s), with the additional peculiarity of the form chicken being used, at present, almost exclusively in the singular number, although, originally, it was, probably, the plural of chick. So Wallis considered it:—"At olim etiam per -en vel -yn formabant pluralia: quorum pauca admodum adhuc retinemus. Ut, an ox, a chick, pluraliter oxen, chicken (sunt qui dicunt in singulari chicken, et in plurali chickens)."—(P. 77). Chick, chick-en, chick-en-s.

Fern.—According to Wallis the -n in fer-n is the -en in oxen, in other words, a plural termination:—"A fere (filix) pluraliter fern (verum nunc plerumque fern utroque numero dicitur, sed et in plurali ferns); nam fere et feres prope obsoleta sunt."—(P. 77.) Subject to this view, the word fer-n-s would exhibit the same phenomenon as the word chicke-n-s. It is doubtful, however, whether Wallis's view be correct. A reason for believing the -n to be radical is presented by the Anglo-Saxon form fearn, and the Old High German, varam.

Women.—Pronounced wimmen, as opposed to the singular form woomman. Probably an instance of accommodation.

Houses.—Pronounced houz-ez. The same peculiarity in the case of s and z, as occurs between f and v in words like life, lives, &c.

Paths, youths.—Pronounced padhz, yoodhz. The same peculiarity in the case of þ and ð, as occurs between s and z in the words house, houses. "Finita in f plerumque alleviantur in plurali numero, substituendo v; ut wife, wives, &c. Eademque alleviatio est etiam in s et th, quamvis retento charactere, in house, cloth, path."—P. 79.

CHAPTER IV.

§ 289. The extent to which there are, in the English language, cases, depends on the meaning which we attach to the word case. In the sentence a house of a father, the idea expressed by the words of a father, is an idea of relation between them and the word house. This idea is an idea of property or possession. The relation between the words father and house may be called the possessive relation. This relation, or connexion, between the two words is expressed by the preposition of.

In a fathers house the idea is, there or thereabouts, the same; the relation or connexion between the two words being the same. The expression, however, differs. In a father's house the relation, or connexion, is expressed, not by a preposition, but by a change of form, father becoming father's.

He gave the house to a father.—Here the words father and house stand in another sort of relationship; the relationship being expressed by the preposition to. The idea to a father differs from the idea of a father, in being expressed in one way only; viz., by the preposition. There is no second mode of expressing it by a change of form, as was done with father's.

The father taught the child.—Here there is neither preposition nor change of form. The connexion between the words father and child is expressed by the arrangement only.

Now if the relation alone between two words constitutes a case, the words or sentences, child; to a father; of a father; and father's, are all equally cases; of which one may be {235}called the accusative, another the dative, a third the genitive, and so on.

Perhaps, however, the relationship alone does not constitute a case. Perhaps there is a necessity of either the addition of a preposition (as in of a father), or of a change in form (as in father's). In this case (although child be not so) father's, of a father, and to a father, are all equally cases.

Now it is a remark, at least as old as Dr. Beattie,^[39] that if the use of a preposition constitute a case, there must be as many cases in a language as there are prepositions, and that "above a man, beneath a man, beyond a man, round about a man, within a man, without a man, shall be cases, as well as of a man, to a man, and with a man."

For etymological purposes it is necessary to limit the meaning of the word case; and, as a sort of definition, it may be laid down that where there is no change of form there is no case. With this remark, the English language may be compared with the Latin.

Here, since in the Latin language there are five changes of form, whilst in English there are but two, there are (as far, at least, as the word pater and father are concerned) three more cases in Latin than in English. It does not, however, follow that because in father we have but two cases, there may not be other words wherein there are more than two.

In order to constitute a case there must be a change of form.—This statement is a matter of definition. A second question, however, arises out of it; viz., whether every change of form constitute a case? In the Greek language there are the words ἔριν (erin), and ἔριδα (erida). Unlike the words father and father's these two words have precisely the same meaning. Each is called an accusative; and each, {236}consequently, is said to be in the same case with the other. This indicates the statement, that in order to constitute a case there must be not only a change of form, but also a change of meaning. Whether such a limitation of the word be convenient, is a question for the general grammarian. At present we merely state that there is no change of case unless there be a change of form. Hence, in respect to the word patribus (and others like it), which is sometimes translated from fathers, and at other times to fathers, we must say, not that in the one case the word is ablative and in the other dative, but that a certain case is used with a certain latitude of meaning. This remark bears on the word her in English. In her book the sense is that of the case currently called genitive. In it moved her, the sense is that of the case currently called the accusative. If we adhere, however, to what we have laid down, we must take exceptions to this mode of speaking. It is not that out of the single form her we can get two cases, but that a certain form has two powers; one that of the Latin genitive, and another that of the Latin accusative.

§ 290. This leads to an interesting question, viz., what notions are sufficiently allied to be expressed by the same form, and in the same case? The word her, in its two senses, may, perhaps, be dealt with as a single case, because the notions conveyed by the genitive and accusative are, perhaps, sufficiently allied to be expressed by the same word. Are the notions, however, of a mistress, and mistresses, so allied? I think not; and yet in the Latin language the same form, dominæ, expresses both. Of dominæ=of a mistress, and of dominæ=mistresses, we cannot say that there is one and the same case with a latitude of meaning. The words were, perhaps, once different. And this leads to the distinction between a real and an accidental identity of form.

In the language of the Anglo-Saxons the genitive cases of the words smith (smið), end (ende), and day (dæg), were, respectively, smithes (smiðes), endes, and dayes (dæges); whilst the nominative plurals were, respectively, smithas (smiðas), endas, and dayas (dægas). A process of change took place, by which the vowel of the last syllable in each {237}word was ejected. The result was, that the forms of the genitive singular and the nominative plural, originally different, became one and the same; so that the identity of the two cases is an accident.

This fact relieves the English grammarian from a difficulty. The nominative plural and the genitive singular are, in the present language of England, identical; the apostrophe in father's being a mere matter of orthography. However, there was once a difference. This modifies the previous statement, which may now stand thus:—for a change of case there must be a change of form existing or presumed.

§ 291. The number of our cases and the extent of language over which they spread.—In the English language there is undoubtedly a nominative case. This occurs in substantives, adjectives, and pronouns (father, good, he) equally. It is found in both numbers.

Accusative.—Some call this the objective case. The words him (singular) and them (plural) (whatever they may have been originally) are now true accusatives. The accusative case is found in pronouns only. Thee, me, us, and you are, to a certain extent, true accusatives.

They are accusative thus far: 1. They are not derived from any other case. 2. They are distinguished from the forms I, my, &c. 3. Their meaning is accusative. Nevertheless, they are only imperfect accusatives. They have no sign of case, and are distinguished by negative characters only.

One word of English is probably a true accusative in the strict sense of the term, viz., the word twain=two. The -n in twai-n is the -n in hine=him and hwone=whom. This we see from the following inflection:—

Although nominative as well as accusative, I have little doubt as to the original character of twégen being accusative. The {238}-n is by no means radical; besides which, it is the sign of an accusative case, and is not the sign of a nominative.

Note.—The words him and them are true accusatives in even a less degree than thee, me, us, and you. The Anglo-Saxon equivalents to the Latin words eos and illos were hi (or hig) and þá (or þæge); in other words, the sign of the accusative was other than the sound of -m. The case which really ended in -m was the so-called dative; so that the Anglo-Saxon forms him (or heom) and þám=the Latin iis and illis.

This fact explains the meaning of the words, whatever they may have been originally, in a preceding sentence. It also indicates a fresh element in the criticism and nomenclature of the grammarian; viz., the extent to which the history of a form regulates its position as an inflection.

Dative.—In the antiquated word whilom (at times), we have a remnant of the old dative in -m. The sense of the word is adverbial; its form, however, is that of a dative case.

Genitive.—Some call this the possessive case. It is found in substantives and pronouns (father's, his), but not in adjectives. It is formed like the nominative plural, by the addition of the lene sibilant (father, fathers; buck, bucks); or if the word end in s, by that of es (boxes, judges, &c.) It is found in both numbers: the men's hearts; the children's bread. In the plural number, however, it is rare; so rare, indeed, that wherever the plural ends in s (as it almost always does), there is no genitive. If it were not so, we should have such words as fatherses, foxeses, princesseses, &c.

Instrumental.—The following extracts from Rask's Anglo-Saxon Grammar, teach us that there exist in the present English two powers of the word spelt t-h-e, or of the so-called definite article.

"The demonstrative pronouns are þæt, se, seó (id, is, ea), which are also used for the article; and þis, þes, þeós (hoc, hic, hæc). They are thus declined:— {239}

"The indeclinable þe is often used instead of þæt, se, seo, in all cases, but especially with a relative signification, and, in later times, as an article. Hence the English article the.

"þy seems justly to be received as a proper ablativus instrumenti, as it occurs often in this character, even in the masculine gender; as, mid þy áþe=with that oath (Inæ Reges, 53). And in the same place in the dative, on þǽm áþe=in that oath."—Pp. 56, 57.

Hence the the that has originated out of the Anglo-Saxon þý is one word; the the that has originated out of the Anglo-Saxon þe, another. The latter is the common article: the former the the in expressions like all the more, all the better=more by all that, better by all that, and the Latin phrases eo majus, eo melius.

That why is in the same case with the instrumental the (=þy) may be seen from the following Anglo-Saxon inflection of the interrogative pronoun:—

Hence, then, in the and why we have instrumental ablatives, or, simply, instrumentals.

§ 292. The determination of cases.—How do we determine cases? In other words, why do we call him and them {240}accusatives rather than datives or genitives? By one of two means; viz., either by the sense or the form.

Suppose that in the English language there were ten thousand dative cases and as many accusatives. Suppose, also, that all the dative cases ended in -m, and all the accusatives in some other letter. It is very evident that, whatever might be the meaning of the words him and them their form would be dative. In this case the meaning being accusative, and the form dative, we should doubt which test to take.

My own opinion is, that it would be convenient to determine cases by the form of the word alone; so that, even if a word had a dative sense only once, where it had an accusative sense ten thousand times, such a word should be said to be in the dative case. Now, as stated above, the words him and them (to which we may add whom) were once dative cases; -m in Anglo-Saxon being the sign of the dative case. In the time of the Anglo-Saxons their sense coincided with their form. At present they are dative forms with an accusative meaning. Still, as the word give takes after it a dative case, we have, even now, in the sentence, give it him, give it them, remnants of the old dative sense. To say give it to him, to them, is unnecessary and pedantic: neither do I object to the expression, whom shall I give it? If ever the formal test become generally recognised and consistently adhered to, him, them, and whom will be called datives with a latitude of meaning; and then the only true and unequivocal accusatives in the English language will be the forms you, thee, us, me, and twain.

My, an accusative form (meh, me, mec), has now a genitive sense. The same may be said of thy.

Me, originally an accusative form (both me and my can grow out of mec and meh), had, even with the Anglo-Saxons, a dative sense. Give it me is correct English. The same may be said of thee.

Her.—For this word, as well as for further details on me and my, see the Chapters on the Personal and Demonstrative Pronouns. {241}

§ 293. When all traces of the original dative signification are effaced, and when all the dative cases in a language are similarly affected, an accusative case may be said to have originated out of a dative.

§ 294. Thus far the question has been concerning the immediate origin of cases: their remote origin is a different matter.

The word um occurs in Icelandic. In Danish and Swedish it is om; in the Germanic languages omme, umbi, umpi, ymbe, and also um. Its meaning is at, on, about. The word whilom is the substantive while=a time or pause (Dan. hvile=to rest), with the addition of the preposition om. That the particular dative form in om has arisen out of the noun plus the preposition is a safe assertion. I am not prepared, however, to account for the formation of all the cases in this manner.

§ 295. Analysis of cases.—In the word children's we are enabled to separate the word into three parts. 1. The root child. 2. The plural signs r and en. 3. The sign of the genitive case, s. In this case the word is said to be analysed, since we not only take it to pieces, but also give the respective powers of each of its elements; stating which denotes the case, and which the number. Although it is too much to say that the analysis of every case of every number can be thus effected, it ought always to be attempted.

§ 296. The true nature of the genitive form in s.—It is a common notion that the genitive form father's is contracted from father his. The expression in our liturgy, for Jesus Christ his sake, which is merely a pleonastic one, is the only foundation for this assertion. As the idea, however, is not only one of the commonest, but also one of the greatest errors in etymology, the following three statements are given for the sake of contradiction to it.

1. The expression the Queen's Majesty is not capable of being reduced to the Queen his Majesty.

2. In the form his itself, the s has precisely the power that it has in father's, &c. Now his cannot be said to arise out of he + his.

3. In all the languages of the vast Indo-European tribe, except the Celtic, the genitive ends in s, just as it does in {242}English; so that even if the words father his would account for the English word father's, it would not account for the Sanskrit genitive pad-as, of a foot; the Zend dughdhar-s, of a daughter; the Lithuanic dugter-s; the Greek ὀδόντ-ος; the Latin dent-is, &c.

For further remarks upon the English genitive, see the Cambridge Philological Museum, vol. ii. p. 246.

CHAPTER V.

§ 297. I, we, us, me, thou, ye.—These constitute the true personal pronouns. From he, she, and it, they differ in being destitute of gender.

These latter words are demonstrative rather than personal, so that there are in English true personal pronouns for the first two persons only.

In other languages the current pronouns of the third person are, as in English, demonstrative rather than personal.

The usual declension of the personal pronouns is exceptionable. I and me, thou and ye, stand in no etymological relations to each other. The true view of the words is, that they are not irregular but defective. I has no oblique, and me no nominative case. And so with respect to the rest.

I, in German ich, Icelandic ek, corresponds with ἐγὼ, and ego of the classical languages; ego and ἐγὼ being, like I, defective in the oblique cases.

My, as stated above, is a form originally accusative, but now used in a genitive sense.

Me.—In Anglo-Saxon this was called a dative form. The fact seems to be that both my and me grow out of an accusative form, meh, mec.

That the sound of k originally belonged to the pronouns me and thee, we learn not only from the Anglo-Saxons mec, þec, meh, þeh, but from the Icelandic mik, þik, and the German mich, dich. This accounts for the form my; since y=ey, and the sounds of y and g are allied. That both me and my can be evolved from mik, we see in the present Scandinavian languages, where, very often even in the same district, mig is pronounced both mey and mee. {244}

We and our.—These words are not in the condition of I and me. Although the fact be obscured, they are really in an etymological relation to each other. This we infer from the alliance between the sounds of w and ou, and from the Danish forms vi (we), vor (our). It may be doubted, however, whether our be a true genitive rather than an adjectival form. In the form ours we find it playing the part, not of a case, but of an independent word. Upon this, however, too much stress cannot be laid. In Danish it takes a neuter form: vor=noster; vort=nostrum. From this I conceive that it agrees, not with the Latin genitive nostrûm, but with the adjective noster.

Us, we, our.—Even us is in an etymological relation to we. That we and our are so, has just been shown. Now in Anglo-Saxon there were two forms of our, viz., úre (=nostrûm), and user (=noster). This connects we and us through our.

From these preliminary notices we have the changes in form of the true personal pronouns, as follows:—

§ 298. We and me have been dealt with as distinct words. But it is only for practical purposes that they can be considered to be thus separate; since the sounds of m and w are allied, and in Sanskrit the singular form ma=I is looked upon as part of the same word with vayam=we. The same is the case with the Greek με (me), and the plural form ἡμεῖς (hæmeis)=we.

You.—As far as the practice of the present mode of speech {245}is concerned, the word you is a nominative form; since we say you move, you are moving, you were speaking.

Why should it not be treated as such? There is no absolute reason why it should not. All that can be said is, that the historical reason and the logical reason are at variance. The Anglo-Saxon form for you was eow, for ye, ge. Neither bear any sign of case at all, so that, form for form, they are equally and indifferently nominative and accusative, as the habit of language may make them. Hence, it, perhaps, is more logical to say that a certain form (you) is used either as a nominative or accusative, than to say that the accusative case is used instead of a nominative. It is clear that you can be used instead of ye only so far as it is nominative in power.

Ye.—As far as the evidence of such expressions as get on with ye is concerned, the word ye is an accusative form. The reasons why it should or should not be treated as such are involved in the previous paragraph.

Me.—Carrying out the views just laid down, and admitting you to be a nominative, or quasi-nominative case, we may extend the reasoning to the word me, and call it also a secondary nominative; inasmuch as such phrases as it is me=it is I are common.

Now to call such expressions incorrect English is to assume the point. No one says that c'est moi is bad French, and that c'est je is good. The fact is, that the whole question is a question of degree. Has or has not the custom been sufficiently prevalent to have transferred the forms me, ye, and you from one case to another, as it is admitted to have done with the forms him and whom, once dative, but now accusative?

Observe.—That the expression it is me=it is I will not justify the use of it is him, it is her=it is he and it is she. Me, ye, you, are what may be called indifferent forms, i. e. nominative as much as accusative, and accusative as much as nominative. Him and her, on the other hand, are not indifferent. The -m and -r are respectively the signs of cases other than the nominative.

Again: the reasons which allow the form you to be {246}considered as a nominative plural, on the strength of its being used for ye, will not allow it to be considered a nominative singular on the strength of its being used for thou. It is submitted to the reader, that in phrases like you are speaking, &c., even when applied to a single individual, the idea is really plural; in other words, that the courtesy consists in treating one person as more than one, and addressing him as such, rather than in using a plural form in a singular sense. It is certain that, grammatically considered, you=thou is a plural, since the verb with which it agrees is plural:—you are speaking, not you art speaking.

CHAPTER VI.

ON THE TRUE REFLECTIVE PRONOUN IN THE GOTHIC LANGUAGES, AND ON ITS ABSENCE IN ENGLISH.

§ 299. A true reflective pronoun is wanting in English. In other words, there are no equivalents to the Latin pronominal forms sui, sibi, se.

Nor yet are there any equivalents in English to the so-called adjectival forms suus, sua, suum: since his and her are the equivalents to ejus and illius, and are not adjectives but genitive cases.

At the first view, this last sentence seems unnecessary. It might seem superfluous to state, that, if there were no such primitive form as se (or its equivalent), there could be no such secondary form as suus (or its equivalent).

Such, however, is not the case. Suus might exist in the language, and yet se be absent; in other words, the derivative form might have continued whilst the original one had become extinct.

Such is really the case with the Old Frisian. The reflective personal form, the equivalent to se, is lost, whilst the reflective possessive form, the equivalent to suus, is found. In the Modern Frisian, however, both forms are lost; as they also are in the present English.

In Mœso-Gothic.—Found in three cases, seina, sis, sik=sui, sibi, se.

In Old High German.—The dative form lost; there being no such word as sir=sis=sibi. Besides this, the genitive {248}or possessive form sin is used only in the masculine and neuter genders.

In Old Frisian.—As stated above, there is here no equivalent to se; whilst there is the form sin=suus.

In Old Saxon.—The equivalent to se, sibi, and sui very rare. The equivalent to suus not common, but commoner than in Anglo-Saxon.

In Anglo-Saxon.—No instance of the equivalent to se at all. The forms sinne=suum, and sinum=suo, occur in Beowulf. In Cædmon cases of sin=suus are more frequent. Still the usual form is his=ejus.

In the Dutch, Danish, and Swedish, the true reflectives, both personal and possessive, occur; so that the modern Frisian and English stand alone in respect to the entire absence of them.—Deutsche Grammatik, iv. 321-348.

The statement concerning the absence of the true reflective in English, although negative, has an important philological bearing on more points than one.

1. It renders the use of the word self much more necessary than it would be otherwise.

2. It renders us unable to draw a distinction between the meanings of the Latin words suus and ejus.

3. It precludes the possibility of the evolution of a middle voice like that of the Old Norse, where kalla-sc=kalla-sik.

CHAPTER VII.

§ 300. The demonstrative pronouns are, 1. He, it. 2. She. 3. This, that. 4. The.

He, she, and it, generally looked on as personal, are here treated as demonstrative pronouns, for the following reasons.

1. The personal pronouns form an extremely natural class, if the pronouns of the two first persons (and se when found in the language) be taken by themselves. This is not the case if they be taken along with he, it, and she. The absence of gender, the peculiarity in their declension, and their defectiveness are marked characters wherein they agree with each other, but not with any other words.

2. The idea expressed by he, it, and she is naturally that of demonstrativeness. In the Latin language is, ea, id; ille, illa, illud; hic, hæc, hoc, are demonstrative pronouns in sense, as well as in declension.

3. The plural forms they, them, in the present English, are the plural forms of the root of that, a true demonstrative pronoun; so that even if he, she, and it could be treated as personal pronouns, it could only be in their so-called singular number.

4. The word she has grown out of the Anglo-Saxon seó. Now seó was in Anglo-Saxon the feminine form of the definite article; the definite article being a demonstrative pronoun.

She.—The Anglo-Saxon form heó, being lost to the language, is replaced by the feminine article seó.

Her.—This is a case, not of the present she, but of the Anglo-Saxon heó: so that she may be said to be defective in {250}the oblique cases and her to be defective in the nominative.

Him.—A true dative form, which has replaced the Anglo-Saxon hine. When used as a dative, it was neuter as well as masculine.

His.—Originally neuter as well as masculine. Now as a neuter, replaced by its—"et quidem ipsa vox his, ut et interrogativum whose, nihil aliud sunt quam hee's, who's, ubi s omnino idem præstat quod in aliis possessivis. Similiter autem his pro hee's eodem errore quo nonnunquam bin pro been; item whose pro who's eodem errore quo done, gone, knowne, growne, &c., pro doen, goen, knowen, vel do'n, go'n, know'n, grow'n; utrobique contra analogiam linguæ; sed usu defenditur."—Wallis, c. v.

It.—Changed from the Anglo-Saxon hit, by the ejection of h. The t is no part of the original word, but a sign of the neuter gender, forming it regularly from he. The same neuter sign is preserved in the Latin id and illud.

Its.—In the course of time the nature of the neuter sign t, in it, the form being found in but a few words, became misunderstood. Instead of being looked on as an affix, it passed for part of the original word. Hence was formed from it the anomalous genitive its, superseding the Saxon his. The same was the case with—

Hers.—The r is no part of the original word, but the sign of the dative case. These formations are of value in the history of cases.

They, their, them.—When hit had been changed into it, when heó had been replaced by she, and when the single form the, as an article, had come to serve for all the cases of all the genders, two circumstances took place: 1. The forms þám and þára as definite articles became superfluous; and, 2. The connexion between the plural forms hí, heom, heora, and the singular forms he and it, grew indistinct. These were conditions favourable to the use of the forms they, them, and their, instead of hí, heom, heora.

Theirs.—In the same predicament with hers and its; either the case of an adjective, or a case formed from a case. {251}

Than or then, and there.—Although now adverbs, they were once demonstrative pronouns, in a certain case and in a certain gender.—Than and then masculine accusative and singular, there feminine dative and singular.

An exhibition of the Anglo-Saxon declension is the best explanation of the English. Be it observed, that the cases marked in italics are found in the present language.

Of this word we meet two forms only, both of the singular number, and both in the nominative case; viz. masc. se; fem. seó (the). The neuter gender and the other cases of the article were taken from the pronoun þæt (that).

1st. That the s is no inflection, but a radical part of the word, like the s in geese.

These facts create difficulties in respect to the word these. Mr. Guest's view is, perhaps, the best; viz. that the plural element of the word is the letter e, and that this -e is the old English and Anglo-Saxon adjective plural; so that thes-e is formed from thes, as gode (=boni) is formed from god (=bonus).

In Old English MSS. this plural in -e is general. It occurs not only in adjectives and pronouns as a regular inflection, but even as a plural of the genitive his, that word being treated as a nominative singular; so that hise is formed from his, as sui from suus, or as eji might have been formed from ejus; provided that in the Latin language this last word had been mistaken for a nominative singular. The following examples are Mr. Guest's.

§ 302. Those.—Perhaps the Anglo-Saxon þá with s added. Perhaps the þás from þis with its power altered. Rask, in his Anglo-Saxon Grammar, writes "from þis we find, in the plural, þæs for þás. From which afterwards, with a distinction in signification, these and those." The English form they is illustrated by the Anglo-Saxon form ðage=þá. The whole doctrine of the forms in question has yet to assume a satisfactory shape.

CHAPTER VIII.

§ 303. In the relative and interrogative pronouns, who, what, whom, whose, we have, expressed by a change of form, a neuter gender, what; a dative case, whom; and a genitive case, whose: the true power of the s (viz. as the sign of a case) being obscured by the orthographical addition of the e mute.

To these may be added, 1. the adverb why, originally the ablative form hvi (quo modo? quâ viâ?). 2. The adverb where, a feminine dative, like there. 3. When, a masculine accusative (in Anglo-Saxon hwæne), and analogous to then.

§ 304. The following points in the history of the demonstrative and relative pronouns are taken from Grimm's Deutsche Grammatik, vol. iii. pp. 1, 2, 3.

Throughout the Indo-European tribe the interrogative or relative idea is expressed by k, or by a modification of k; e.g., qu, hv, or h; as Sanskrit, kas, who; kataras, which of two; katama, which of many.—Lithuanic, kas, who; koks, of what sort; kokelys, how great; kaip, how.—Slavonic: kto, who, Russian and Polish; kdo, who, Bohemian; kotory, which, Russian; kolik, how great.—Quot, qualis, quantus, Latin.—Κόσος, κοῖος, κότε, Ionic Greek; in the other dialects, however, πότερος, πόσος, &c.—Gothic: hvas, who, Mœso-Gothic; huer, Old High German; hvaþar, which of two, Mœso-Gothic; huëdar, Old High German; hvem, hvad, huanne, huar, Norse; what, why, which, where, &c., English.

Throughout the Indo-European tribe the demonstrative idea is expressed by t, or by a modification of it; as, Sanskrit, tat, that; tata-ras, such a one out of two.—Lithuanic, tas, he; toks, such; tokelys, so great; taip, so.—Slavonic, t' or {256}ta, he; taku, such; tako, so.—Tot, talis, tantum, Latin.—Τόσος, τοῖος, τότε, Greek; this, that, thus, English, &c.

The two sounds in the Danish words hvi, hvad, &c., and the two sounds in the English, what, when (Anglo-Saxon, hwæt, hwæne), account for the forms why and how. In the first the w alone, in the second the h alone, is sounded. The Danish for why is hvi, pronounced vi; in Swedish the word is hu.

§ 305. The following remarks (some of them not strictly etymological) apply to a few of the remaining pronouns. For further details, see Grimm, D. G. iii. 4.

Same.—Wanting in Anglo-Saxon, where it was replaced by the word ylca, ylce. Probably derived from the Norse.

Self.—In myself, thyself, herself, ourselves, yourselves, a substantive (or with a substantival power), and preceded by a genitive case. In himself and themselves an adjective (or with an adjectival power), and preceded by an accusative case. Itself is equivocal, since we cannot say whether its elements are it and self, or its and self; the s having been dropped in utterance. It is very evident that either the form like himself, or the form like thyself, is exceptionable; in other words, that the use of the word is inconsistent. As this inconsistency is as old as the Anglo-Saxons, the history of the word gives us no elucidation. In favour of the forms like myself (self being a substantive), are the following facts:—

2. The Middle High German phrases, mîn lîp, dîn lîp, my body, thy body, equivalent in sense to myself, thyself.

3. The circumstance that if self be dealt with as a substantive, such phrases as my own self, his own great self, &c., can be used; whereby the language is a gainer.

"Vox self, pluraliter selves, quamvis etiam pronomen a quibusdam censeatur (quoniam ut plurimum per Latinum ipse redditur), est tamen plane nomen substantivum, cui quidem vix aliquod apud Latinos substantivum respondet; proxime tamen accedet vox persona vel propria persona, ut my self, thy self, our selves, your selves, &c. (ego ipse, tu ipse, nos ipsi, {257}vos ipsi, &c.), ad verbum mea persona, tua persona, &c. Fateor tamen himself, itself, themselves vulgo dici pro his-self, its-self, theirselves; at (interposito own) his own self, &c., ipsius propria persona, &c."—Wallis, c. vii.

Whit.—As in the phrase not a whit. This enters in the compound pronouns aught and naught.

One.—As in the phrase one does so and so. From the French on. Observe that this is from the Latin homo, in Old French hom, om. In the Germanic tongues man is used in the same sense: man sagt=one says=on dit. One, like self and other, is so far a substantive, that it is inflected. Gen. sing, one's own self: plural, my wife and little ones are well.

Derived pronouns.—Any, in Anglo-Saxon, ænig. In Old High German we have einîc=any, and einac=single. In Anglo-Saxon ânega means single. In Middle High German einec is always single. In New High German einig means, 1. a certain person (quidam), 2. agreeing; einzig, meaning single. In Dutch ênech has both meanings. This indicates the word án, one, as the root of the word in question.—Grimm, D. G. iii. 9.

Compound pronouns.—Which, as has been already stated more than once, is most incorrectly called the neuter of who. Instead of being a neuter, it is a compound word. The adjective leiks, like, is preserved in the Mœso-Gothic words galeiks, and missaleiks. In Old High German the form is lih, in Anglo-Saxon lic. Hence we have Mœso-Gothic, hvêleiks; Old High German, huëlih; Anglo-Saxon, huilic and hvilc; Old Frisian, hwelik; Danish, hvilk-en; German, welch; Scotch, whilk; English, which. (Grimm, D. G., iii. 47). The same is the case with—

1. Such.—Mœso-Gothic, svaleiks; Old High German, sôlih; Old Saxon, sulîc; Anglo-Saxon, svilc; German, solch; English, such. (Grimm, D. G. iii. 48). Rask's derivation of the Anglo-Saxon swilc from swa-ylc, is exceptionable.

2. Thilk.—An old English word, found in the provincial dialects, as thick, thuck, theck, and hastily derived by Tyrwhitt, {258}Ritson, and Weber, from së ylca, is found in the following forms: Mœso-Gothic, þêleiks; Norse, þvilikr. (Grimm, iii. 49.)

3. Ilk.—Found in the Scotch, and always preceded by the article; the ilk, or that ilk, meaning the same. In Anglo-Saxon this word is ylca, preceded also by the article se ylca, seó ylce, þæt ylce. In English, as seen above, the word is replaced by same. In no other Gothic dialect does it occur. According to Grimm, this is no simple word, but a compound one, of which some such word as ei is the first, and lîc the second element. (Deutsche Grammatik, iii. 50.)

Aught.—In Mœso-Gothic is found the particle aiv, ever, but only in negative propositions; ni (not) preceding it. Its Old High German form is êo, io; in Middle High German, ie in New High German, je; in Old Saxon, io; in Anglo-Saxon, â; in Norse, æ. Combined with this particle the word whit (thing) gives the following forms: Old High German, éowiht; Anglo-Saxon, âviht; Old Frisian, âwet; English, aught. The word naught is aught preceded by the negative particle. (Deutsche Grammatik, iii. 52.)

Each.—The particle gi enters, like the particle in the composition of pronouns. Old High German, êogalîher, every one; êocalih, all; Middle High German, iegelich; New High German, jeglich; Anglo-Saxon, ælc; English, each; the l being dropped, as in which and such. Ælc, as the original of the English each and the Scotch ilka,^[42] must by no means be confounded with the word ylce, the same. (Grimm, D. G. iii. 54.)

Every, in Old English, everich, everech, everilk one, is ælc, preceded by the particle ever. (Grimm, D. G. iii. 54.)

Either.—Old High German, êogahuëdar; Middle High German, iegewëder; Anglo-Saxon, æghväðer, ægðer; Old Frisian, eider.

Neither.—The same, with the negative article prefixed. Neither : either :: naught : aught.

Other, whether.—These words, although derived forms, being simpler than some that have preceded, might fairly {259}have been dealt with before. They make, however, a transition from the present to the succeeding chapter, and so find a place here.

A. First, it may be stated of them that the idea which they express is not that of one out of many, but that of one out of two.

1. In Sanskrit there are two forms, ^a) kataras, the same word as whether, meaning which out of two; ^b) katamas, which out of many. So also êkateras, one out of two; êkatamas, one out of many. In Greek, the Ionic form κότερος (πότερος); in Latin, uter, neuter, alter; and in Mœso-Gothic, hvathar, have the same form and the same meaning.

2. In the Scandinavian language the word anden, Dano-Saxon annar, Iceland corresponds to the English word second, and not the German zweite: e. g., Karl den Anden, Charles the Second. Now anthar is the older form of other.

B. Secondly, it may be stated of them, that the termination -er is the same termination that we find in the comparative degree.

1. The idea expressed by the comparative degree is the comparison, not of many, but of two things; this is better than that.

2. In all the Indo-European languages where there are pronouns in -ter, there is also a comparative degree in -ter. See next chapter.

3. As the Sanskrit form kataras corresponds with the comparative degree, where there is the comparison of two things with each other; so the word katamas is a superlative form; and in the superlative degree lies the comparison of many things with each other.

Hence other and whether (to which may be added either and neither) are pronouns with the comparative form.

Other has the additional peculiarity of possessing the plural form others. Hence, like self, it is, in the strictest sense, a substantival pronoun.

CHAPTER IX.

§ 306. Preparatory to the consideration of the degrees of comparison, it is necessary to make some remarks upon a certain class of words, which, with considerable differences of signification, all agree in one fact, viz., all terminate in -er, or t-er.

3. Certain adjectives, with the form of the comparative, but the power of the positive degree; as upp-er, und-er, inn-er, out-er, hind-er.

4. All adjectives of the comparative degree; as wis-er, strong-er, bett-er, &c.

Now what is the idea common to all these words, expressed by the sign -er, and connecting the four divisions into one class? It is not the mere idea of comparison; although it is the comparative degree, to the expression of which the affix in question is more particularly applied. Bopp, who has best generalised the view of these forms, considers the fundamental idea to be that of duality. In the comparative degree we have a relation between one object and some other object like it, or a relation between two single elements of comparison: A is wiser than B. In the superlative degree we have a relation between one object and all others like it, or a relation between one single and one complex element of comparison: A is wiser than B, C, D, &c.

"As in comparatives a relation between two, and in superlatives a relation between many, lies at the bottom, it is {261}natural that their suffixes should be transferred to other words, whose chief notion is individualised through that of duality or plurality."—Vergleichende Grammatik, § 292, Eastwick's and Wilson's Translation.

1. The Sanskrit forms kataras=which of two persons? a comparative form; katamas=which of more than two persons? a superlative form. Similarly, êkataras=one of two persons; êkatamas=one of more than two persons.

2. The Greek forms, ἑκάτερος=each or either out of two persons; ἕκαστος=each or any out of more than two persons.

§ 307. The more important of the specific modifications of the general idea involved in the comparison of two objects are,—

1. Contrariety; as in inner, outer, under, upper, over. In Latin the words for right and left end in -er,—dexter, sinister.

2. Choice in the way of an alternative; as either, neither, whether, other.

An extension of the reasoning probably explains forms like the Greek ἀμφό-τερ-ος, and the plural possessive forms νωΐ-τερ-ος, ἡμέ-τερ-ος, &c, which, like our own forms in -r, (ou-r, you-r) correspond in termination with the comparative degree (σοφώ-τερ-ος, wiser). Words, also, like hither and thither are instances of what is probably the effect of a similar association of ideas.

§ 308. A confirmation of Bopp's view is afforded by the Laplandic languages. Herein the distinction between one of two and one of more than two is expressed by affixes; and these affixes are the signs of the comparative and superlative: gi=who; gua-bba=who of two; gutte-mush=who of many.

3. Precisely as the words guabba and guttemush are formed, so also are the regular degrees of adjectives. {262}

The following extracts from Stockfleth's Lappish Grammar were probably written without any reference to the Sanskrit or Greek. "Guabba, of which the form and meaning are comparative, appears to have originated in a combination of the pronoun gi, and the comparative affix -abbo."—"Guttemush, of which the form and meaning are superlative, is similarly derived from the pronoun gutte, and the superlative affix -mush."—Grammatik i det Lappiske Sprog, §§ 192, 193.

§ 309. Either, neither, other, whether.—It has just been stated that the general fundamental idea common to all these forms is that of choice between one of two objects in the way of an alternative. Thus far the termination -er in either, &c., is the termination -er in the true comparatives, brav-er, wis-er, &c. Either and neither are common pronouns. Other, like one, is a pronoun capable of taking the plural form of a substantive (others), and also that of the genitive case (the other's money, the other's bread). Whether is a pronoun in the almost obsolete form whether (=which) of the two do you prefer, and a conjunction in sentences like whether will you do this or not? The use of the form others is recent. "They are taken out of the way as all other."—Job. "And leave their riches for other."—Psalms.

CHAPTER X.

§ 310. The proper preliminary to the study of the comparative and quasi-comparative forms in English is the history of the inflection or inflections by which they are expressed. There is no part of our grammar where it is more necessary to extend our view beyond the common limit of the Gothic stock of languages, than here.

In the Sanskrit language the signs of the comparative degree are two:—1. -tara, as punya=pure; punya-tara=purer; 2. -îyas, as kśipra=swift; kśêpîyas=swifter. Of these the first is the most in use.

The same forms occur in the Zend; as husko=dry; huskô-tara=drier; -îyas, however, is changed into -is.

In the classical languages we have the same forms. 1. in uter, neuter, alter, πότερος, λεπτότερος. 2. In the adverb magis, Lat. In Bohemian and Polish, -ssj and -szy correspond with the Sanskrit forms -îyas.

Thus we collect, that, expressive of the comparative degree, there are two parallel forms; viz., the form in tr, and the form in s; of which one is the most in use in one language, and the other in another.

§ 311. Before we consider the Gothic forms of the comparative, it may be advisable to note two changes to which it is liable. 1. The change of s into r; the Latin word meliorem being supposed to have been originally meliosem, and the s in nigrius, firmius, &c., being considered not so much the sign of the neuter gender as the old comparative s in its oldest form. 2. The ejection of t, as in the Latin words inferus, superus, compared with the Greek λεπτότερος (leptoteros). {264}

§ 312. Now, of the two parallel forms, the Gothic one was the form s; the words other and whether only preserving the form tr. And here comes the application of the remarks that have just gone before. The vast majority of our comparatives end in r, and so seem to come from tr rather than from s. This, however, is not the case. The r in words like sweeter is derived, not from tar—t, but from s, changed into r. In Mœso-Gothic the comparative ended in s (z); in Old High German the s has become r: Mœso-Gothic aldiza, batiza, sutiza; Old High German, altiro, betsiro, suatsiro; English, older, better, sweeter.

The importance of a knowledge of the form in s is appreciated when we learn that, even in the present English, there are vestiges of it.

§ 313. Comparison of adverbs.—The sun shines bright.—Herein the word bright means brightly; and although the use of the latter word would have been the more elegant, the expression is not ungrammatical; the word bright being looked upon as an adjectival adverb.

The sun shines to-day brighter than it did yesterday, and to-morrow it will shine brightest.—Here also the sense is adverbial; from whence we get the fact, that adverbs take degrees of comparison.

Now let the root mag-, as in magnus, μέγας, and mikil (Norse), give the idea of greatness. In the Latin language we have from it two comparative forms: 1. the adjectival comparative major=greater; 2. the adverbial comparative magis=more (plus). The same takes place in Mœso-Gothic: maiza means greater, and is adjectival; mais means more, and is adverbial. The Anglo-Saxon forms are more instructive still; e.g., þäs þe mâ=all the more, þäs þe bet=all the better, have a comparative sense, but not a comparative form, the sign r being absent. Now, compared with major, and subject to the remarks that have gone before, the Latin magis is the older form. With mâ and bet, compared with more and better, this may or may not be the case. Mâ and bet may each be one of two forms; 1. a positive used in a comparative sense; 2. a true comparative, which has lost {265}its termination. The present section has been written not for the sake of exhausting the subject, but to show that in the comparative degree there were often two forms; of which one, the adverbial, was either more antiquated, or more imperfect than the other: a fact bearing upon some of the forthcoming trains of etymological reasoning.

§ 314. Change of vowel.—By reference to Rask's Grammar, § 128, it may be seen that in the Anglo-Saxon there were, for the comparative and superlative degrees, two forms; viz. -or and -re, and -ost and -este, respectively.

By reference to p. 159 of the present volume, it may be seen that the fulness or smallness of a vowel in a given syllable may work a change in the nature of the vowel in a syllable adjoining. In the Anglo-Saxon the following words exhibit a change of vowel.

Of this change, the word last quoted is a still-existing specimen, as old, elder and older, eldest and oldest. Between the two forms there is a difference in meaning, elder being used as a substantive, and having a plural form, elders.

§ 315. The previous section has stated that in Anglo-Saxon there were two forms for the comparative and superlative degrees, one in -re and -este, the other in -or and -ost, respectively. Now the first of these was the form taken by adjectives; as se scearpre sweord=the sharper sword, and se scearpeste sweord=the sharpest sword. The second, on the other hand, was the form taken by adverbs; as, se sweord scyrð scearpor=the sword cuts sharper, and se sweord scyrð scearpost=the sword cuts sharpest.

The adjectival form has, as seen above, a tendency to make the vowel of the preceding syllable small: old, elder. {266}

The adverbial form has a tendency to make the vowel of the preceding syllable full.

Of this effect on the part of the adverbial form the adverbial comparative rather is a specimen. We pronounce the a as in father, or full. Nevertheless, the positive form is small, the a being pronounced as the a in fate.

The word rather means quick, easy=the classical root ῥαδ- in ῥάδιος. What we do quickly and willingly we do preferably. Now if the word rather were an adjective, the vowel of the comparative would be sounded as the a in fate. As it is, however, it is adverbial, and as such is properly sounded as the a in father.

The difference between the action of the small vowel in -re, and of the full in -or, effects this difference.

§ 316. Excess of expression.—Of this two samples have already been given: 1. in words like songstress; 2. in words like children. This may be called excess of expression; the feminine gender, in words like songstress, and the plural number, in words like children, being expressed twice over. In the vulgarism betterer for better, and in the antiquated forms worser for worse, and lesser for less, we have, in the case of the comparatives, as elsewhere, an excess of expression. In the Old High German we have the forms betsërôro, mêrôro, êrërëra=better, more, ere.

§ 317. Better.—Although in the superlative form best there is a slight variation from the strict form of that degree, the word better is perfectly regular. So far, then, from truth are the current statements that the comparison of the words good, better, and best is irregular. The inflection is not irregular, but defective. As the statement that applies to good, better, and best applies to many words besides, it will be well in this place, once for all, to exhibit it in full.

§ 318. Difference between a sequence in logic and a sequence in etymology.—The ideas or notions of thou, thy, thee, are ideas between which there is a metaphysical or logical connexion. The train of such ideas may be said to form a sequence and such a sequence may be called a logical one.

The forms (or words) thou, thy, thee, are forms or words {267}between which there is a formal or an etymological connexion. A train of such words may be called a sequence, and such a sequence may be called an etymological one.

In the case of thou, thy, thee, the etymological sequence tallies with the logical one.

The ideas of I, my, and me are also in a logical sequence: but the forms I, my, and me are not altogether in an etymological one.

In the case of I, my, me, the etymological sequence does not tally (or tallies imperfectly) with the logical one.

This is only another way of saying that between the words I and me there is no connexion in etymology.

It is also only another way of saying, that, in the oblique cases, I, and, in the nominative case, me, are defective.

Now the same is the case with good, better, bad, worse, &c. Good and bad are defective in the comparative and superlative degrees; better and worse are defective in the positive; whilst between good and better, bad and worse, there is a sequence in logic, but no sequence in etymology.

To return, however, to the word better; no absolute positive degree is found in any of the allied languages, and in none of the allied languages is there found any comparative form of good. Its root occurs in the following adverbial forms: Mœso-Gothic, bats; Old High German, pats; Old Saxon and Anglo-Saxon, bet; Middle High German, baz; Middle Dutch, bat, bet.—Grimm, D. G. iii. 604.

§ 319. Worse.—Mœso-Gothic, vairsiza; Old High German, wirsiro; Middle High German, wirser; Old Saxon, wirso; Anglo-Saxon, vyrsa; Old Norse, vërri; Danish, værre; and Swedish, värre. Such are the adjectival forms. The adverbial forms are Mœso-Gothic, vairs; Old High German, virs; Middle High German, wirs; Anglo-Saxon, vyrs: Old Norse, vërr; Danish, værre; Swedish, värre.—Grimm, D. G. iii. 606. Whether the present form in English be originally adjectival or adverbial is indifferent; since, as soon as the final a of vyrsa was omitted, the two words would be the same. The forms, however, vairsiza, wirser, worse, and vërri, make the word one of the most perplexing in the language. {268}

If the form worse be taken without respect to the rest, the view of the matter is simply that in the termination s we have a remnant of the Mœso-Gothic forms, like sutiza, &c., in other words, the old comparative in s.

Wirser and vairsiza traverse this view. They indicate the likelihood of the s being no sign of the degree, but a part of the original word. Otherwise the r in wirser, and the z in vairsiza, denote an excess of expression.

The analogies of songstress, children, and betsërôro show that excess of expression frequently occurs.

Finally the euphonic processes of the Scandinavian languages tell us that, even had there been an s, it would, in all probability, have been ejected. These difficulties verify the statement that the word worse is one of the most perplexing in the language.

§ 320. Much, more.—Here, although the words be unlike each other, there is a true etymological relation. Mœso-Gothic, mikils; Old High German, mihhil; Old Saxon, mikil; Anglo-Saxon, mycel; Old Norse, mickill; Scotch, muckle and mickle (all ending in l): Danish, megen, m.; meget, n.; Swedish, mycken, m.; myckett, n. (where no l is found). Such is the adjectival form of the positive, rarely found in the Modern Gothic languages, being replaced in German by gross, in English by great, in Danish by stor. The adverbial forms are miök and miög, Norse; much, English. It is remarkable that this last form is not found in Anglo-Saxon, being replaced by sâre, Germ, sehr.—Grimm, D. G. iii. 608.

The adverbial and the Norse forms indicate that the l is no part of the original word. Comparison with other Indo-European languages gives us the same circumstance: Sanskrit, maha; Latin, mag-nus; Greek, μέγας (megas).

There is in Mœso-Gothic the comparative form máiza, and there is no objection to presuming a longer form, magiza; since in the Greek form μείζων, compared with μέγας, there {269}is a similar disappearance of the g. In the Old High German we find mêro, corresponding with máiza, Mœso-Gothic, and with more, English.

Mickle (replaced by great) expresses size; much, quantity; many, number. The words more and most apply equally to number and quantity. I am not prepared either to assert or to deny that many, in Anglo-Saxon mænig, is from the same root with much. Of the word mâ notice has already been taken. Its later form, moe, occurs as late as Queen Elizabeth, with an adjectival as well as an adverbial sense.

§ 321. Little, less.—Like much and more, these words are in an etymological relation to each other. Mœso-Gothic, leitils; Old High German, luzil; Old Saxon, luttil; Anglo-Saxon, lytel; Middle High German, lützel; Old Norse, lîtill. In these forms we have the letter l. Old High German Provincial, luzíc; Old Frisian, litich; Middle Dutch, luttik; Swedish, liten; Danish, liden.—Deutsche Grammatik, iii. 611. From these we find that the l is either no part of the original word, or one that is easily got rid of. In Swedish and Danish there are the forms lille and liden; whilst in the neuter form, lidt, the d is unpronounced. Even the word liden the Danes have a tendency to pronounce leen. My own notion is that these changes leave it possible for less to be derived from the root of little. According to Grimm, the Anglo-Saxon lässa is the Gothic lasivôza, the comparative of lasivs=weak.—Deutsche Grammatik, iii. 611. In Anglo-Saxon there was the adjectival form læssa, and the adverbial form læs. In either case we have the form s.

§ 322. Near, nearer.—Anglo-Saxon, neah; comparative, nearre, near, nyr; superlative, nyhst, nehst. Observe, in the Anglo-Saxon positive and superlative, the absence of the r. This shows that the English positive near is the Anglo-Saxon comparative nearre, and that in the secondary comparative nearer, we have an excess of expression. It may be, however, that the r in near is a mere point of orthography, and that it is not pronounced. The fact that in the English language the words father and farther are, for the most part, pronounced alike, is the key to the forms near and nearer. {270}

§ 323. Farther.—Anglo-Saxon feor, fyrre, fyrrest. The th seems euphonic, inserted by the same process that gives the δ in ἄνδρος.

Further.—Confounded with farther, although in reality from a different word, fore. Old High German, furdir; New High German, der vordere; Anglo-Saxon, fyrðre.

§ 324. Former.—A comparative formed from the superlative; forma being such. Consequently, an instance of excess of expression, combined with irregularity.

Languages have a comparative without a superlative degree; no language has a superlative degree without having also a comparative one.

§ 325. In Mœso-Gothic spêdists means last, and spêdiza=later. Of the word spêdists two views may be taken. According to one it is the positive degree with the addition of st; according to the other, it is the comparative degree with the addition only of t. Now, Grimm and others lay down as a rule, that the superlative is formed, not directly from the positive, but indirectly through the comparative.

With the exception of worse and less, all the English comparatives end in r: yet no superlative ends in rt, the form being, not wise, wiser, wisert, but wise, wiser, wisest. This fact, without invalidating the notion just laid down, gives additional importance to the comparative forms in s; since it is from these, before they have changed to r, that we must suppose the superlatives to have been derived. The theory being admitted, we can, by approximation, determine the comparative antiquity of the superlative degree. It was introduced into the Indo-European tongues after the establishment of the comparative, and before the change of -s into -r. I give no opinion as to the truth of this theory.

CHAPTER XI.

§ 326. The history of the superlative form, accurately parallel with what has been stated of the comparative, is as follows:—

In Sanskrit there is, 1. the form tama, 2. the form ishta; the first being the commonest. The same is the case in the Zend.

Each of these appears again in the Greek. The first, as τατ (tat), in λεπτότατος (leptotatos); the second, as ιστ (ist), in οἴκτιστος (oiktistos). For certain reasons, Grimm thinks that the tat stands for tamt, or tant.

In Latin, words like intimus, extimus, ultimus, preserve im; whilst venustus, vetustus, and robustus, are considered as positives, preserving the superlative form -st.

Just as in inferus and nuperus, there was the ejection of the t in the comparative ter, so in infimus, nigerrimus, &c., is there the ejection of the same letter in the superlative tim.

This gives us, as signs of the superlative, 1. tm; 2. st; 3. m, t being lost; 4. t, m being lost.

Of the first and last of these, there are amongst the true superlatives, in English, no specimens.

Of the third, there is a specimen in the Anglo-Saxon se forma, the first, from the root fore, as compared with the Latin primus, and the Lithuanic pirmas.

Of the English superlatives, the only ones that demand a detailed examination are those that are generally despatched without difficulty; viz., the words in most; such as midmost, foremost, &c. The current view is the one adopted by Rask in his Anglo-Saxon Grammar (§ 133), viz., that they are {272}compound words, formed from simple ones by the addition of the superlative term most. Grimm's view is opposed to this. In appreciating Grimm's view, we must bear in mind the phenomena of excess of expression; at the same time we must not depart from the current theory without duly considering the fact stated by Rask; which is, that we have in Icelandic the forms nærmeir, fjærmeir, &c., nearer, and farther, most unequivocally compounded of near and more, and of far and more.

Let especial notice be taken of the Mœso-Gothic forms fruma, first; aftuma, last; and of the Anglo-Saxon forms forma, aftema, aftermost; ufema, upmost; hindema, hindmost; midema, midmost; innema, inmost; ûtema, outmost; siðema, last; latema, last; niðema, nethermost. These account for the m.

Add to this, with an excess of expression, the letters st. This accounts for the whole form, as mid-m-ost, in-m-ost, &c. Such is Grimm's view.

Furthermost, innermost, hindermost.—Here there is a true addition of most, and an excess of inflection, a superlative form being added to a word in the comparative degree.

Former.—Here, as stated before, a comparative sign is added to a word in the superlative degree.

§ 327. The combination st occurs in other words besides those of the superlative degree; amongst others, in certain adverbs and prepositions, as among, amongst; while, whilst; between, betwixt.—Its power here has not been well explained.

CHAPTER XII.

§ 328. In one sense the cardinal numbers form no part of a work on etymology. They are single words, apparently simple, and, as such, appertaining to a dictionary rather than to a grammar.

In another sense they are strictly etymological. They are the basis of the ordinals, which are formed from them by derivation. Furthermore, some of them either have, or are supposed to have, certain peculiarities of form which can be accounted for only by considering them derivatives, and that of a very peculiar kind.

§ 329. It is an ethnological fact, that the numerals are essentially the same throughout the whole Indo-European class of languages. The English three is the Latin tres, the Sanskrit tri, &c. In the Indo-European languages the numerals agree, even when many common terms differ.

And it is also an ethnological fact, that in a great many other groups of languages the numerals differ, even when many of the common terms agree. This is the case with many of the African and American dialects. Languages alike in the common terms for common objects differ in respect to the numerals.

What is the reason for this inconsistency in the similarity or dissimilarity of the numerals as compared with the similarity or dissimilarity of other words? I believe that the following distinction leads the way to it:—

The word pair also=2; but not absolutely, not unequivocally, and only in a secondary manner. {274}

Hence the distinction between absolute terms expressive of number, and secondary terms expressive of number.

When languages separate from a common stock before the use of certain words is fixed as absolute, there is room for considerable latitude in the choice of numerals; e.g., whilst with one tribe the word pair=two, another tribe may use the word couple, a third brace, and so on. In this case dialects that agree in other respects may differ in respect to their numerals.

When, on the other hand, languages separate from a common stock after the meaning of such a word as two has been fixed absolutely, there is no room for latitude; and the numerals agree where the remainder of the language differs.

4. Four=quatuor, τέτταρα. This is apparently problematical. Nevertheless, the assumed changes can be verified by the following forms:—

β. Πίσυρες, Æolic. Illustrates the change between τ- and π- (allied to f-), within the pale of the classical languages.

β. Pump, Welsh. These account for the change from the n + t in πέντε to m + p.

The change from the π- of πέντε to the qu- of quinque is the change so often quoted by Latin and Celtic scholars between p and k: ἵππος, ἵκκος, equus.

This form is difficult. The Mœso-Gothic form is sibun, without a -t-; the Norse, syv, without either -t- or -n (=-m). A doubtful explanation of the form seven, &c., will be found in the following chapter. {275}

9. Nine=ἐννέα, novem. The Mœso-Gothic form is nigun, the Icelandic niu. In the Latin novem the v=the g of nigun. In the English and Greek it is wanting. The explanation of the -n and -m will be found in the following chapter.

10. Ten=δέκα, decem. The Mœso-Gothic form is tihun; wherein the h=the c of decem and the κ of δέκα. The Icelandic form is tiu, and, like δέκα, is without the -n (or -m). The hypothesis as to the -m or -n will be given in the next chapter.

α. The e is ein=one. Einlif, ein-lef, eilef, eilf, elf, Old High German; andlova, Old Frisian; end-leofan, endlufan, Anglo-Saxon. This is universally admitted.

β. The -lev- is a modification of the root laib-an=manere=to stay=to be over. Hence eleven=one over (ten). This is not universally admitted.

γ. The -n has not been well accounted for. It is peculiar to the Low Germanic dialects.—Deutsche Grammatik, ii. 946.

12. Twelve=the root two + the root laib=two over (ten). Tvalif, Mœso-Gothic; zuelif, Old High German; toll, Swedish. The same doubts that apply to the doctrine of the -lv- in eleven representing the root -laib, apply to the -lv- in twelve.—Deutsche Grammatik, ii. 946.

30. Thirty=3 × 10, or three decads. This difference in the decimal power of the syllables -teen and -ty is illustrated by—

α. The Mœso-Gothic.—Here we find the root tig- used as a true substantive, equivalent in form as well as power to the Greek δέκ-ας. Tváim tigum þusandjom=duobus decadibus myriadum. (Luke xiv. 31.) Jêrê þrijê tigivé=annorum duarum decadum. (Luke iii. 23.) þrins tiguns silubrinaize=tres decadas argenteorum. (Matthew xxvii. 3, 9.)—Deutsche Grammatik, ii. 948. {276}

β. The Icelandic.—"The numbers from 20 to 100 are formed by means of the numeral substantive, tigr, declined like viðr, and naturally taking the word which it numerically determines in the genitive case.

"This is the form of the inflection in the best and oldest MSS. A little later was adopted the indeclinable form tigi, which was used adjectivally."—Det Oldnorske Sprogs Grammatik, af P. A. Munch, og C. B. Unger, Christiania, 1847.

§ 330. Generally speaking, the greater part of the numerals are undeclined, even in inflected languages. As far as number goes, this is necessary.

As to the inflection of gender and cases, there is no reason why all the numerals should not be as fully inflected as the Latin unus, una, unum, unius.

CHAPTER XIII.

§ 331. The remarks at the close of the last chapter but one indicated the fact that superlative forms were found beyond the superlative degree. The present chapter shows that they are certainly found in some, and possibly in all of the ordinal numbers.

First.—In Mœso-Gothic, fruma, frumist; in Anglo-Saxon, forma, fyrmest; in Old High German, vurist; in Old Norse, fyrst; in New High German, erst. In all these words, whether in m, in mst, or in st, there is a superlative form. The same is the case with pratamas, Sanskrit; fratemas, Zend; πρῶτος, Greek; primus, Latin; primas, Lithuanic. Considering that, compared with the other ordinals, the ordinal of one is a sort of superlative, this is not at all surprising.

Between the words one and first there is no etymological relation. This is the case in most languages. Unus, primus, ἑῖς, πρῶτος, &c.

§ 332. Second.—Between this word and its cardinal, two, there is no etymological connexion. This is the case in many, if not in most, languages. In Latin the cardinal is duo, and the ordinal secundus, a gerund of sequor, and meaning the following. In Anglo-Saxon the form was se oðer=the other. In the present German, the ordinal is zweite, a word etymologically connected with the cardinal zwei=two.

Old High German, andar; Old Saxon, othar; Old Frisian, other; Middle Dutch, ander. In all these words we have the comparative form -ter; and considering that, compared with the word first, the word second is a sort of {278}comparative, there is nothing in the circumstance to surprise us. The Greek forms δεύτερος and ἕτερος, the Latin alter, and the Lithuanic antras, are the same.

§ 333. With the third ordinal number begin difficulties: 1. in respect to their form; 2. in respect to the idea conveyed by them.

1. Comparing third, fourth, fifth, &c., with three, four, and five, the formation of the ordinal from the cardinal form may seem simply to consist in the addition of d or th. Such, however, is far from being the case.

2. Arguing from the nature of the first two ordinals, namely, the words first and second, of which one has been called a superlative and the other a comparative, it may seem a simple matter to associate, in regard to the rest, the idea of ordinalism with the idea of comparison. A plain distinction, however, will show that the case of the first two ordinals is peculiar. First is a superlative, not as compared with its cardinal, one, but as compared with the other numerals. Second, or other, is a comparative, not as compared with its cardinal, two, but as compared with the numeral one. Now it is very evident, that, if the other ordinals be either comparatives or superlatives, they must be so, not as compared with one another, but as compared with their respective cardinals. Sixth, to be anything like a superlative, must be so when compared with six.

§ 334. Now there are, in etymology, two ways of determining the affinity of ideas. The first is the metaphysical, the second the empirical, method.

This is better than that, is a sentence which the pure metaphysician may deal with. He may first determine that there is in it the idea of comparison; and next that the comparison is the comparison between two objects, and no more than two. This idea he may compare with others. He may determine, that, with a sentence like this is one and that is the other, it has something in common; since both assert something concerning one out of two objects. Upon this connexion in sense he is at liberty to reason. He is at liberty to conceive that in certain languages words expressive {279}of allied ideas may also be allied in form. Whether such be really the case, he leaves to etymologists to decide.

The pure etymologist proceeds differently. He assumes the connexion in meaning from the connexion in form. All that he at first observes is, that words like other and better have one and the same termination. For this identity he attempts to give a reason, and finds that he can best account for it by presuming some affinity in sense. Whether there be such an affinity, he leaves to the metaphysician to decide. This is the empirical method.

At times the two methods coincide, and ideas evidently allied are expressed by forms evidently allied.

At times the connexion between the ideas is evident; but the connexion between the forms obscure: and vice versâ. Oftener, however, the case is as it is with the subjects of the present chapter. Are the ideas of ordinalism in number, and of superlativeness in degree, allied? The metaphysical view, taken by itself, gives us but unsatisfactory evidence; whilst the empirical view, taken by itself, does the same. The two views, however, taken together, give us evidence of the kind called cumulative, which is weak or strong according to its degree.

Compared with three, four, &c., all the ordinals are formed by the addition of th, or t; and th, ð, t, or d, is the ordinal sign, not only in English, but in the other Gothic languages. But, as stated before, this is not the whole of the question.

The letter t is found, with a similar power, 1. In Latin, as in tertius, quartus, quintus, sextus; 2. Greek, as in τρίτος (tritos), τέταρτος (tetartos), πέμπτος (pemptos), ἕκτος (hectos), ἔννατος (ennatos), δέκατος (dekatos); 3. Sanskrit, as in tritiyas, ćatuŕtas, shasht´as=third, fourth, sixth; 4. In Zend, as in thrityas=the third, haptathas=the seventh; 5. In Lithuanic, as ketwirtas=fourth, penktas=fifth, szesztas=sixth; 6. In Old Slavonic, as in cétvertyi=fourth, pjatyi=fifth, shestyi=sixth, devjatyi=ninth, desjatyi=tenth. Speaking more generally, it is found, with a similar force, throughout the Indo-European stock.

The following forms indicate a fresh train of reasoning. {280}The Greek ἑπτὰ (hepta), and Icelandic sjau, have been compared with the Latin septem and the Anglo-Saxon seofon. In the Greek and Icelandic there is the absence, in the Latin and Anglo-Saxon the presence, of a final liquid (m or n).

Again, the Greek forms ἐννέα (ennea), and the Icelandic níu=nine, have been compared with the Latin novem and the Gothic nigun.

Thirdly, the Greek δέκα (deka), and the Icelandic tíu, have been compared with the Latin decem and the Gothic tihun=ten.

These three examples indicate the same circumstance; viz. that the m or n, in seven, nine, and ten, is no part of the original word.

§ 335. The following hypotheses account for these phenomena; viz. that the termination of the ordinals is the superlative termination -tam: that in some words, like the Latin septimus, the whole form is preserved; that in some, as in τέταρτος=fourth, the t only remains; and that in others, as in decimus, the m alone remains. Finally, that in seven, nine, and ten, the final liquid, although now belonging to the cardinal, was once the characteristic of the ordinal number. For a fuller exhibition of these views, see Grimm, Deutsche Grammatik, iii. 640.

CHAPTER XIV.

§ 336. In the generality of grammars the definite article the, and the indefinite article an, are the very first parts of speech that are considered. This is exceptionable. So far are they from being essential to language, that, in many dialects, they are wholly wanting. In Greek there is no indefinite, in Latin there is neither an indefinite nor a definite article. In the former language they say ἀνήρ τις=a certain man: in the Latin the words filius patris mean equally the son of the father, a son of a father, a son of the father, or the son of a father. In Mœso-Gothic and in Old Norse, there is an equal absence of the indefinite article; or, at any rate, if there be one at all, it is a different word from what occurs in English. In these the Greek τις is expressed by the Gothic root sum.

Now, as it is very evident that, as far as the sense is concerned, the words some man, a certain man, and a man, are, there or thereabouts, the same, an exception may be taken to the statement that in Greek and Mœso-Gothic there is no indefinite article. It may, in the present state of the argument, be fairly said that the words sum and τις are pronouns with a certain sense, and that a and an are no more; consequently, that in Greek the indefinite article is τις, in Mœso-Gothic sum, and in English a or an,

A distinction, however, may be made. In the expression ἀνήρ τις (anær tis)=a certain man, or a man, and in the expression sum mann, the words sum and τις preserve their natural and original meaning; whilst in a man and an ox the words a and an are used in a secondary sense. These words, as is currently known, are one and the same, the n, in the form a, being ejected through a euphonic process. They are, moreover, the same words with the numeral one; {282}Anglo-Saxon, án; Scotch, ane. Now, between the words a man and one man, there is a difference in meaning; the first expression being the most indefinite. Hence comes the difference between the English and the Mœso-Gothic expressions. In the one the word sum has a natural, in the other the word an has a secondary power.

The same reasoning applies to the word the. Compared with a man, the words the man are very definite. Compared, however, with the words that man, they are the contrary. Now, just as an and a have arisen out of the numeral one, so has the arisen out of the demonstrative pronoun þæt, or at least from some common root. It will be remembered that in Anglo-Saxon there was a form þe, undeclined, and common to all the cases of all the numbers.

In no language in its oldest stage is there ever a word giving, in its primary sense, the ideas of a and the. As tongues become modern, some noun with a similar sense is used to express them. In the course of time a change of form takes place, corresponding to the change of meaning; e. g., one becomes an, and afterwards a. Then it is that articles become looked upon as separate parts of speech, and are dealt with accordingly. No invalidation of this statement is drawn from the Greek language. Although the first page of the etymology gives us ὁ, ἡ, τὸ (ho, hæ, to), as the definite articles, the corresponding page in the syntax informs us, that, in the oldest stage of the language, ὁ (ho)=the, had the power of οὗτος (howtos)=this.

The origin of the articles seems uniform. In German ein, in Danish en, stand to one in the same relation that an does. The French un, Italian and Spanish uno, are similarly related to unus=one.

And as, in English the, in German der, in Danish den, come from the demonstrative pronouns, so in the classical languages are the French le, the Italian il and lo, and the Spanish el, derived from the Latin demonstrative, ille.

In his Outlines of Logic, the present writer has given reasons for considering the word no (as in no man) an article.

That the, in expressions like all the more, all the better, &c., is no article, has already been shown.

CHAPTER XV.

§ 337. Compared with the words lamb, man, and hill, the words lambkin, mannikin, and hillock convey the idea of comparative smallness or diminution. Now, as the word hillock=a little hill differs in form from hill we have in English a series of diminutive forms, or diminutives.

The English diminutives may be arranged according to a variety of principles. Amongst others:

1. According to their form.—The word hillock is derived from hill, by the addition of a syllable. The word tip is derived from top, by the change of a vowel.

2. According to their meaning.—In the word hillock there is the simple expression of comparative smallness in size. In the word doggie for dog, lassie for lass, the addition of the -ie makes the word not so much a diminutive as a term of tenderness or endearment. The idea of smallness, accompanied, perhaps, with that of neatness, generally carries with it the idea of approbation. The word clean in English, means, in German, little=kleine. The feeling of protection which is extended to small objects engenders the notion of endearment. In Middle High German we have vaterlìn=little father, mütterlìn=little mother. In Middle High German there is the diminutive sunnelìn; and the French soleil is from the Latin form solillus. In Slavonic the word slunze=sun is a diminutive form.

The Greek word μείωσις (meiôsis) means diminution; the Greek word ὑποκόρισμα means an endearing expression. Hence we get names for the two kinds of diminutives; viz., the term meiotic for the true diminutives, and the term hypocoristic for the diminutives of endearment.—Grimm, Deutsche Grammatik, iii. 664. {284}

3. According to their historical origin.—The syllable -ock, as in hillock, is of Anglo-Saxon and Gothic origin. The -et, as in lancet, is of French and classical origin.

4. According as they affect proper names or common names.—Hawkin, Perkin, Wilkin, &c. In these words we have the diminutives of Hal, Peter, Will, &c.

1. Those formed by a change of vowel.—Tip, from top. The relation of the feminine to the masculine is allied to the ideas conveyed by many diminutives. Hence in the word kit, from cat, it is doubtful whether there be meant a female cat or a little cat. Kid is a diminutive form of goat.

2. Those formed by the addition of a letter or letters.—Of the diminutive characteristics thus formed the commonest, beginning from the simpler forms, are

Ie.—Almost peculiar to the Lowland Scotch; as daddie, lassie, minnie, wifie, mousie, doggie, boatie, &c.—Deutsche Grammatik, iii. 686.

Kin.—Lambkin, mannikin, ladikin, &c. As is seen above, common in proper names.

En.—Chicken, kitten, from cock, cat. The notion of diminution, if indeed that be the notion originally conveyed, lies not in the -en, but in the vowel. In the word chicken, from cock, observe the effect of the small vowel on the c.

Et, as in trumpet, lancet, pocket; the word pock, as in meal-pock=a meal-bag, being found in the Scottish. From the French -ette, as in caissette, poulette.

The forms -rel, as in cockerel, pickerel, and -let, as in streamlet, require a separate consideration. The first has nothing to do with the Italian forms acquerella and coserella—themselves, perhaps, of Gothic, rather than of classical origin.

In the Old High-German there are a multitude of diminutive forms in -l; as ouga=an eye, ougili=a little eye, lied=a song, liedel=a little song. "In Austria and Bavaria {285}are the forms mannel, weibel, hundel, &c., or mannl, weibl, hundl, &c. In some districts there is an r before the l, as madarl=a little maid, muadarl=a little mother, briadarl=a little brother, &c. This is occasioned by the false analogy of the diminutives of the derived form in r."—Deutsche Grammatik, iii. p. 674. This indicates the nature of words like cockerel.

Even in English the diminutive power of -el can be traced in the following words:—

Soare=a deer in its third year. Sor-rel=a deer in its second year.—See Love's Labour Lost, with the note.

Hurdle; in Dutch horde; German, hurde. Hording, without the -l, is used in an allied sense by builders in English.

In the words in point we must assume an earlier form, cocker and piker, to which the diminutive form -el is affixed. If this be true, we have, in English, representatives of the diminutive form -l, so common in the High Germanic dialects. Wolfer=a wolf, hunker=a haunch, flitcher=a flitch, teamer=a team, fresher=a frog,—these are north country forms of the present English.^[43]

The termination -let, as in streamlet, seems to be double, and to consist of the Gothic diminutive -l, and the French diminutive -t.

§ 339. Augmentatives.—Compared with capello=a hat, the Italian word capellone=a great hat is an augmentative. The augmentative forms, pre-eminently common in the Italian language, often carry with them a depreciating sense.

The termination -rd (in Old High German, -hart), as in drunkard, braggart, laggard, stinkard, carries with it this idea of depreciation. In buzzard, and reynard, the name of the fox, it is simply augmentative. In wizard, from witch, it has the power of a masculine form.

The termination -rd, taken from the Gothic, appears in {286}the modern languages of classical origin: French, vieillard; Spanish, codardo. From these we get at, second-hand, the word coward.—Deutsche Grammatik, iii. 707.

The word sweetheart is a derived word of this sort, rather than a compound word; since in Old High German and Middle High German, we have the corresponding form liebhart. Now the form for heart is in German not hart, but herz.

Words like braggadocio, trombone, balloon, being words of foreign origin, prove nothing as to the further existence of augmentative forms in English.

§ 340. Patronymics.—In the Greek language the notion of lineal descent, in other words, the relation of the son to the father, is expressed by a particular termination; as, Πηλεὺς (Peleus), Πηλείδης (Peleidæs), the son of Peleus. It is very evident that this mode of expression is very different from either the English form Johnson, or Gaelic MacDonald. In these last-named words, the words son and Mac mean the same thing; so that Johnson and MacDonald are not derived, but compound words. This Greek way of expressing descent is peculiar, and the words wherein it occurs are classed together by the peculiar name patronymic, from patær=a father, and onoma=a name. Is there anything in English corresponding to the Greek patronymics? It was for the sake of this question that the consideration of the termination -ling, as in duckling, &c., was deferred.

The termination -ling, like the terminations -rel and -let, is compound. Its simpler form is -ing. This, from being affixed to the derived forms in -l, has become -ling.

In Anglo-Saxon the termination -ing is as truly patronymic as -ιδης is in Greek. In the Bible-translation the son of Elisha is called Elising. In the Anglo-Saxon Chronicle occur such genealogies as the following:—Ida wæs Eopping, Eoppa Êsing, Êsa Inging, Inga Angenviting, Angenvit Alocing, Aloc Beonocing, Beonoc Branding, Brand Bældæging, Bældæg Vódening, Vóden Friðowulfing, Friðowulf Finning, Finn Godwulfing, Godwulf Geating=Ida was the son of Eoppa, Eoppa of Esing, Esing of Inga, Inga of Angenvit, {287}Angenvit of Aloc, Aloc of Beonoc, Beonoc of Brand, Brand of Bældag, Bældag of Woden, Woden of Friðowulf, Friðowulf of Finn, Finn of Godwulf, Godwulf of Geat.—In Greek, Ἴδα ἦν Ἐοππείδης, Ἔοππα Ἠσείδης, Ἤσα Ἰγγείδης, Ἴγγα Ἀγγενφιτείδης, &c. In the plural number these forms denote the race of; as Scyldingas=the Scyldings, or the race of Scyld, &c. Edgar Atheling means Edgar of the race of the nobles. The primary of -ing and -l-ing is descent or relationship; from these comes the idea of youth and endearment, and thence the true diminutive idea. In darling, stripling, duckling, gosling (pr. gesling), kitling (pr. for kitten), nestling, yearling, chickling, fatling, fledgling, firstling, the idea of descent still remains. In hireling the idea of diminution is accompanied with the idea of contempt. In changeling we have a Gothic termination and a classical root. See, for the full exposition of this view, Deutsche Grammatik, ii. 349-364, iii. 682.

The word silverlings has troubled the commentators. Burst their silverbins has been proposed as the true reading. The word, however, is a true diminutive, as siluparlinc, silarbarling=a small silver coin, Old High German.

A good chapter on the English diminutives may be seen in the Cambridge Philological Museum, vol. i. p. 679.

CHAPTER XVI.

§ 341. These have been illustrated by Mr. Guest in the Transactions of the Philological Society.

The only word in the present English that requires explanation is the name of the principality Wales.

1. The form is plural, however much the meaning may be singular; so that the -s in Wale-s is the -s in fathers, &c.

2. It has grown out of the Anglo-Saxon from wealhas=foreigners, the name by which the Welsh are spoken of by the Germans of England, just as the Italians are called Welsh by the Germans of Germany: wal-nuts=foreign nuts.

3. The transfer of the name of the people inhabiting a certain country to the country so inhabited, was one of the commonest processes in both Anglo-Saxon and Old English.—Guest, Phil. Trans.

CHAPTER XVII.

ON THE CONNEXION BETWEEN THE NOUN AND VERB, AND ON THE INFLECTION OF THE INFINITIVE MOOD.

§ 342. In order to understand clearly the use of the so-called infinitive mood in English, it is necessary to bear in mind two facts, one a matter of logic, the other a matter of history.

In the way of logic, the difference between a noun and a verb is less marked than it is in the way of grammar.

Grammatically, the contrast is considerable. The inflection of nouns expresses the ideas of sex as denoted by gender, and of relation in place as denoted by cases. That of verbs rarely expresses sex, and never position. On the other hand, however, it expresses what no noun ever does or can express; e.g., the relation of the agent to the individual speaking, by means of person; the time in which acts take place, by means of tense; and the conditions of their occurrence, by means of mood.

The idea of number is the only one that, on a superficial view, is common to these two important parts of speech.

Logically, the contrast is inconsiderable. A noun denotes an object of which either the senses or the intellect can take cognizance, and a verb does no more. To move=motion, to rise=rising, to err=error, to forgive=forgiveness. The only difference between the two parts of speech is this, that, whereas a noun may express any object whatever, verbs can only express those objects which consist in an action. And it is this superadded idea of action that superadds to the verb the phenomena of tense, mood, person, and voice; in other words, the phenomena of conjugation.

§ 343. A noun is a word capable of declension only. A {290}verb is a word capable of declension and conjugation also. The fact of verbs being declined as well as conjugated must be remembered. The participle has the declension of a noun adjective, the infinite mood the declension of a noun substantive. Gerunds and supines, in languages where they occur, are only names for certain cases of the verb.

Although in all languages the verb is equally capable of declension, it is not equally declined. The Greeks, for instance, used forms like

oftener than the Romans. The fact of there being an article in Greek may account for this.

§ 344. Returning, however, to the illustration of the substantival character of the so-called infinitive mood, we may easily see—

α. The name of any action may be used without any mention of the agent. Thus, we may speak of the simple fact of walking or moving, independently of any specification of the walker or mover.

β. That, when actions are spoken of thus indefinitely, the idea of either person or number has no place in the conception; from which it follows that the so-called infinitive mood must be at once impersonal, and without the distinction of singular, dual, and plural.

γ. That, nevertheless, the ideas of time and relation in space have place in the conception. We can think of a person being in the act of striking a blow, of his having been in the act of striking a blow, or of his being about to be in the act of striking a blow. We can also think of a person being in the act of doing a good action, or of his being from the act of doing a good action.

This has been written to show that verbs of languages in general are as naturally declinable as nouns. What follows will show that the verbs of the Gothic languages in particular were actually declined, and that fragments of this declension remain in the present English.

§ 345. The inflection of the verb in its impersonal (or {291}infinitive form) consisted, in full, of three cases, a nominative (or accusative), a dative, and a genitive. The genitive is put last, because its occurrence in the Gothic language is the least constant.

Be it observed, that the -en in words like strengthen, &c., is a derivational termination, and by no means a representation of the Anglo-Saxon infinitive inflection. The Anglo-Saxon infinitive inflection is lost in the present English, except in certain provincial dialects.

In Anglo-Saxon the dative of the infinitive verb ended in -nne, and was (as a matter of syntax) generally, perhaps always, preceded by the preposition to.

The genitive, ending in -es, occurs only in Old High German and Modern High German, plâsannes, weinnenes.

§ 346. With these preliminaries we can take a clear view of the English infinitives. They exist under two forms, and are referable to a double origin.

1. The independent form.—This is used after the words can, may, shall, will, and some others, as, I can speak, I may go, I shall come, I will move. Here there is no preposition, and the origin of the infinitive is from the form in -an.

2. The prepositional form.—This is used after the majority of English verbs, as I wish to speak, I mean to go, I intend to come, I determine to move. Here we have the preposition to and the origin of the infinitive is from the form in -nne.

are very remarkable. They exhibit the phenomena of a nominative case having grown not only out of a dative but out of a dative plus its governing preposition.

CHAPTER XVIII.

§ 347. Of number, person, mood, tense, and conjugation, special notice is taken in their respective chapters. Of the divisions of verbs into active and passive, transitive and intransitive, unless there be an accompanying change of form, etymology takes no cognisance. The forces of the auxiliary verbs, and the tenses to which they are equivalent, are also points of syntax rather than of etymology.

Four classes, however, of derived verbs, as opposed to simple, especially deserve notice.

I. Those ending in -en; as soften, whiten, strengthen, &c. Here it has been already remarked that the -en is a derivational affix; and not a representative of the Anglo-Saxon infinitive form -an (as lufian, bærnan=to love, to burn), and the Old English -en (as tellen, loven).

II. Transitive verbs derived from intransitives by a change of the vowel of the root.

In Anglo-Saxon these words were more numerous than they are at present. The following list is taken from the Cambridge Philological Museum, ii. 386.

All these intransitives form their præterite by a change of vowel, as sink, sank; all the transitives by the addition of d or t, as fell, fell'd.

III. Verbs derived from nouns by a change of accent; as to survéy, from a súrvey. For a fuller list see the Chapter on Derivation. Walker attributes the change of accent to the influence of the participial termination -ing. All words thus affected are of foreign origin.

IV. Verbs formed from nouns by changing a final sharp consonant into its corresponding flat one; as,

CHAPTER XIX.

§ 348. Compared with the Latin, the Greek, the Mœso-Gothic, and almost all the ancient languages, there is, in English, in respect to the persons of the verbs, but a very slight amount of inflection. This may be seen by comparing the English word call with the Latin voco.

Here the Latins have different forms for each different person, whilst the English have forms for two only; and even of these one (callest) is becoming obsolete. With the forms of voco marked in italics there is, in the current English, nothing correspondent.

In the word am, as compared with are and art, we find a sign of the first person singular.

In the Modern English, the Old English, and the Anglo-Saxon, the peculiarities of our personal inflections are very great. This may be seen from the following tables of comparison:—

Herein remark; 1. the Anglo-Saxon addition of t in the second person singular; 2. the identity in form of the three persons of the plural number; 3. the change of -að into -en in the Old English plural; 4. the total absence of plural forms in the Modern English; 5. the change of the th into s, in loveth and loves. These are points bearing especially upon the history of the English persons. The following points indicate a more general question.

1. The full form prennames in the newer Old High German, as compared with sókjam in the old Mœso-Gothic.

3. The difference between the Old Saxon and the Anglo-Saxon in the second person singular; the final t being absent in Old Saxon.

4. The respective powers of M in the first, of S in the second, and of T (or its allied sounds) in the third persons singular; {296}of MES in the first, of T (or its allied sounds) in the second, and of ND in the third persons plural. In this we have a regular expression of the persons by means of regular signs; and this the history of the personal terminations verifies.

§ 349. First person singular.—That the original sign of this person was M we learn from the following forms: dadâmi, Sanskrit; dadhâmi, Zend; δίδωμι, Greek; dumi, Lithuanic; damy, Slavonic=I give. The Latin language preserves it in sum and inquam, and in the first persons of tenses, like legam, legebam, legerem, legissem. The form im=I am occurs in Mœso-Gothic; and the words stom=I stand, lirnem=I shall learn, in Old High German. The word am is a fragmentary specimen of it in our own language.

Plural.—The original sign MES. Dadmas, Sanskrit; δίδομες, afterwards δίδομεν, Greek; damus, Latin=we give. The current form in Old High German.

These forms in M may or may not be derived from the pronoun of the first person; mâ, Sanskrit; me, Latin, English, &c.

Second person singular.—The original sign S. Dadasi, Sanskrit; δίδως, Greek; das, Latin; dasi, Slavonic. Preserved in the Gothic languages.

Plural.—The original sign T, or an allied sound. Dadyata, Sanskrit; daidhyâta, Zend; δίδοτε, Greek; datis, Latin; d[ou]kite, Lithuanic; dashdite, Slavonic=ye give. Current in the Gothic languages.

These forms in T and S may or may not be derived from the pronoun of the second person; tva, Sanskrit; σὺ, Greek; thou, English.

Third person singular.—-The original sign T. Dadati, Sanskrit; dadhâiti, Zend; δίδωτι, Old Greek; dat, Latin; d[ou]sti, Lithuanic; dasty, Slavonic=he gives. Preserved in the Gothic languages.

Plural.—The original sign NT. Dadenti, Zend; δίδοντι, afterwards διδοῦσι, Greek; dant, Latin=they give. In Mœso-Gothic and Old High German.

The preceding examples are from Grimm and Bopp. To them add the Welsh form carant=they love, and the Persian budend=they are. {297}

The forms in T and NT may or may not be derived from the demonstrative pronoun ta, Saxon; τὸ, Greek; that, English, &c.

§ 350. The present state of the personal inflection in English, so different from that of the older languages, has been brought about by two processes.

I. Change of form.—^a) The ejection of -es in -mes, as in sôkjam and köllum, compared with prennames; ^b) the ejection of -m, as in the first person singular, almost throughout; ^c) the change of -s into -r, as in the Norse kallar, compared with the Germanic sôkeis; ^d) the ejection of -d from -nd, as in loven (if this be the true explanation of that form) compared with prennant; ^e) the ejection of -nd, as in kalla; ^f) the addition of -t, as in lufast and lovest. In all these cases we have a change of form.

II. Confusion or extension.—In vulgarisms like I goes, I is, one person is used instead of another. In vulgarisms like I are, we goes, one number is used instead of another. In vulgarisms like I be tired, or if I am tired, one mood is used instead of another. In vulgarisms like I give for I gave, one tense is used for another. In all this there is confusion. There is also extension: since, in the phrase I is, the third person is used instead of the first; in other words, it is used with an extension of its natural meaning. It has the power of the third person + that of the first. In the course of time one person may entirely supplant, supersede, or replace another. The application of this is as follows:—

The only person of the plural number originally ending in ð is the second; as sókeiþ, prennat, kalliþ, lufiað; the original ending of the first person being -mes, or -m, as prennames, sôkjam, köllum. Now, in Anglo-Saxon, the first person ends in ð, as lufiað. Has -m, or -mes, changed to ð, or has the second person superseded the first? The latter alternative seems the likelier.

I call, first person singular.—The word call is not one person more than another. It is the simple verb, wholly uninflected. It is very probable that the first person was the {298}one where the characteristic termination was first lost. In the Modern Norse language it is replaced by the second: Jeg taler=I speak, Danish.

Thou callest, second person singular.—The final -t appears throughout the Anglo-Saxon, although wanting in Old Saxon. In Old High German it begins to appear in Otfrid, and is general in Notker. In Middle High German and New High German it is universal.—Deutsche Grammatik, i. 1041. 857.

He calleth, or he calls, third person singular.—The -s in calls is the -th in calleth, changed. The Norse form kallar either derives its -r from the -th by way of change, or else the form is that of the second person replacing the first.

Lufiað, Anglo-Saxon, first person plural.—The second person in the place of the first. The same in Old Saxon.

Lufiað, Anglo-Saxon, third person plural.—Possibly changed from -ND, as in sôkjand. More probably the second person.

Loven, Old English.—For all the persons of the plural. This form may be accounted for in three ways: 1. The -m of the Mœso-Gothic and High Old German became -n; as it is in the Middle and Modern German, where all traces of the original -m are lost. In this case the first person has replaced the other two. 2. The -nd may have become -n; in which case it is the third person that replaces the others. 3. The indicative form loven may have arisen out of a subjunctive one; since there was in Anglo-Saxon the form lufion, or lufian, subjunctive. In the Modern Norse languages the third person replaces the other two: Vi tale, I tale, de tale=we talk, ye talk, they talk.

§ 352. The person in -T.—Art, wast, wert, shalt, wilt. Here the second person singular ends, not in -st, but in -t. A reason for this (though not wholly satisfactory) we find in the Mœso-Gothic and the Icelandic.

In those languages the form of the person changes with the tense, and the second singular of the præterite tense of one conjugation is, not -s, but -t; as Mœso-Gothic, svôr=I swore, svôrt=thou swarest, gráip=I griped, gráipt=thou gripedst; Icelandic, brannt=thou burnest, gaft=thou {299}gavest. In the same languages ten verbs are conjugated like præterites. Of these, in each language, skal is one.

§ 353. Thou spakest, thou brakest, thou sungest.^[45]—In these forms there is a slight though natural anomaly. They belong to the class of verbs which form their præterite by changing the vowel of the present; as sing, sang, &c. Now, all words of this sort in Anglo-Saxon formed their second singular præterite, not in -st, but in -e; as þú funde=thou foundest, þú sunge=thou sungest. The English termination is derived from the present. Observe that this applies only to the præterites formed by changing the vowel. Thou loved'st is Anglo-Saxon as well as English, viz., þú lufodest.

§ 354. In the northern dialects of the Anglo-Saxon the -ð of plurals like lufiað=we love becomes -s. In the Scottish this change was still more prevalent:

CHAPTER XX.

§ 355. The inflection of the present tense, not only in Anglo-Saxon, but in several other languages as well, has been given in the preceding chapter. As compared with the present plural forms, we love, ye love, they love, both the Anglo-Saxon we lufiað, ge lufiað, hi lufiað, and the Old English we loven, ye loven, they loven, have a peculiar termination for the plural number which the present language wants. In other words, the Anglo-Saxon and the Old English have a plural personal characteristic, whilst the Modern English has nothing to correspond with it.

The word personal is printed in italics. It does not follow, that, because there is no plural personal characteristic, there is also no plural characteristic.

There is no reason against the inflection of the word love running thus—I love, thou lovest, he loves; we lave, ye lave, they lave; in other words, there is no reason against the vowel of the root being changed with the number. In such a case there would be no personal inflection, though there would be a plural, or a numeral, inflection.

Now, in Anglo-Saxon, with a great number of verbs such a plural inflection not only actually takes place, but takes place most regularly. It takes place, however, in the past tense only. And this is the case in all the Gothic languages as well as in Anglo-Saxon. Amongst the rest, in—

In all the Anglo-Saxon words, it may be remarked that the change is from a to u, and that both the vowels are short, or dependent. Also, that the vowel of the present tense is i short; as swim, sing, &c. The Anglo-Saxon form of run is yrnan.

In the following words the change is from the Anglo-Saxon á to the Anglo-Saxon ī. In English, the regularity of the change is obscured by a change of pronunciation.

From these examples the reader has himself drawn his inference; viz. that words like

generally called double forms of the past tense, were originally different numbers of the same tense, the forms in u, as swum, and the forms in i, bit, being plural.

CHAPTER XXI.

§ 357. Between the second plural imperative, and the second plural indicative, speak ye and ye speak, there is no difference of form. Between the second singular imperative speak, and the second singular indicative, speakest, there is a difference in form. Still, as the imperative form speak is distinguished from the indicative form speakest by the negation of a character rather than by the possession of one, it cannot be said that there is in English any imperative mood.

§ 358. If he speak, as opposed to if he speaks, is characterised by a negative sign only, and consequently is no true example of a subjunctive. Be, as opposed to am, in the sentence if it be so, is an uninflected word used in a limited sense, and consequently no true example of a subjunctive.

The only true subjunctive inflection in the English language is that of were and wert, as opposed to the indicative forms was and wast.

CHAPTER XXII.

§ 359. The nature of tenses in general is best exhibited by reference to the Greek; since in that language they are more numerous, and more strongly marked than elsewhere.

I strike, I struck.—Of these words, the first implies an action taking place at the time of speaking, the second marks an action that has already taken place.

These two notions of present and of past time, being expressed by a change of form, are true tenses. They are however, the only true tenses in our language. In I was beating, I have beaten, I had beaten, and I shall beat, a difference of time is expressed; but as it is expressed by a combination of words, and not by a change of form, no true tenses are constituted.

In Greek the case is different. Τύπτω (typtô)=I beat; ἔτυπτον (etypton)=I was beating; τύψω (typsô)=I shall beat; ἔτυψα (etypsa)=I beat; τέτυφα (tetyfa)=I have beaten; ἐτετύφειν (etetyfein)=I had beaten. In these words we have, of the same mood, the same voice, and the same conjugation, six different tenses;^[47] whereas, in English, there are but two. The forms τέτυφα and ἔτυψα are so strongly marked, that we recognise them wheresoever they occur. The first is formed by a reduplication of the initial τ, and, consequently, may be called the reduplicate form. As a tense it is called the perfect. In the form ἔτυψα an ε is prefixed, and an σ is added. In the allied language of Italy {304}the ε disappears, whilst the σ (s) remains. Ἔτυψα is said to be an aorist tense. Scripsi : scribo :: ἔτυψα : τύπτω.

§ 360. Now in the Latin language a confusion takes place between these two tenses. Both forms exist. They are used, however, indiscriminately. The aorist form has, besides its own, the sense of the perfect. The perfect has, besides its own, the sense of the aorist. In the following pair of quotations, vixi, the aorist form, is translated I have lived, while tetigit, the perfect form, is translated he touched.

When a difference of form has ceased to express a difference of meaning, it has become superfluous. This is the case with the two forms in question. One of them may be dispensed with; and the consequence is, that, although in the Latin language both the perfect and the aorist forms are found, they are, with few exceptions, never found in the same word. Wherever there is the perfect, the aorist is wanting, and vice versâ. The two ideas I have struck and I struck are merged into the notion of past time in general, and are expressed by one of two forms, sometimes by that of the Greek perfect, and sometimes by that of the Greek aorist. On account of this the grammarians have cut down the number of Latin tenses to five; forms like cucurri and vixi being dealt with as one and the same tense. The true view is, that in curro the aorist form is replaced by the perfect, and in vixi the perfect form is replaced by the aorist.

§ 361. In the present English there is no undoubted perfect or reduplicate form. The form moved corresponds in meaning not with τέτυφα and momordi, but with ἔτυψα and vixi. Its sense is that of ἔτυψα, and not that of τέτυφα. The notion given by τέτυφα we express by the circumlocution I have beaten. We have no such form as bebeat or memove. In the Mœso-Gothic, however, there was a true reduplicate form; in other words, a perfect tense as well as an aorist. It {305}is by the possession of this form that the verbs of the first six conjugations are characterized.

In Mœso-Gothic, as in Latin, the perfect forms have, besides their own, an aorist sense, and vice versâ.

In Mœso-Gothic, as in Latin, the two forms are dealt with as a single tense; láilô being called the præterite of láia, and svôr the præterite of svara. The true view, however, is that in Mœso-Gothic, as in Latin, there are two past tenses, each having a certain latitude of meaning, and each, in certain words, replacing the other.

The reduplicate form, in other words, the perfect tense, is current in none of the Gothic languages except the Mœso-Gothic. A trace of it is found in the Anglo-Saxon of the seventh century in the word heht, which is considered to be hê-ht, the Mœso-Gothic háiháit, vocavi. This statement is taken from the Cambridge Philological Museum, ii. 378. Did from do is also considered to be a reduplicate form.

§ 362. In the English language the tense corresponding with the Greek aorist and the Latin forms like vixi, is formed after two modes; 1, as in fell, sang, and took, from fall, sing, and take, by changing the vowel of the present: 2, as in moved and wept, from move and weep, by the addition of d or t; the d or t not being found in the original word, but being a fresh element added to it. In forms, on the contrary, like sang and fell, no addition being made, no new element appears. The {306}vowel, indeed, is changed, but nothing is added. Verbs, then, of the first sort, may be said to form their præterites out of themselves; whilst verbs of the second sort require something from without. To speak in a metaphor, words like sang and fell are comparatively independent. Be this as it may, the German grammarians call the tenses formed by a change of vowel the strong tenses, the strong verbs, the strong conjugation, or the strong order; and those formed by the addition of d or t, the weak tenses, the weak verbs, the weak conjugation, or the weak order. Bound, spoke, gave, lay, &c., are strong; moved, favoured, instructed, &c., are weak. For the proof that the division of verbs into weak and strong is a natural division, see the Chapter on Conjugation.

CHAPTER XXIII.

§ 363. The strong præterites are formed from the present by changing the vowel, as sing, sang, speak, spoke.

The first point in the history of these tenses that the reader is required to be aware of, is stated in the Chapter upon the Numbers, viz., that, in Anglo-Saxon, several præterites change, in their plural, the vowel of their singular; as

As a general rule, the second singular has the same vowel with the plural persons, as burne, thou burntest, plural burnon, we burnt.

The bearing of this fact upon the præterites has been indicated in p. 300. In a great number of words we have a double form, as ran and run, sang and sung, drank and drunk, &c. One of these forms is derived from the singular, and the other from the plural. I cannot say at what period the difference of form ceased to denote a difference of sense.

In cases where but one form is preserved, that form is not necessarily the singular one. For instance, Ic fand, I found, we fundon, we found, are the Anglo-Saxon forms. Now the present word found comes, not from the singular fand, but from the plural fund; although in the Lowland Scotch dialect and in the old writers, the singular form occurs.

Even in the present English it will be found convenient to {308}call the forms like sang and drank the singular, and those like sung and bound the plural forms.

Be it observed, that, though this fact accounts for most of our double forms, it will not account for all. In the Anglo-Saxon, Ic sprǽc, I spake, we sprǽcon, we spake. There is no change of number to account for the two forms spake and spoke.

§ 364. Contains the two words fall and fell, hold and held, where the sound of o is changed into that of ĕ. Here must be noticed the natural tendency of a to become o; since the forms in Anglo-Saxon are, Ic fealle, I fall; Ic feoll, I fell; Ic healde, I hold; Ic heold, I held.

§ 365. Here the præterite ends in -ew. Words of this class are distinguished from those of the third Class by the different form of the present tense.

In these words the w has grown out of a g, as may be seen from the Anglo-Saxon forms. The word see (saw) belongs to this class: since, in Anglo-Saxon, we find the forms geseáh and gesegen, and in the Swedish the præterite form is saag.

§ 366. Here an o before w, in the present, becomes e before w in the præterite; as

§ 367. Contains the single word let, where a short e in the {309}present remains unchanged in the præterite. In the Anglo-Saxon the present form was Ic læte, the præterite Ic lét.

§ 368. Contains the single word beat, where a long e remains unchanged. In Anglo-Saxon the forms were Ic beate, Ic beot.

§ 369. Present come, præterite came, participle come. In Anglo-Saxon, cume, com, cumen.

§ 370. In this class we have the sounds of the ee, in feet, and of the a in fate (spelt ea or a), changed into o or oo. As several words in this class have a second form in a, the præterite in o or oo will be called the primary, the præterite in a the secondary form.

The præterite of stand was originally long. This we collect {310}from the spelling, and from the Anglo-Saxon form stód. The process that ejects the nd is the same process that, in Greek, converts ὀδόντ-ος into ὀδούς.

§ 371. In this class the sound of the ee in feet, and the a in fate (spelt ea), is changed into a. Several words of this class have secondary forms. Further details may be seen in the remarks that come after the following list of verbs.

Here observe,—1. That in speak, cleave, steal, the ea has the same power with the ee in freeze and seethe; so that it may be dealt with as the long (or independent) sound of the i in bid, sit, give.

2. That the same view may be taken of the ea in break, although the word by some persons is pronounced brake. Gabrika, gabrak, Mœso-Gothic; briku, brak, Old Saxon; brece, brac, Anglo-Saxon. Also of bear, tear, swear, wear. In the provincial dialects these words are even now pronounced beer, teer, sweer. The forms in the allied languages are, in {311}respect to these last-mentioned words, less confirmatory; Mœso-Gothic, svara, báira; Old High German, sverju, piru.

3. That the ea in tread was originally long; Anglo-Saxon, tredan, trede, trǽd, treden.

4. Lie.—Here the sound is diphthongal, having grown out of the Anglo-Saxon forms licgan, lǽg, legen.

5. Sat.—The original præterite was long. This we collect from the spelling sate, and from the Anglo-Saxon sǽt.

§ 372. A, as in fate, is changed either into the o in note, or the oo in book. Here it should be noticed that, unlike break and swear, &c., there is no tendency to sound the a of the present as ee, neither is there, as was the case with clove and spoke, any tendency to secondary forms in a. A partial reason for this lies in the original nature of the vowel. The original vowel in speak was e. If this was the é fermé of the French, it was a sound from which the a in fate and the ee in feet might equally have been evolved. The vowel sound of the verbs of the present class was that of a for the present and that of ó for the præterite forms; as wace, wóc, grafe, gróf. Now of these two sounds it may be said that the a has no tendency to become the ee in feet, and that the ó has no tendency to become the a in fate.

The sounds that are evolved from the accentuated ó, are the o in note and the oo in book.

§ 373. Containing the single word strike, struck, stricken. It is only in the Middle High German, the Middle Dutch, the New High German, the Modern Dutch, and the English, that {312}this word is found in its præterite forms. These are, in Middle High German, streich; New High German, strich; Middle Dutch, strêc; Modern Dutch, strîk. Originally it must have been referable to the ninth class.

§ 374. In this class we first find the secondary forms accounted for by the difference of form between the singular and plural numbers. The change is from the i in bite to the o in note, and the i in pit. Sometimes it is from the i in bit to the a in bat. The Anglo-Saxon conjugation (A) may be compared with the present English (B).

1. That, with the exception of the word slit, the i is sounded as a diphthong.

2. That, with the exception of bat and slat, it is changed into o in the singular and into ĭ in the plural forms.

4. That, even with the word shone, the o was originally long. This is known from the final -e mute, and from the Anglo-Saxon form scéan; Mœso-Gothic, skáin; Old Norse, skein.

6. That the statement last made shows that even bat and slat were once in the same condition with arose and smote, the Anglo-Saxon forms being arás, smát, bát, slát.

§ 375. In this class i is generally short; originally it was always so. In the singular form it becomes ă, in the plural, ŭ.

1. That, with the exceptions of bind, find, grind, and wind, the vowels are short (or dependent) throughout.

2. That, with the exception of run and burst, the vowel of the present tense is either the i or e.

3. That i short changes into a for the singular, and into u for the plural forms.

4. That e changes into o in the singular forms; these being the only ones preserved.

5. That the i in bind, &c., changes into ou in the plural forms; the only ones current.

6. That the vowel before m or n is, with the single exception of run, always i.

7. That the vowel before l and r is, with the single exception of burst, always e.

8. That, where the i is sounded as in bind, the combination following is -nd.

9. That ng being considered as a modification of k (the Norse and Mœso-Gothic forms being drecka and drikjan), it may be stated that i short, in the twelfth class, precedes either a liquid or a mute of series k.

From these observations, even on the English forms only, we find thus much regularity; and from these observations, even on the English forms only, we may lay down a rule like the following: viz. that i or u, short, before the consonants m, n, {315}or ck, is changed into a for the singular, and into u for the plural forms; that i long, or diphthongal, becomes ou; that e before l becomes o; and that u before r remains unchanged.

This statement, however, is nothing like so general as the one that, after a comparison of the older forms and the allied languages, we are enabled to make. Here we are taught,

1. That, in the words bind, &c., the i was once pronounced as in till, fill; in other words, that it was the simple short vowel, and not the diphthong ey; or at least that it was treated as such.

When the vowel ĭ of the present took the sound of the i in bite, the ŭ in the præterite became the ou in mouse. From this we see that the words bind, &c., are naturally subject to the same changes with spin, &c., and that, mutatis mutandis, they are so still.

2. That the e in swell, &c., was once ĭ. This we collect from the following forms:—hilpa, Mœso-Gothic; hilfu, Old High German; hilpu, Old Saxon; hilpe, Middle High German; hilpe, Old Frisian. Suillu=swell, Old High German. Tilfu=delve, Old High German; dilbu, Old Saxon. Smilzu, Old High German=smelt or melt. This shows that originally the vowel i ran throughout, but that before l and r it was changed into e. This change took place at different periods in different dialects. The Old Saxon preserved the {316}i longer than the Anglo-Saxon. It is found even in the middle High German; in the new it has become e; as schwelle, schmelze. In one word milk, the original i is still preserved; although in Anglo-Saxon it was e; as melce, mealc=milked, mulcon. In the Norse the change from i to e took place full soon, as svëll=swells. The Norse language is in this respect important.

4. That a change between a and o took place by times. The Anglo-Saxon præterite of swelle is sweoll; whilst ongon, bond, song, gelomp, are found in the same language for ongan, band, sang, gelamp.—Rask's Anglo-Saxon Grammar, p. 90.

The Anglo-Saxon form is yrnan; in the præterite arn, urnon. A transposition has since taken place. The word run seems to have been originally no present, but a præterite form.

6. That burst is only an apparent exception. Before r, ĕ, ĭ, ŭ, are pronounced alike. We draw no distinction between the vowels in pert, flirt, hurt. The Anglo-Saxon forms are, berste, byrst, bærse, burston, borsten.

§ 376. Contains the single word choose, in the præterite chose; in Anglo-Saxon, ceóse, ceás.

CHAPTER XXIV.

§ 377. The præterite tense of the weak verbs is formed by the addition of -d or -t. If necessary, the syllable -ed is substituted for -d.

The current statement that the syllable -ed, rather than the letter -d, is the sign of the præterite tense, is true only in regard to the written language. In stabbed, moved, bragged, whizzed, judged, filled, slurred, slammed, shunned, barred, strewed, the e is a point of spelling only. In language, except in declamation, there is no second vowel sound. The -d comes in immediate contact with the final letter of the original word, and the number of syllables remains the same as it was before.

When, however, the original word ends in -d or -t, as slight or brand, then, and then only (and that not always), is there the addition of the syllable -ed; as in slighted, branded. This is necessary, since the combinations slightt and brandd are unpronounceable.

Whether the addition be -d or -t depends upon the flatness or sharpness of the preceding letter.

After b, v, th (as in clothe), g, or z, the addition is -d. This is a matter of necessity. We say stabd, môvd, clôthd, braggd, whizzd, because stabt, môvt, clotht, braggt, whizzt, are unpronounceable.

After l, m, n, r, w, y, or a vowel, the addition is also -d. This is the habit of the English language. Filt, slurt, strayt, &c., are as pronounceable as filld, slurrd, strayd, &c. It is the habit, however, of the English language to prefer the latter forms. All this, as the reader has probably observed, is merely the reasoning concerning the s, in words like {318}father's, &c., applied to another letter and to another part of speech.

For some historical notices respecting the use of -d, -t, and -ed, in the spelling of the English præterites and participles, the reader is referred to the Cambridge Philological Museum, vol. i. p. 655.

§ 378. The verbs of the weak conjugation fall into three classes. In the first there is the simple addition of -d, -t, or -ed.

To this class belong the greater part of the weak verbs and all verbs of foreign origin.

§ 379. In the second class, besides the addition of -t or -d, the vowel is shortened. It also contains those words which end in -d or -t, and at the same time have a short vowel in the præterite. Such, amongst others, are cut, cost, &c., where the two tenses are alike, and bend, rend, &c., where the præterite is formed from the present by changing -d into -t, as bent, rent, &c.

In the following list, the words ending in -p are remarkable; since, in Anglo-Saxon, each of them had, instead of a weak, a strong præterite.

In this class we sometimes find -t where the -d is expected; the forms being left and dealt, instead of leaved and dealed. {319}

§ 380. Third class.—In the second class the vowel of the present tense was shortened in the præterite. In the third class it is changed.

To this class belong the remarkable præterites of the verbs seek, beseech, catch, teach, bring, think, and buy, viz., sought, besought, caught, taught, brought, thought, and bought. In all these, the final consonant is either g or k, or else a sound allied to those mutes. When the tendency of these sounds to become h and y, as well as to undergo farther changes, is remembered, the forms in point cease to seem anomalous. In wrought, from work, there is a transposition. In laid and said the present forms make a show of regularity which they have not. The true original forms should be legde and sægde, the infinitives being lecgan, secgan. In these words the i represents the semivowel y, into which the original g was changed. The Anglo-Saxon forms of the other words are as follows:—

§ 381. Out of the three classes into which the weak verbs in Anglo-Saxon are divided, only one takes a vowel before the d or t. The other two add the syllables -te, or -de, to the last letter of the original word. The vowel that, in one out of the three Anglo-Saxon classes, precedes d is o. Thus we have lufian, lufode; clypian, clypode. In the other two classes the forms are respectively bærnan, bærnde; and tellan, tealde, no vowel being found. The participle, however, as stated above, ended, not in -de or -te, but in -d or -t; and in two out of the three classes it was preceded by a vowel, gelufod, bærned, geteald. Now in those conjugations where no vowel preceded the d of the præterite, and where the original word ended in -d or -t, a difficulty, which has already been indicated, arose. To add the sign of the præterite to a word like eard-ian (to dwell) was an easy matter, inasmuch as {320}eardian was a word belonging to the first class, and in the first class the præterite was formed in -ode. Here the vowel o kept the two d's from coming in contact. With words, however, like métan and sendan, this was not the case. Here no vowel intervened; so that the natural præterite forms were met-te, send-de, combinations wherein one of the letters ran every chance of being dropped in the pronunciation. Hence, with the exception of the verbs in the first class, words ending in -d or -t in the root admitted no additional d or t in the præterite. This difficulty, existing in the present English as it existed in the Anglo-Saxon, modifies the præterites of most words ending in -t or -d.

In several words there is the actual addition of the syllable -ed; in other words d is separated from the last letter of the original word by the addition of a vowel; as ended, instructed, &c. Of this e two views may be taken.

1. It may be derived from the original o in -ode, the termination of the first class in Anglo-Saxon. This is the opinion which we form when the word in question is known to have belonged to the Anglo-Saxon language, and, in it, to the first class. Ended, planted, warded, hated, heeded, are (amongst others) words of this sort; their Anglo-Saxon forms being endode, plantode, weardode, hatode, and eahtode, from endian, plantian, weardian, hatian, and eahtian.

2. The form may be looked upon, not as that of the præterite, but as that of the participle in a transferred sense. This is the view when we have two forms, one with the vowel, and the other without it, as bended and bent, wended and went, plighted and plight.

A. In several words the final -d is changed into -t, as bend, bent; rend, rent; send, sent; gild, gilt; build, built; spend, spent, &c.

B. In several words the vowel of the root is changed; as feed, fed; bleed, bled; breed, bred; meet, met; speed, sped; rēad, rĕad, &c. Words of this last-named class cause occasional difficulty to the grammarian. No addition is made to the root, and, in this circumstance, they agree with the strong verbs. Moreover, there is a change of the vowel. {321}In this circumstance also they agree with the strong verbs. Hence with forms like fed and led we are in doubt as to the conjugation. This doubt we have three means of settling, as may be shown by the word beat.

a. By the form of the participle.—The -en in beaten shows that the word beat is strong.

b. By the nature of the vowel.—The weak form of to beat would be bet, or beăt, after the analogy of feed and rēad. By some persons the word is pronounced bet, and with those who do so the word is weak.

c. By a knowledge of the older forms.—The Anglo-Saxon form is beáte, beot. There is no such a weak form as beáte, bætte. The præterite of sendan is sende, weak. There is in Anglo-Saxon no such form as sand, strong.

In all this we see a series of expedients for separating the præterite form from the present, when the root ends with the same sound with which the affix begins.

The change from a long vowel to a short one, as in feed, fed, &c., can only take place where there is a long vowel to be changed.

Where the vowels are short, and, at the same time, the word ends in -d, the -d of the present may become -t in the præterite. Such is the case with bend, bent.

When there is no long vowel to shorten, and no -d to change into -t, the two tenses, of necessity, remain alike; such is the case with cut, cost, &c.

Words like planted, heeded, &c., belong to the first class. Words like feed, lead, to the second class. Bend and cut belong also to the second class; they belong to it, however, by what may be termed an etymological fiction. The vowel would be changed if it could.

§ 382. Made, had.—In these words there is nothing remarkable but the ejection of a consonant. The Anglo-Saxon forms are macode and hæfde, respectively. The words, however, in regard to the amount of change, are not upon a par. The f in hæfde was probably sounded as v. Now v {322}is a letter excessively liable to be ejected, which k is not. K, before it is ejected, is generally changed into either g or y.

Would, should, could.—It must not be imagined that could is in the same predicament with these words. In will and shall the -l is part of the original word. This is not the case with can. For the form could, see the Chapter upon Irregularity.

Aught.—In Anglo-Saxon áhte, the præterite of the present form áh, plural ágan.—As late as the time of Elizabeth we find owe used for own. The present form own seems to have arisen from the plural ágen. Aught is the præterite of the Anglo-Saxon áh; owed of the English owe=debeo; owned of the English own=possideo. The word own, in the expression to own to a thing, has a totally different origin. It comes from the Anglo-Saxon an (plural, unnon)=I give, or grant=concedo.

Durst.—The verb dare is both transitive and intransitive. We can say either I dare do such a thing, or I dare (challenge) such a man to do it. This, in the present tense, is unequivocally correct. In the past the double power of the word dare is ambiguous; still it is, to my mind at least, allowable. We can certainly say I dared him to accept my challenge; and we can, perhaps, say I dared venture on the expedition. In this last sentence, however, durst is the preferable expression.

Now, although dare is both transitive and intransitive, durst is only intransitive. It never agrees with the Latin word provoco; only with the Latin word audeo. Moreover, the word durst has both a present and a past sense. The difficulty which it presents consists in the presence of the -st, letters characteristic of the second person singular, but here found in all the persons alike; as I durst, they durst, &c.

The Mœso-Gothic forms are dar, dart? dar, daúrum, daúruþ, daúrun, for the persons of the present tense; and daúrsta, daúrstês, daúrsta, &c., for those of the præterite. The same is the case throughout the Germanic languages. No -s, however, appears in the Scandinavian; the præterites being þorði and törde, Icelandic and Danish. The Anglo-Saxon is dear=I dare, dearst=thou darest, durron=we, {323}ye, or they dare; subjunctive, durre, dorste, dorston. Old Saxon, present, dar; præterite dursta. The Mœso-Gothic tense, daúrsta, instead of daúrda, shows the antiquity of this form in -s.

The readiest mode of accounting for the form in question is to suppose that the second singular has been extended over all the other persons. This view, however, is traversed by the absence of the -s in the Mœso-Gothic present. The form there (real or presumed) is not darst, but dart. Of this latter form, however, it must be remarked that its existence is hypothetical.

In Matthew xxvi. 67, of the Mœso-Gothic Gospel of Ulphilas, is found the form kaúpastêdun, instead of kaúpatidédun, the præterite plural of kaúpatjan=to beat. Here there is a similar insertion of the -s.—Deutsche Grammatik, i. 848, 852, 853.

Must.—A form common to all persons, numbers, and tenses. That neither the -s nor the -t are part of the original root, is indicated by the Scandinavian form maae (Danish), pronounced moh; præterite maatte.

The readiest mode of accounting for the -s in must, is to presume that it belongs to the second singular, extended to the other persons, mo-est=must. Irrespective, however, of other objections, this view is traversed by the forms môtan, Mœso-Gothic (an infinitive), and mót, Mœso-Gothic, Old Saxon, and Anglo-Saxon (a first person present). These neutralise the evidence given by the Danish form maae, and indicate that the -t is truly a part of the original root.

Now, the -t being considered as part of the root, the -s cannot be derived from the second singular; inasmuch as it precedes, instead of following the -t.

At one time, for want of a better theory, I conceived, that in the word in point (and also in durst and a few others), we had traces of the Scandinavian passive. This notion I have, for evident reasons, abandoned.

In p. 298 it was stated that the Mœso-Gothic termination of the second singular of the strong præterites was -t. It is {324}here mentioned that must is a præterite form. Now the final letter of the root mot, and the sign of the second singular of the strong præterite, are the same, -t. Now, as -t cannot be immediately added to t, the natural form of the second singular mót-t is impracticable. Hence, before the -t of the second person, the -t of the root is changed, so that, instead of máimáit-t, bigat-t, fáifalþ-t, láilot-t, &c., we have máimáis-t, bigas-t, fáifals-t, láilos-t, &c., Mœso-Gothic.—See Deutsche Grammatik, 844.

The euphonic reason for the -s, in must, is sufficient to show that it is in a different predicament from durst.

The provincial form mun, there or thereabouts equivalent in meaning to must, has no etymological connexion with this last named word. It is a distinct word, in Scandinavian monne.

Wist.—In its present form a regular præterite from wiss=know. The difficulties of this word arise from the parallel forms wit (as in to wit), and wot=knew. The following are the forms of this peculiar word:—

In Mœso-Gothic, 1 sing. pres. ind. váit; 2. do., váist; 1. pl. vitum; præterite 1. s. vissa; 2 vissêss; 1. pl. vissêdum. From the form váist we see that the second singular is formed after the manner of must; that is, váist stands instead of váit-t. From the form vissêdum we see that the præterite is not strong, but weak; therefore that vissa is euphonic for vista.

In Anglo-Saxon.—Wât, wâst, witon, wiste and wisse, wiston.—Here the double forms, wiste and wisse, verify the statement concerning the Mœso-Gothic vissa.

In Icelandic.—Veit, veizt, vitum, vissi. Danish ved, vide, vidste. Observe the form vidste; since, in it, the -d of the root (in spelling, at least), is preserved. The -t of the Anglo-Saxon wiste is the -t, not of the root, but of the inflection.

In respect to the four forms in question, viz., wit, wot, wiss, wist; the first seems to be the root; the second a strong præterite regularly formed, but used (like οἶδα in Greek) with a present sense; the third a weak præterite, of which the -t has been ejected by a euphonic process, used also with a {325}present sense; the fourth is a second singular from wiss after the manner of wert from were, a second singular from wit after the manner of must, a secondary præterite from wiss, or finally, the form wisse, anterior to the operation of the euphonic process that ejected the -t.

Do.—In the phrase this will do=this will answer the purpose, the word do is wholly different from the word do, meaning to act. In the first case it is equivalent to the Latin valere; in the second to the Latin facere. Of the first the Anglo-Saxon inflection is deáh, dugon, dohte, dohtest, &c. Of the second it is dó, dóð, dyde, &c. I doubt whether the præterite did, as equivalent to valebat=was good for, is correct. In the phrase it did for him=it finished him, either meaning may be allowed.

In the present Danish they write duger, but say duer: as duger et noget?=Is it worth anything? pronounced dooer deh note? This accounts for the ejection of the g. The Anglo-Saxon form deah does the same.

Is the word weak?—This is the view that arises from the form did. The participle done traverses this view.

Is the word strong?—In favour of this notion we have the English participle done, and the præterite second singular in Old High German tâti. Against it are the Old Saxon dédos, and the Anglo-Saxon dydest, as second singulars.

Is there a reduplication?—If this were the case, we might assume such a form as dôan, dáidô, for the Mœso-Gothic. This view, however, is traversed by the substantival forms dêds, Mœso-Gothic; tât, Old High German; dæd, Anglo-Saxon; which show that the second -d is part of the original word.

The true nature of the form did has yet to be exhibited.—See Deutsche Grammatik, i. 1041.

Mind—mind and do so and so.—In this sentence the word mind is wholly different from the noun mind. The Anglo-Saxon forms are geman, gemanst, gemunon, without the -d; this letter occurring only in the præterite tense (gemunde, {326}gemundon), of which it is the sign. Mind is, then, a præterite form with a present sense; whilst minded (as in he minded his business) is an instance of excess of inflection; in other words, it is a præterite formed from a præterite.

A præterite formed upon a præterite may also be called a secondary præterite; just as the word theirs, derived from their (a case formed from a case), is called a secondary genitive.

In like manner the present form mind is not a genuine present, but a præterite with a present sense; its form being taken as the test. Presents of this sort may be called transformed præterites.

It is very evident that the præterites most likely to become present are those of the strong class. In the first place, the fact of their being præterite is less marked. The word tell carries with it fewer marks of its tense than the word moved. In the second place they can more conveniently give rise to secondary præterites. A weak præterite already ends in -d or -t. If this be used as a present, a second -d or -t must be appended.

Hence it is that all the transposed præterites in the Gothic tongues were, before they took the present sense, not weak, but strong. The word in question, mind (from whence minded), is only an apparent exception to this statement.

Now the words shall, can, owe (whence aught), dare, may, man (of the Anglo-Saxon geman, the origin of mind), are, (irrespective of their other peculiarities), for certain etymological reasons, looked upon as præterite forms with a present sense.

And the words should, could, aught, dared (or durst), must, wist, might, mind, are, for certain etymological reasons, looked upon as secondary præterites.

This fact alters our view of the form minded. Instead of being a secondary præterite, it is a tertiary one. Geman (the apparent present) being dealt with as a strong præterite with a present sense, mind (from the Anglo-Saxon gemunde) is the secondary præterite, and minded (from the English mind) is a tertiary præterite. To analyse the word, the {327}præterite is first formed by the vowel a, then by the addition of -d, and, thirdly, by the termination -ed; man, mind, minded.

The proof of this we collect from the second persons singular, Mœso-Gothic. The second singular præterite of the strong class is -t; of the weak class, -es; of the present, both weak and strong, -s. Now the second singular of the words in point is skal-t, kan-t, áih-t, dar-t? mag-t, man-t, respectively.—Deutsche Grammatik, i. 852.

Besides this, in Anglo-Saxon, the plural forms are those of the strong præterites. See Rask, p. 79.

Yode.—The obsolete præterite of go, now replaced by went, the præterite of wend. Regular, except that the initial g has become y.

CHAPTER XXV.

§ 383. The current statement respecting verbs like sing and fall, &c., is that they are irregular. How far this is the case may be seen from a review of the twelve classes in Mœso-Gothic, where the change of the vowel is subject to fewer irregularities than elsewhere. In the first six conjugations the præterite is replaced by a perfect tense. Consequently, there is a reduplication. Of these the fifth and sixth superadd to the reduplication a change of the vowel.

Exhibited in a tabular form, the changes of the vowels in Mœso-Gothic are as follows:—

§ 384. Such is the arrangement of the strong verbs in Mœso-Gothic, with which the arrangement of the strong verbs in the other Gothic languages may or may not coincide.

For a full and perfect coincidence three things are necessary:—1. the coincidence of form; 2. the coincidence of distribution; 3. the coincidence of order.

1. Coincidence of form..—Compared with the Mœso-Gothic rinna, rann, runnum, runnans, the Old High German inflection coincides most rigidly; e.g., rinnu, ran, runnumês, runnanê. The vowel is the same in the two languages, and it is similarly changed in each. It is very evident that this might be otherwise. The Mœso-Gothic i might have become e, or the u might have become o. In this case, the formula for the two languages would not have been the same. Instead of i, a, u, u (see the tabular arrangement), serving for the Old High German as well as the Mœso-Gothic, the formula would have been, for the Mœso-Gothic, i, a, u, u, and for the Old High German e, a, u, u, or i, a, o, o. The forms in this latter case would have been equivalent, but not the same.

2. Coincidence of distribution.—A given number of words in the Mœso-Gothic form their præterites by changing i into a; in other words, a given number of verbs in Mœso-Gothic are inflected like rinna and rann. The same is the case with the Old High German. Now if these words are the same in the two languages, the Mœso-Gothic and the Old High German (as far as the agreement extends) coincide in the distribution of their verbs; that is, the same words are arranged in the same class, or (changing the phrase) are distributed alike.

3. Coincidence of order.—The conjugation to which the Mœso-Gothic words rinna and rann belong is the twelfth. The same is the case in Old High German. It might, {330}however, have been the case that in Old High German the class corresponding with the twelfth in Mœso-Gothic was the first, second, third, or any other.

Now a coincidence of form, a coincidence of distribution, and a coincidence of order, in all the classes of all the Gothic languages, is more than can be expected. If such were the case, the tenses would be identical throughout.

Coincidence of form is infringed upon by the simple tendency of sounds to change. Hilpa in Mœso-Gothic is helpe in Anglo-Saxon: hulpans in Mœso-Gothic is holfanêr in Old High German, and holpen in Anglo-Saxon. A change, however, of this sort is insufficient to affect the arrangement. Helpan, in Anglo-Saxon, is placed in the same class with spinnan; and all that can be said is, that the Mœso-Gothic i is, in Anglo-Saxon, represented not by i exclusively, but sometimes by i and sometimes by ĕ.

Coincidence of distribution is of great etymological importance. A word may in one stage of a language take the form of one conjugation, and in another that of another. The word climban is, in Anglo-Saxon, placed in the same conjugation with drincan, &c. For this there was a reason; viz., the fact of the i being short. For the i being short there was a reason also. The b preceded the vowel a, and consequently was sounded. This was the case whether the word was divided clim-ban or climb-an. An, however, was no part of the original word, but only the sign of the infinitive mood. As such it became ejected. The letter b then came at the end of the word; but as the combination mb, followed by nothing was unstable, b was soon lost in pronunciation. Now b being lost, the vowel which was once short became lengthened, or rather it became the sound of the diphthong ei; so that the word was no longer called clĭmb, but clime. Now the words that follow the analogy of spin, span ,&c. (and consequently constitute the twelfth class), do so, not because the vowel is i, but because it is a short i; and when the i is sounded like a diphthong, the præterite is formed differently. The Anglo-Saxon præterite of climban was sounded clŏmm, and rhymed to from; the English præterite (when strong) of {331}climb is sounded clōmbe, rhyming to roam. The word climb, which was once classed with spin and sing, is now to be classed with arise and smite; in other words, it is distributed differently.

Coincidence in the order of the classes is violated when a class which was (for instance) the third in one language becomes, in another language the fourth, &c. In Mœso-Gothic the class containing the words smeita, smáit, smitum, smitans, is the eighth. This is a natural place for it. In the class preceding it, the vowel is the same in both numbers. In the classes that follow it, the vowel is changed in the plural. The number of classes that in Mœso-Gothic change the vowel is five; viz., the eighth, ninth, tenth, eleventh, and twelfth. Of these the eighth is the first. The classes where the change in question takes place form a natural subdivision, of which the eighth class stands at the head. Now in Anglo-Saxon the vowel is not changed so much as in the Mœso-Gothic. In words like choose, give, and steal, the vowel remains unaltered in the plural. In Mœso-Gothic, however, these words are, respectively, of the ninth, tenth, and eleventh classes. It is not till we get to the eleventh that the Anglo-Saxon plurals take a fresh vowel. As the presence or absence of a change of vowel naturally regulates the order of the classes, the eighth class in Mœso-Gothic becomes the eleventh in Anglo-Saxon. If it were not so, the classes where a change took place in the plural would be separated from each other.

The later the stage of the language, the less complete the coincidence in the classes.

Of the present arrangement, the twelfth class coincides most throughout the Gothic languages.

In the word climb, a reason was given for its having changed from the twelfth class to the eleventh class. This, in the present state of our knowledge, cannot always be done.

These statements are made lest the reader should expect to find between the English and the Anglo-Saxon classification anything more than a partial coincidence. A detailed exhibition of the English conjugations would form a work of {332}itself. Moreover, the present classes of the strong verbs must, to a great degree, be considered as provisional.

Observe, that it is the classes of the strong verbs that are provisional. With the great divisions into weak and strong, the case is far otherwise. The general assertions which will be made in p. 333, respecting the strong conjugation, show most cogently that the division is a natural one.

§ 385. Preliminary, however, to making them, the reader's attention is directed to the following list of verbs. In the present English they all form the præterite in -d or -t; in Anglo-Saxon, they all form it by a change of the vowel. In other words they are weak verbs that were once strong.

§ 386. The first of the general statements made concerning strong verbs, with a view of proving that the order is natural, shall be the one arising out of the preceding list of præterites.

IV. No new word is ever, upon its importation, inflected according to the strong conjugation. It is always weak. As early as A.D. 1085, the French word adouber=to dubb, was introduced into English. Its præterite was dubbade.^[56]

V. All derived words are inflected weak. The intransitive forms drink and lie, are strong; the transitive forms drench and lay, are weak.

The fourth statement will again be recurred to. The present object is to show that the division into strong and weak is natural.

§ 387. Obsolete forms.—Instead of lept, slept, mowed, snowed, &c., we find, in the provincial dialects and in the older writers, the strong forms lep, step, mew, snew, &c. This is no more than what we expect. Here there are two forms, and each form is of a different conjugation.

§ 388. Double Forms.—In lep and mew we have two forms, of which one only is current. In swoll and swelled, in clomb and climbed, and in hung and hanged, we have two forms, of which both are current. These latter are true double forms. Of double forms there are two kinds.

1. Those like swoll and swelled; where there is the same tense, but a different conjugation.

2. Those like spoke and spake; where the tense is the same and the conjugation the same, but where the form is different.

The bearings of these double forms (which, however, are points of general rather than of English grammar) are as follows. Their number in a given language may be very great, and the grammarian of a given language may call them, not double forms of the same tense, but different tenses. Let the number of words like swoll and swelled be multiplied by 1000. The chances are, that, in the present state of etymology, they would be called first præterites and second præterites. The bearing of this remark upon the so-called aorists and futures of the Greek language is evident. I think that a writer in the Cambridge Philological Museum^[57] indicates the true nature of those tenses. They are the same tense in a different conjugation, and differ from swoll and swelled only in the frequency of their occurrence.

Difference of form, and difference of conjugation, may each simulate a difference of tense.

CHAPTER XXVI.

§ 389. In § 361 the distinction between irregularity and defectiveness was slightly foreshadowed. In pp. 243, 267, it was exhibited in its principles. In the present chapter the difference is more urgently insisted on.

The words that have hitherto served as illustrations are the personal pronouns I and me, and the adjectives good, better, and best. See the sections referred to above.

The view of these words was as follows: viz., that none of them were irregular, but that they were all defective. Me wanted the nominative, I the oblique cases. Good was without a comparative, better and best had no positive degree.

Now me and better may be said to make good the defectiveness of I and good; and I and good may be said to replace the forms wanting in me and better. This gives us the principle of compensation. To introduce a new term, I and me, good and better, may be said to be complementary to each other.

What applies to nouns applies to verbs also. Go and went are not irregularities. Go is (at least in the present stage of our language) defective in the past tense. Went (at least in its current sense) is without a present. The two words, however, compensate their mutual deficiencies, and are to each other complementary.

The distinction between defectiveness and irregularity, is the first instrument of criticism for coming to true views concerning the proportion of the regular and irregular verbs.

The second instrument of criticism in determining the irregular verbs, is the meaning that we attach to terms. {336}

It is very evident that it is in the power of the grammarian to raise the number of etymological irregularities to any amount, by narrowing the definition of the word irregular; in other words, by framing an exclusive rule. The current rule of the common grammarians is that the præterite is formed by the addition of -t, or -d, or -ed. Now this position is sufficiently exclusive; since it proscribes not only the whole class of strong verbs, but also words like bent and sent, where -t exists, but where it does not exist as an addition. The regular forms, it may be said, should be bended and sended.

Exclusive, however, as the rule in question is, it is plain that it might be made more so. The regular forms might, by the fiat of a rule, be restricted to those in -d. In this case words like wept and burnt would be added to the already numerous list of irregulars.

Finally, a further limitation might be made, by laying down as a rule that no word was regular, unless it ended in -ed.

Thus much concerning the modes of making rules exclusive, and, consequently, of raising the amount of irregularities. This is the last art that the philosophic grammarian is ambitious of acquiring. True etymology reduces irregularity by making the rules of grammar, not exclusive, but general. The quantum of irregularity is in the inverse proportion to the generality of our rules. In language itself there is no irregularity. The word itself is only another name for our ignorance of the processes that change words; and, as irregularity is in the direct proportion to the exclusiveness of our rules, the exclusiveness of our rules is in the direct proportion to our ignorance of etymological processes.

The explanation of some fresh terms will lead us towards (but not to) the definition of the word irregular.

I. Vital and obsolete processes.—The word moved is formed from move, by the addition of -d. The addition of -d is the process by which the present form is rendered præterite. The word fell is formed from fall, by changing a into e. The change of vowel is the process by which the present form is {337}rendered præterite. Of the two processes the result is the same. In what respect do they differ?

For the sake of illustration, let a new word be introduced into the language. Let a præterite tense of it be formed. This præterite would be formed, not by changing the vowel, but by adding -d. No new verb ever takes a strong præterite. The like takes place with nouns. No new substantive would form its plural, like oxen or geese, by adding -en, or by changing the vowel. It would rather, like fathers and horses, add the lene sibilant.

Now, the processes that change fall, ox, and goose into fell, oxen, and geese, inasmuch as they cease to operate on the language in its present stage, are obsolete processes; whilst those that change move into moved, and horse into horses, operating on the language in its present stage, are vital processes.

A definition of the word irregular might be so framed as to include all words whose forms could not be accounted for by the vital processes. Such a definition would, in the present English, make words like bent, sought, &c. (the euphonic processes being allowed for), regular, and all the strong verbs irregular.

The very fact of so natural a class as that of the strong verbs being reduced to the condition of irregulars, invalidates such a definition as this.

II. Processes of necessity as opposed to processes of habit.—The combinations -pd-, -fd-, -kd-, -sd-, and some others, are unpronounceable. Hence words like step, quaff, back, kiss, &c., take after them the sound of -t: stept, quafft, &c. (the sound being represented), being their præterites, instead of stepd, quaffd. Here the change from -d (the natural termination) to -t is a matter (or process) of necessity. It is not so with words like weep and wept, &c. Here the change of vowel is not necessary. Weept might have been said if the habit of the language had permitted.

A definition of the word irregular might be so framed as to include all words whose natural form was modified by any euphonic process whatever. In this case stept (modified by a {338}process of necessity), and wept (modified by a process of habit), would be equally irregular.

A less limited definition might account words regular as long as the process by which they are deflected from their natural form was a process of necessity. Those, however, which were modified by a process of habit it would class with the irregulars.

Definitions thus limited arise from ignorance of euphonic processes, or rather from an ignorance of the generality of their operation.

III. Ordinary processes as opposed to extraordinary processes.—The whole scheme of language is analogical. A new word introduced into a language takes the forms of its cases or tenses, &c., from the forms of the cases or tenses, &c., of the old words. The analogy is extended. Now few forms (if any) are so unique as not to have some others corresponding with them; and few processes of change are so unique as not to affect more words than one. The forms wept and slept correspond with each other. They are brought about by the same process; viz. by the shortening of the vowel in weep and sleep. The analogy of weep is extended to sleep, and vice versâ. Changing our expression, a common influence affects both words. The alteration itself is an ultimate fact. The extent of its influence is an instrument of classification. When processes affect a considerable number of words, they may be called ordinary processes; as opposed to extraordinary processes, which affect one or few words.

When a word stands by itself, with no other corresponding to it, we confess our ignorance, and say that it is affected by an extraordinary process, by a process peculiar to itself, or by a process to which we know nothing similar.

A definition of the word irregular might be so framed as to include all words affected by extraordinary processes; the rest being considered regular.

IV. Positive processes as opposed to ambiguous processes.—The words wept and slept are similarly affected. Each is changed from weep and sleep respectively; and we know that {339}the process which affects the one is the process that affects the other also. Here there is a positive process.

Reference is now made to words of a different sort. The nature of the word worse is explained in p. 267, and the reader is referred to the section. There the form is accounted for in two ways, of which only one can be the true one. Of the two processes, each might equally have brought about the present form. Which of the two it was, we are unable to say. Here the process is ambiguous.

A definition of the word irregular might be so framed as to include all words affected by ambiguous processes.

V. Normal processes as opposed to processes of confusion.—Let a certain word come under class A. Let all words under class A be similarly affected. Let a given word come under class A. This word will be affected even as the rest of class A is affected. The process affecting, and the change resulting, will be normal, regular, or analogical.

Let, however, a word, instead of really coming under class A, appear to do so. Let it be dealt with accordingly. The analogy then is a false one. The principle of imitation is a wrong one. The process affecting is a process of confusion.

Examples of this (a few amongst many) are words like songstress, theirs, minded, where the words songstr-, their-, and mind-, are dealt with as roots, which they are not.

Ambiguous processes, extraordinary processes, processes of confusion—each, or all of these are legitimate reasons for calling words irregular. The practice of etymologists will determine what definition is most convenient.

With extraordinary processes we know nothing about the word. With ambiguous processes we are unable to make a choice. With processes of confusion we see the analogy, but, at the same time, see that it is a false one.

§ 390. Could.—With all persons who pronounce the l this word is truly irregular. The Anglo-Saxon form is cuðe. The -l is inserted by a process of confusion.

Can, cunne, canst, cunnon, cunnan, cuðe, cuðon, cuð—such are the remaining forms in Anglo-Saxon. None of them {340}account for the -l. The presence of the -l makes the word could irregular. No reference to the allied languages accounts for it.

Notwithstanding this, the presence of the -l is accounted for. In would and should the -l has a proper place. It is part of the original words, will and shall. A false analogy looked upon could in the same light. Hence a true irregularity; provided that the L be pronounced.

The L, however, is pronounced by few, and that only in pursuance to the spelling. This reduces the word could to an irregularity, not of language, but only of orthography.

That the mere ejection of the -n in can, and that the mere lengthening of the vowel, are not irregularities, we learn from a knowledge of the processes that convert the Greek ὀδόντος (odontos) into ὀδούς (odows).

§ 391. The verb quoth is truly defective. It is found in only one tense, one number, and one person. It is the third person singular of the præterite tense. It has the further peculiarity of preceding its pronoun. Instead of saying he quoth, we say quoth he. In Anglo-Saxon, however, it was not defective. It was found in the other tenses, in the other number, and in other moods. Ic cweðe, þu cwyst, he cwyð. Ic cwæð, þú cwæðe, he cwæð, we cwædon, ge cwædon, hi cwædon. Imperative, cweð. Participle, gecweden. In the Scandinavian it is current in all its forms. There, however, it means, not to speak but to sing. As far as its conjugation goes, it is strong. As far as its class goes, it follows the form of speak, spoke. Like speak, its Anglo-Saxon form is in æ, as cwæð. Like one of the forms of speak, its English form is in o, as quoth, spoke.

The whole of the present chapter is indicative of the nature of irregularity, and of the elements that should enter into the definition of it, rather than exhaustive of the detail.

The principle that I recognise for myself is to consider no word irregular unless it can be proved so. This view includes the words affected by ambiguous processes, and by processes of confusion, and no others. The words affected by {341}extraordinary processes form a provisional class, which a future increase of our etymological knowledge may show to be regular. Worse and could (its spelling being considered) are the fairest specimens of our irregulars. The class, instead of filling pages, is exceedingly limited.

CHAPTER XXVII.

§ 392. Meseems.—Equivalent to it seems to me; mihi videtur, φαίνεταί μοι. The verb seems is intransitive; consequently the pronoun me has the power of a dative case. The pronoun it is not required to accompany the verb.

§ 393. Methinks.—In Anglo-Saxon there are two forms; þencan=to think, and þincan=to seem. It is from the latter form that the verb in methinks comes. Such being the case, it is intransitive, and consequently the pronoun me has the power of a dative case. The pronoun it is not required to accompany the verb.

§ 394. Me listeth, or me lists.—Equivalent to it pleases me=me juvat. Anglo-Saxon lystan=to wish, to choose, also to please, to delight; Norse, lysta. Unlike the other two, the verb is transitive, so that the pronoun me has the power of an accusative case. The pronoun it is not required to accompany the verb.

These three are the only true impersonal verbs in the English language. They form a class by themselves, because no pronoun accompanies them, as is the case with the equivalent expressions it appears, it pleases, and with all the other verbs in the language.

In the old language impersonal verbs, or rather the impersonal use of verbs, was commoner than at present.

Strictly speaking, the impersonal verbs are a part of syntax rather than of etymology.

CHAPTER XXVIII.

§ 395. The verb substantive is generally dealt with as an irregular verb. This is inaccurate. The true notion is that the idea of being or existing is expressed by four different verbs, each of which is defective in some of its parts. The parts, however, that are wanting in one verb, are made up by the inflections of one of the others. There is, for example, no præterite of the verb am, and no present of the verb was. The absence, however, of the present form of was is made up by the word am, and the absence of the præterite form of am is made up by the word was.

§ 396. Was.—Defective, except in the præterite tense, where it is found both in the indicative and conjunctive.

In the older stages of the Gothic languages the word has both a full conjugation and a regular one. In Anglo-Saxon it has an infinitive, a participle present, and a participle past. In Mœso-Gothic it is inflected throughout with -s; as visa, vas, vêsum, visans. In that language it has the power of the Latin maneo = to remain. The -r first appears in the Old High German; wisu, was, wârumês, wësaner. In Norse the s entirely disappears, and the word is inflected with r throughout; vera, var, vorum, &c.

§ 397. Be.—Inflected in Anglo-Saxon throughout the present tense, both indicative and subjunctive; found also as an {345}infinitive beón, as a gerund to beonne, and as a participle beonde. In the present English its inflection is as follows:—

The line in Milton beginning If thou beest he—(P. L. b. ii.), leads to the notion that the antiquated form beest is not indicative, but conjunctive. Such, however, is not the case: býst in Anglo-Saxon is indicative, the conjunctive form being beó.—And every thing that pretty bin (Cymbeline).—Here the word bin is the conjunctive plural, in Anglo-Saxon béon; so that the words every thing are to be considered equivalent to the plural form all things. The phrase in Latin would stand thus, quotquot pulcra sint; in Greek thus, ἁ ἄν κάλα ᾖ. The indicative plural is, in Anglo-Saxon, not beón, but beóð and beó.

§ 398. In the Deutsche Grammatik, i. 1051, it is stated that the Anglo-Saxon forms beó, bist, bið, beoð, or beó, have not a present, but a future sense; that whilst am means I am, beó means I shall be; and that in the older languages it is only where the form am is not found that be has the power of a present form. The same root occurs in the Slavonic and Lithuanic tongues with the same power; as, esmi=I am; búsu=I shall be, Lithuanic.—Esmu=I am; buhshu=I shall be, Livonic.—Jesm=I am; budu=I shall be, Slavonic.—Gsem=I am; budu=I shall be, Bohemian. This, however, proves, not that there is in Anglo-Saxon a future tense (or form), but that the word beó has a future sense. There is no fresh tense where there is no fresh form.

The following is a specimen of the future power of beón in Anglo-Saxon:—"Hi ne beóð na cílde, soðlice, on domesdæge, ac beóð swa micele menn swa swa hi, migton beón gif hi full weoxon on gewunlicre ylde."—Ælfric's Homilies. "They will not be children, forsooth, on Domesday, but will be as much {346}(so muckle) men as they might be if they were full grown (waxen) in customary age."

§ 399. If we consider the word beón like the word weorðan (see below) to mean not so much to be as to become, we get an element of the idea of futurity. Things which are becoming anything have yet something further to either do or suffer. Again, from the idea of futurity we get the idea of contingency, and this explains the subjunctive power of be. In English we often say may for shall, and the same was done in Anglo-Saxon.—"Ic ðe secge, heò is be ðam húse ðe Fegor hátte, and nán man nis ðe hig wíte (shall, may know) ær ðám myclan dóme."—Ælfric's Homilies, 44.

§ 400. Am.—Of this form it should be stated, that the letter -m is no part of the original word. It is the sign of the first person, just as it is in all the Indo-European languages.

It should also be stated, that, although the fact be obscured, and although the changes be insufficiently accounted for, the forms am, art, are, and is, are not, like am and was, parts of different words, but forms of one and the same word; in other terms, that, although between am and be there is no etymological connexion, there is one between am and is. This we collect from the comparison of the Indo-European languages.

In English and Anglo-Saxon the word is found in the {347}present indicative only. In English it is inflected through both numbers; in Anglo-Saxon in the singular number only. The Anglo-Saxon plurals are forms of the German seyn, a verb whereof we have, in the present English, no vestiges.

Worth.—In the following lines of Scott, the word worth=is, and is a fragment of the regular Anglo-Saxon verb weorðan=to be, or to become; German, werden.

CHAPTER XXIX.

§ 401. The present participle, called also the active participle and the participle in -ing, is formed from the original word by adding -ing; as, move, moving. In the older languages the termination was more marked, being -nd. Like the Latin participle in -ns, it was originally declined. The Mœso-Gothic and Old High German forms are habands and hapêntêr=having, respectively. The -s in the one language, and the -êr in the other, are the signs of the case and gender. In the Old Saxon and Anglo-Saxon the forms are -and and -ande; as bindand, bindande=binding. In all the Norse languages, ancient and modern, the -d is preserved. So it is in the Old Lowland Scotch, and in many of the modern provincial dialects of England, where strikand, goand, is said for striking, going. In Staffordshire, where the -ing is pronounced -ingg, there is a fuller sound than that of the current English. In Old English the form in -nd is predominant, in Middle English, the use fluctuates, and in New English the termination -ing is universal. In the Scotch of the modern writers we find the form -in.

It is with the oblique cases of the present participles of the classical languages, rather than with the nominative, that we must compare the corresponding participle in Gothic; e.g., {349}ἔχοντ-ος (ekhontos), Greek; habent-is, Latin; hapênt-êr, Old High German.

§ 402. It has often been remarked that the participle is used in many languages as a substantive. This is true in Greek,

§ 403. But it is also stated, that, in the English language, the participle is used as a substantive in a greater degree than elsewhere, and that it is used in several cases and in both numbers, e.g.,

Some remarks of Mr. R. Taylor, in the Introduction to his edition of Tooke's Diversions of Purley, modify this view. According to these, the -ing in words like rising is not the -ing of the present participle; neither has it originated in the Anglo-Saxon -end. It is rather the -ing in words like morning, which is anything but a participle of the non-existent verb morn, and which has originated in the Anglo-Saxon substantival termination -ung. Upon this Rask writes as follows:—"Gitsung, gewilnung=desire; swutelung=manifestation; clænsung=a cleansing; sceawung=view, contemplation; eorð beofung=an earthquake; gesomnung=an assembly. This termination is chiefly used in forming substantives from verbs of the first class in -ian; as, hálgung=consecration, from hálgian=to consecrate. These verbs are all feminine."—Anglo-Saxon Grammar, p. 107.

Now, whatever may be the theory of the origin of the termination -ing in old phrases like rising early is healthy, it cannot apply to expressions of recent introduction. Here the direct origin in -ung is out of the question. {350}

1. That the older forms in -ing are substantival in origin, and=the Anglo-Saxon -ung.

2. That the latter ones are participial, and have been formed on a false analogy.

CHAPTER XXX.

§ 404. The participle in -en.—In the Anglo-Saxon this participle was declined like the adjectives. Like the adjectives, it is, in the present English, undeclined.

In Anglo-Saxon it always ended in -en, as sungen, funden, bunden. In English this -en is often wanting, as found, bound; the word bounden being antiquated. Words where the -en is wanting may be viewed in two lights; 1, they may be looked upon as participles that have lost their termination; 2, they may be considered as præterites with a participial sense.

§ 405. Drank, drunk, drunken.—With all words wherein the vowel of the plural differs from that of the singular, the participle takes the plural form. To say I have drunk, is to use an ambiguous expression; since drunk may be either a participle minus its termination, or a præterite with a participial sense. To say I have drank, is to use a præterite for a participle. To say I have drunken, is to use an unexceptionable form.

In all words with a double form, as spake and spoke, brake and broke, clave and clove, the participle follows the form in o, as spoken, broken, cloven. Spaken, braken, claven, are impossible forms. There are degrees in laxity of language, and to say the spear is broke is better than to say the spear is brake.

These two statements bear upon the future history of the præterite. That of the two forms sang and sung, one will, in the course of language, become obsolete is nearly certain; and, as the plural form is also that of the participle, it is the plural form which is most likely to be the surviving one. {352}

§ 406. As a general rule, we find the participle in -en wherever the præterite is strong; indeed, the participle in -en may be called the strong participle, or the participle of the strong conjugation. Still the two forms do not always coincide. In mow, mowed, mown; sow, sowed, sown; and several other words, we find the participle strong, and the præterite weak. I remember no instances of the converse. This is only another way of saying that the præterite has a greater tendency to pass from strong to weak than the participle.

§ 407. In the Latin language the change from s to r, and vice versâ, is very common. We have the double forms arbor and arbos, honor and honos, &c. Of this change we have a few specimens in English. The words rear and raise, as compared with each other, are examples. In Anglo-Saxon a few words undergo a similar change in the plural number of the strong præterites.

This accounts for the participial form forlorn, or lost, in New High German verloren. In Milton's lines,

§ 408. The participle in -d, -t, or -ed.—In the Anglo-Saxon this participle was declined like the adjective. Like the adjective, it is, in the present English, undeclined.

In Anglo-Saxon it differed in form from the præterite, inasmuch as it ended in -ed, or -t, whereas the præterite ended in -ode, -de, or -te: as, lufode, bærnde, dypte, præterites; gelufod, bærned, dypt, participles.

As the ejection of the e reduces words like bærned and bærnde to the same form, it is easy to account for the present {353}identity of form between the weak præterites and the participles in -d: e. g., I moved, I have moved, &c.

§ 409. In the older writers, and in works written, like Thomson's Castle of Indolence, in imitation of them, we find prefixed to the præterite participle the letter y-, as yclept=called: yclad=clothed: ydrad=dreaded.

The following are the chief facts and the current opinion concerning this prefix:—

1. It has grown out of the fuller forms ge-: Anglo-Saxon, ge-: Old Saxon, gi-: Mœso-Gothic, ga-: Old High German, ka-, cha-, ga-, ki-, gi-.

3. It occurs, with a few fragmentary exceptions, in none of the Scandinavian languages of the Gothic stock.

4. In Anglo-Saxon it occasionally indicates a difference of sense; as hâten=called, ge-hâten=promised, boren=borne, ge-boren=born.

6. Its power, in the case of nouns, is generally some idea of association, or collection.—Mœso-Gothic, sinþs=a journey, ga-sinþa=a companion; Old High German, perc=hill; ki-perki (ge-birge)=a range of hills.

7. But it has also a frequentative power; a frequentative power which is, in all probability, secondary to its collective power: since things which recur frequently recur with a tendency to collection or association; Middle High German, ge-rassel=rustling; ge-rumpel=c-rumple.

This power is also a collective, since every quality is associated with the object that possesses it: a sea with waves=a wavy sea. {354}

9. Hence it is probable that the ga-, ki-, or gi-, Gothic, is the cum of Latin languages. Such is Grimm's view, as given in Deutsche Grammatik, i. 1016.

Concerning this, it may be said that it is deficient in an essential point. It does not show how the participle past is collective. Undoubtedly it may be said that every such participle is in the condition of words like ge-feax and ge-heort; i. e., that they imply an association between the object and the action or state. But this does not seem to be Grimm's view; he rather suggests that the ge- may have been a prefix to verbs in general, originally attached to all their forms, but finally abandoned everywhere except in the case of the participle. The theory of this prefix has yet to assume a satisfactory form.

CHAPTER XXXI.

§ 410. In the following words, amongst many others, we have palpable and indubitable specimens of composition. Day-star, vine-yard, sun-beam, apple-tree, ship-load, silver-smith, &c. The words palpable and indubitable have been used, because, in many cases, as will be seen hereafter, it is difficult to determine whether a word be a true compound or not.

Now, in each of the compounds quoted above, it may be seen that it is the second word which is qualified, or defined, by the first, and that it is not the first which is qualified or defined, by the second. Of yards, beams, trees, loads, smiths, there may be many sorts, and, in order to determine what particular sort of yard, beam, tree, load, or smith, may be meant, the words vine, sun, apple, ship, and silver, are prefixed. In compound words it is the first term that defines or particularises the second.

§ 411. That the idea given by the word apple-tree is not referable to the words apple and tree, irrespective of the order in which they occur, may be seen by reversing the position of them. The word tree-apple, although not existing in the language, is as correct a word as thorn-apple. In tree-apple, the particular sort of apple meant is denoted by the word tree, and if there were in our gardens various sorts of plants called apples, of which some grew along the ground and others upon trees, such a word as tree-apple would be required in order to be opposed to earth-apple, or ground-apple, or some word of the kind.

In the compound words tree-apple and apple-tree, we have the same elements differently arranged. However, as the {356}word tree-apple is not current in the language, the class of compounds indicated by it may seem to be merely imaginary. Nothing is farther from being the case. A tree-rose is a rose of a particular sort. The generality of roses being on shrubs, this grows on a tree. Its peculiarity consists in this fact, and this particular character is expressed by the word tree prefixed. A rose-tree is a tree of a particular sort, distinguished from apple-trees, and trees in general (in other words, particularised or defined) by the word rose prefixed.

A ground-nut is a nut particularised by growing in the ground. A nut-ground is a ground particularised by producing nuts.

A finger-ring, as distinguished from ear-rings, and from rings in general (and so particularised), is a ring for the finger. A ring finger, as distinguished from fore-fingers, and from fingers in general (and so particularised), is a finger whereon rings are worn.

§ 412. At times this rule seems to be violated. The words spitfire and daredevil seem exceptions to it. At the first glance it seems, in the case of a spitfire, that what he (or she) spits is fire; and that, in the case of a daredevil, what he (or she) dares is the devil. In this case the initial words spit and dare, are particularised by the final ones fire and devil. The true idea, however, confirms the original rule. A spitfire voids his fire by spitting. A daredevil, in meeting the fiend, would not shrink from him, but would defy him. A spitfire is not one who spits fire, but one whose fire is spit. A daredevil is not one who dares even the devil, but one by whom the devil is even dared.

§ 413. Of the two elements of a compound word, which is the most important? In one sense the latter, in another sense the former. The latter word is the most essential; since the general idea of trees must exist before it can be defined or particularised; so becoming the idea which we have in apple-tree, rose-tree, &c. The former word, however, is the most influential. It is by this that the original idea is qualified. The latter word is the staple original element: the former is the superadded influencing element. Compared with each {357}other, the former element is active, the latter passive. Etymologically speaking, the former element, in English compounds, is the most important.

§ 414. Most numerous are the observations that bear upon the composition of words; e.g., how nouns combine with nouns, as in sunbeam; nouns with verbs, as in daredevil, &c. It is thought sufficient in the present work to be content with, 1. defining the meaning of the term composition; 2. explaining the nature of some obscure compounds.

Composition is the joining together, in language, of two different words, and treating the combination as a single term. Observe the words in italics.

In language.—A great number of our compounds, like the word merry-making, are divided by the sign -, or the hyphen. It is very plain that if all words spelt with a hyphen were to be considered as compounds, the formation of them would be not a matter of speech, or language, but one of writing or spelling. This distinguishes compounds in language from mere printers' compounds.

Different.—In Old High German we find the form sëlp-sëlpo. Here there is the junction of two words, but not the junction of two different ones. This distinguishes composition from gemination.—Grimm, Deutsche Grammatik, iii. 405.

Words.—In father-s, clear-er, four-th, &c., there is the addition of a letter or a syllable, and it may be even of the part of a word. There is no addition, however, of a whole word. This distinguishes composition from derivation.

Treating the combination as a single term.—In determining, in certain cases, between derived words and compound words, there is an occasional perplexity; the perplexity, however, is far greater in determining between a compound word and two words. In the eyes of one grammarian the term mountain height may be as truly a compound word as sunbeam. In the eyes of another grammarian it may be no compound word, but two words, just as Alpine height is two words; mountain being dealt with as an adjective. It is in the determination of this that the accent plays an important part. This fact was foreshadowed in the Chapter upon Accents. {358}

§ 415. The attention of the reader is drawn to the following line, slightly altered, from Churchill:—

On each of the syllables rest, friend, spare, prec-, breath, there is an accent. Each of these syllables must be compared with the one that precedes it; rest with then, friend with my, and so on throughout the line. Compared with the word and, the word spare is not only accented, but the accent is conspicuous and prominent. There is so little on and, and so much on spare, that the disparity of accent is very manifest.

Now, if in the place of and, there was some other word, a word not so much accented as spare, but still more accented than and, this disparity would be diminished, and the accents of the two words might be said to be at par, or nearly so. As said before, the line was slightly altered from Churchill, the real reading being

In the true reading we actually find what had previously only been supposed. In the words spare, spare, the accents are nearly at par. Such the difference between accent at par and disparity of accent.

Good illustrations of the parity and disparity of accent may be drawn from certain names of places. Let there be such a sentence as the following: the lime house near the bridge north of the new port. Compare the parity of accent on the separate words lime and house, bridge and north, new and port, with the disparity of accent in the compound words Límehouse, Brídgenorth, and Néwport. The separate words beef steak, where the accent is nearly at par, compared with the compound word sweépstakes, where there is a great disparity of accent, are further illustrations of the same difference.

§ 416. The difference between a compound word and two words is greatest where the first is an adjective. This we see in comparing such terms as the following: bláck bírd, meaning a bird that is black, with bláckbird=the Latin merula; or blúe béll, meaning a bell that is blue, with blúebell, the flower. {359}Expressions like a shárp edgéd instrument, meaning an instrument that is sharp and has edges, as opposed to a shárp-edged instrument, meaning an instrument with sharp edges, further exemplify this difference.

Subject to four small classes of exceptions, it may be laid down, that, in the English language, there is no composition unless there is either a change of form or a change of accent.

The reader is now informed, that unless, in what has gone before, he has taken an exception to either a statement or an inference, he has either seen beyond what has been already laid down by the author, or else has read him with insufficient attention. This may be shown by drawing a distinction between a compound form and a compound idea.

In the words a red house, each word preserves its natural and original meaning, and the statement is that a house is red. By a parity of reasoning a mad house should mean a house that is mad; and, provided that each word retain its natural meaning and its natural accent, such is the fact. Let a house mean, as it often does, a family. Then the phrase, a mad house, means that the house, or family, is mad, just as a red house means that the house is red. Such, however, is not the current meaning of the word. Every one knows that a mad house means a house for mad men; in which case it is treated as a compound word, and has a marked accent on the first syllable, just as Límehouse has. Now, compared with the word red house, meaning a house of a red colour, and compared with the words mad house, meaning a deranged family, the word mádhouse, in its common sense, expresses a compound idea; as opposed to two ideas, or a double idea. The word beef steak is evidently a compound idea; but, as there is no disparity of accent, it is not a compound word. Its sense is compound; its form is not compound, but double. This indicates the objection anticipated, which is this: viz., that a definition, which would exclude such a word as beef steak from the list of compounds, is, for that very reason, exceptionable. I answer to this, that the term in question is a compound idea, and not a compound form; in other words, that it is a compound in logic, but not a compound in etymology. {360}Now etymology, taking cognisance of forms only, has nothing to do with ideas, except so far as they influence forms.

Such is the commentary upon the words, "treating the combination as a single term;" in other words, such the difference between a compound word and two words. The rule, being repeated, stands (subject to the four classes of exceptions) thus: There is no true composition without either a change of form or a change of accent. As I wish to be clear upon this point, I shall illustrate the statement by its application.

The word trée-rose is often pronounced trée róse; that is, with the accent at par. It is compound in the one case; it is two words in the other.

The words mountain ash and mountain height are generally (perhaps always) pronounced with an equal accent on the syllables mount- and ash, mount- and height, respectively. In this case the word mountain must be dealt with as an adjective, and the words considered as two. The word moúntain wave is often pronounced with a visible diminution of accent on the last syllable. In this case there is a disparity of accent, and the word is compound.

§ 417. The following quotation indicates a further cause of perplexity in determining between compound words and two words:—

To speak first of the word (or words) gallant mast. If gallant mean brave, there are two words. If the words be two, there {361}is a stronger accent on mast. If the accent on mast be stronger, the rhyme with fast is more complete; in other words, the metre favours the notion of the words being considered as two. Gallant-mast, however, is a compound word, with an especial nautical meaning. In this case the accent is stronger on gal- and weaker on -mast. This, however, is not the state of things that the metre favours. The same applies to mountain wave. The same person who in prose would throw a stronger accent on mount- and a weaker one on wave (so dealing with the word as a compound), might, in poetry, make the words two, by giving to the last syllable a parity of accent.

The following quotation from Ben Jonson may be read in two ways; and the accent may vary with the reading.

§ 418. On certain words wherein the fact of their being compound is obscured.—Composition is the addition of a word to a word, derivation is the addition of letters or syllables to a word. In a compound form each element has a separate and independent existence; in a derived form, only one of the elements has such. Now it is very possible that in an older stage of a language two words may exist, may be put together, and may so form a compound; at the time in point each word having a separate and independent existence: whilst, in a later stage of language, only one of these words may have a separate and independent existence, the other having become obsolete. In this case a compound word would take the appearance of a derived one, since but one of its elements could be exhibited as a separate and independent word. Such is the case with, amongst others, the word bishopric. In the present language the word ric has no separate and independent existence. For all this, the word {362}is a true compound, since, in Anglo-Saxon, we have the noun ríce as a separate, independent word, signifying kingdom or domain.

Again, without becoming obsolete, a word may alter its form. This is the case with most of our adjectives in -ly. At present they appear derivative; their termination -ly having no separate and independent existence. The older language, however, shows that they are compounds; since -ly is nothing else than -lic, Anglo-Saxon; -lih, Old High German; -leiks, Mœso-Gothic;=like, or similis, and equally with it an independent separate word.

For the following words a separate independent root is presumed rather than shown. It is presumed, however, on grounds that satisfy the etymologist.

Mis-, as in misdeed, &c.—Mœso-Gothic, missô=in turns; Old Norse, â mis=alternately; Middle High German, misse=mistake. The original notion alternation, thence change, thence defect. Compare the Greek ἄλλως.—Grimm, Deutsche Grammatik, ii. 470.

Dom, as in wisdom, &c.—The substantive dôm presumed.—Deutsche Grammatik, ii. 491.

Hood and head, as in Godhead, manhood, &c.—The substantive háids=person, order, kind, presumed.—Deutsche Grammatik, ii. 497. Nothing to do with the word head.

Ship, as in friendship.—Anglo-Saxon, -scipe and -sceäft; German, -schaft; Mœso-Gothic, gaskafts=a creature, or creation. The substantive skafts or skap presumed. The -skip or -scape in landskip is only an older form.—Deutsche Grammatik, ii. 522.

Less, as in sleepless, &c., has nothing to do with less. Derived from láus, lôs, destitute of=Latin, expers.—Deutsche Grammatik, ii. 565.

For the further details, which are very numerous, see the Deutsche Grammatik, vol. iii.

§ 419. "Subject to four classes of exceptions, it may be laid down that there is no true composition unless there is either a change of form or a change of accent."—Such is the statement made in p. 359. The first class of exceptions consists {363}of those words where the natural tendency to disparity of accent is traversed by some rule of euphony. For example, let two words be put together, which at their point of contact form a combination of sounds foreign to our habits of pronunciation. The rarity of the combination will cause an effort in utterance. The effort in utterance will cause an accent to be laid on the latter half of the compound. This will equalize the accent, and abolish the disparity. The word monkshood, the name of a flower (aconitum napellus), where, to my ear at least, there is quite as much accent on the -hood as on the monks-, may serve in the way of illustration. Monks is one word, hood another. When joined together, the h- of the -hood is put in immediate opposition with the -s of the monks-. Hence the combination monkshood. At the letters s and h is the point of contact. Now the sound of s followed immediately by the sound of h is a true aspirate. But true aspirates are rare in the English language. Being of rare occurrence, the pronunciation of them is a matter of attention and effort; and this attention and effort creates an accent which otherwise would be absent. Hence words like monkshóod, well-héad, and some others.

Real reduplications of consonants, as in hop-pole, may have the same parity of accent with the true aspirates: and for the same reasons. They are rare combinations that require effort and attention.

The second class of exceptions contains those words wherein between the first element and the second there is so great a disparity, either in the length of the vowel, or the length of the syllable en masse, as to counteract the natural tendency of the first element to become accented. One of the few specimens of this class (which after all may consist of double words) is the term upstánding. Here it should be remembered, that words like hapházard, foolhárdy, uphólder, and withhóld come under the first class of the exceptions.

The third class of exceptions contains words like perchánce and perháps. In all respects but one these are double words, just as by chance is a double word. Per, however, differs from by in having no separate existence. This sort of words {364}we owe to the multiplicity of elements (classical and Gothic) in the English language.

To anticipate objections to the rule respecting the disparity of accent, it may be well to state in fresh terms a fact already indicated, viz., that the same combination of words may in one sense be compound, and in the other double (or two). An uphill game gives us the combination up + hill as a compound. He ran up hill gives us the combination up + hill as two words. So it is with down + hill, down + right, and other words. Man-servant, cock-sparrow, &c., are double or compound, as they are pronounced mán-sérvant, mán-servant, cóck-spárrow, or cóck-sparrow.

The fourth class is hypothetical. I can, however, imagine that certain compounds may, if used almost exclusively in poetry, and with the accent at par, become so accented even in the current language.

§ 420. For a remark on the words peacock, peahen, see the Chapter upon Gender.—If these words be rendered masculine or feminine by the addition of the elements -cock and -hen, the statements made in the beginning of the present chapter are invalidated. Since, if the word pea- be particularized, qualified, or defined by the words -cock and -hen, the second term defines or particularises the first, which is contrary to the rule of p. 355. The truth, however, is, that the words -cock and -hen are defined by the prefix pea-. Preparatory to the exhibition of this, let us remember that the word pea (although now found in composition only) is a true and independent substantive, the name of a species of fowl, like pheasant, partridge, or any other appellation. It is the Latin pavo, German pfau. Now, if the word peacock mean a pea (pfau or pavo) that is a male, then do wood-cock, black-cock, and bantam-cock, mean woods, blacks, and bantams that are male. Or if the word peahen mean a pea (pfau or pavo) that is female, then do moorhen and guineahen mean moors and guineas that are female. Again, if a peahen mean a pea (pfau or pavo) that is female, then does the compound pheasant-hen mean the same as hen-pheasant; which is not the case. The fact is that peacock means a cock that is a pea (pfau or pavo); {365}peahen means a hen that is a pea (pfau or pavo); and, finally, peafowl means a fowl that is a pea (pfau or pavo). In the same way moorfowl means, not a moor that is connected with a fowl, but a fowl that is connected with a moor.

§ 421. It must be clear, ex vi termini, that in every compound word there are two parts; i. e., the whole or part of the original, and the whole or part of the superadded word. In the most perfect forms of inflection there is a third element, viz., a vowel, consonant, or syllable that joins the first word with the second.

In the older forms of all the Gothic languages the presence of this third element was the rule rather than the exception. In the present English it exists in but few words.

b. The -in- in night-in-gale is most probably such a connecting element. Compare the German form nacht-i-gale, and remember the tendency of vowels to take the sound of -ng before g.

§ 422. Improper compounds.—The -s- in words like Thur-s-day, hunt-s-man, may be one of two things.

a. It may be the sign of the genitive case, so that Thursday=Thoris dies. In this case the word is an improper compound, since it is like the word pater-familias in Latin, in a common state of syntactical construction.

b. It may be a connecting sound, like the -i- in nacht-i-gale. Reasons for this view occur in the following fact:—

In the Modern German languages the genitive case of feminine nouns ends otherwise than in -s. Nevertheless, the sound of -s- occurs in composition equally, whether the noun it follows be masculine or feminine. This fact, as far as it goes, makes it convenient to consider the sound in question as a connective rather than a case. Probably, it is neither one nor the other exactly, but the effect of a false analogy.

§ 423. Decomposites.—"Composition is the joining together of two words."—See p. 357.

In the first edition the sentence ran "two or more" words; being so written to account for compounds like mid-ship-man, {366}gentle-man-like, &c., where the number of verbal elements seems to amount to three.

Nevertheless, the caution was unnecessary. Compound radicals like midship and gentleman, are, for the purposes of composition, single words. Compounds wherein one element is compound are called decomposites.

§ 424. The present chapter closes with the notice of two classes of words. They are mentioned now, not because they are compounds, but because they can be treated of here more conveniently than elsewhere.

There are a number of words which are never found by themselves; or, if so found, have never the same sense that they have in combination. Mark the word combination. The terms in question are points of combination, not of composition: since they form not the parts of words, but the parts of phrases. Such are the expressions time and tide—might and main—rede me my riddle—pay your shot—rhyme and reason, &c. These words are evidently of the same class, though not of the same species with bishopric, colewort, spillikin, gossip, mainswearer, and the words quoted in p. 362. These last-mentioned terms give us obsolete words preserved in composition. The former give us obsolete words preserved in combination.

The other words are etymological curiosities. They may occur in any language. The English, however, from the extent of its classical element, is particularly abundant in them. It is a mere accident that they are all compound words.

CHAPTER XXXII.

§ 425. Derivation, like etymology, is a word used in a wide and in a limited sense. In the wide sense of the term every word, except it be in the simple form of a root, is a derived word. In this sense the cases, numbers, and genders of nouns, the persons, moods, and tenses of verbs, the ordinal numbers, the diminutives, and even the compound words, are alike matters of derivation. In the wide sense of the term the word fathers, from father, is equally in a state of derivation with the word strength, from strong.

In the use of the word, even in its limited sense, there is considerable laxity and uncertainty.

Gender, number, case.—These have been called the accidents of the noun, and these it has been agreed to separate from derivation in its stricter sense, or from derivation properly so called, and to class together under the name of declension. Nouns are declined.

Person, number, tense, voice.—These have been called the accidents of a verb, and these it has been agreed to separate from derivation properly so called, and to class together under the name of conjugation. Verbs are conjugated.

Conjugation and declension constitute inflection. Nouns and verbs, speaking generally, are inflected.

Inflection, a part of derivation in its wider sense, is separated from derivation properly so called, or from derivation in its limited sense.

The degrees of comparison, or certain derived forms of adjectives; the ordinals, or certain derived forms of the numerals; the diminutives, &c., or certain derived forms of the substantive, have been separated from derivation properly {368}so called. I am not certain, however, that for so doing there is any better reason than mere convenience. By some the decrees of comparison are considered as points of inflection.

Derivation proper, the subject of the present chapter, comprises all the changes that words undergo, which are not referable to some of the preceding heads. As such, it is, in its details, a wider field than even composition. The details, however, are not entered into.

§ 426. Derivation proper may be divided according to a variety of principles. Amongst others,

I. According to the evidence.—In the evidence that a word is not simple, but derived, there are at least two degrees.

A. That the word strength is a derived word I collect to a certainty from the word strong, an independent form, which I can separate from it. Of the nature of the word strength there is the clearest evidence, or evidence of the first degree.

B. Fowl, hail, nail, sail, tail, soul; in Anglo-Saxon, fugel, hægel, nægel, segel, tægel, sawel. —These words are by the best grammarians considered as derivatives. Now, with these words I can not do what was done with the word strength, I can not take from them the part which I look upon as the derivational addition, and after that leave an independent word. Strength - th is a true word; fowl or fugel - l is no true word. If I believe these latter words to be derivations at all, I do it because I find in words like handle, &c., the -l as a derivational addition. Yet, as the fact of a word being sometimes used as a derivational addition does not preclude it from being at other times a part of the root, the evidence that the words in question are not simple, but derived, is not cogent. In other words, it is evidence of the second degree.

II. According to the effect.—The syllable -en in the word whiten changes the noun white into a verb. This is its effect. We may so classify as to arrange combinations like -en (whose effect is to give the idea of the verb) in one order; whilst combinations like th (whose effect is, as in the word strength, to give the idea of abstraction) form another order.

III. According to the form.—Sometimes the derivational {369}element is a vowel (as the -ie in doggie); sometimes a consonant combined: in other words, a syllable (as the -en in whiten); sometimes a change of vowel without any addition (as the i in tip, compared with top); sometimes a change of consonant without any addition (as the z in prize, compared with price; sometimes it is a change of accent, like a súrvey, compared with to survéy. To classify derivations in this manner is to classify them according to their form. For the detail of the derivative forms, see Deutsche Grammatik, ii. 89-405.

IV. According to the historical origin of the derivational elements.—For this see the Chapter upon Hybridism.

V. According to the number of the derivational elements.—In fisher, as compared with fish, there is but one derivational affix. In fishery, as compared with fish, the number of derivational elements is two.

§ 427. The list (taken from Walker) of words alluded to in p. 293, is as follows:—

§ 428. Churl, earl, owl, fowl, hail, nail, sail, snail, tail, hazel, needle, soul, teazle, fair, beam, bottom, arm, team, worm, heaven, morn, dust, ghost, breast, rest, night, spright, blind, harp, flax, fox, finch, stork, &c. All these words, for certain etymological reasons, are currently considered, by the latest philologists, as derivatives. Notwithstanding the general prevalence of a fuller form in the Anglo-Saxon, it is clear that, in respect to the evidence, they come under division B.

§ 429. Forms like tip, from top, price and prize, &c., are of importance in general etymology. Let it be received as a theory (as with some philologists is really the case) that fragmentary sounds like the -en in whiten, the -th in strength, &c., were once words; or, changing the expression, let it be considered that all derivation was once composition. Let this view be opposed. The first words that are brought to militate against it are those like tip and prize, where, instead of any addition, there is only a change; and, consequently, no vestiges of an older word. This argument, good as far as it goes, is rebutted in the following manner. Let the word top have attached to it a second word, in which second word there is a small vowel. Let this small vowel act upon the full one in top, changing it to tip. After this, let the second word be ejected. We then get the form tip by the law of accommodation, and not as an immediate sign of derivation. The i in chick (from cock) may be thus accounted for, the -en in chicken being supposed to have exerted, first, an influence of accommodation, and afterwards to have fallen off. The i in chick may, however, be accounted for by simple processes.

§ 430. In words like bishopric, and many others mentioned in the last chapter, we had compound words under the appearance of derived ones; in words like upmost, and many others, we have derivation under the appearance of composition.

CHAPTER XXXIII.

§ 431. Adverbs.—The adverbs are capable of being classified after a variety of principles.

Firstly, they may be divided according to their meaning. In this case we speak of the adverbs of time, place, number, manner. This division is logical rather than etymological.

A division, however, which although logical bears upon etymology, is the following:—

Well, better, ill, worse.—Here we have a class of adverbs expressive of degree, or intensity. Adverbs of this kind are capable of taking an inflection, viz., that of the comparative and superlative degrees.

Now, then, here, there.—In the idea expressed by these words there are no degrees of intensity. Adverbs of this kind are incapable of taking any inflection.

Words like better and worse are adjectives or adverbs as they are joined to nouns or verbs.

Adverbs differ from nouns and verbs in being susceptible of one sort of inflection only, viz., that of degree.

Secondly, adverbs may be divided according to their form and origin. This is truly an etymological classification.

A Better, worse.—Here the combination of sounds gives equally an adjective and an adverb. This book is better than that—here better agrees with book, and is therefore adjectival. This looks better than that—here better qualifies looks, and is therefore adverbial. Again; to do a thing with violence is equivalent to do a thing violently. This shows how adverbs may arise out of cases. In words like the English better, the Latin vi=violenter, the Greek κάλον=κάλως, we have {372}adjectives in their degrees, and substantives in their cases, with adverbial powers. In other words, nouns are deflected from their natural sense to an adverbial one. Adverbs of this kind are adverbs of deflection.

B Brightly, bravely.—Here an adjective is rendered adverbial by the addition of the derivative syllable -ly. Adverbs like brightly, &c., may (laxly speaking) be called adverbs of derivation.

C Now.—This word has not satisfactorily been shown to have originated as any other part of speech but as an adverb. Words of this sort are adverbs absolute.

When, now, well, worse, better.—here the adverbial expression consists in a single word, and is simple. To-day, yesterday, not at all, somewhat—here the adverbial expression consists of a compound word, or a phrase. This indicates the division of adverbs into simple and complex.

§ 432. The adverbs of deflection (of the chief importance in etymology) may be arranged after a variety of principles. I. According to the part of speech from whence they originate. This is often an adjective, often a substantive, at times a pronoun, occasionally a preposition, rarely a verb. II. According to the part of the inflection from whence they originate. This is often an ablative case, often a neuter accusative, often a dative, occasionally a genitive.

Else, unawares, eftsoons.—These are the genitive forms of adjectives. By rights is a word of the same sort.

Little, less, well.—Neuter accusatives of adjectives. Bright, in the sun shines bright, is a word of the same class. The {373}neuter accusative is a common source of adverbs in all tongues.

§ 433. Darkling.—This is no participle of a verb darkle, but an adverb of derivation, like unwaringun=unawares, Old High German; stillinge=secretly, Middle High German; blindlings=blindly, New High German; darnungo=secretly, Old Saxon; nichtinge=by night, Middle Dutch; blindeling=blindly, New Dutch; bæclinga=backwards, handlunga=hand to hand, Anglo-Saxon; and, finally, blindlins, backlins, darklins, middlins, scantlins, stridelins, stowlins, in Lowland Scotch.—Deutsche Grammatik, iii. 236.

§ 434. "Adverbs like brightly may (laxly speaking) be called adverbs of derivation." Such the assertion made a few paragraphs above. The first circumstance that strikes the reader is, that the termination -ly is common both to adjectives and to adverbs. This termination was once an independent word, viz., leik. Now, as -ly sprung out of the Anglo-Saxon -lice, and as words like early, dearly, &c., were originally arlîce, deorlîce, &c., and as arlîce, deorlîce, &c., were adjectives, the adverbs in -ly are (strictly speaking) adverbs, not of derivation, but of deflection.

It is highly probable that not only the adverbs of derivation, but that also the absolute adverbs, may eventually be reduced to adverbs of deflection. For now, see Deutsche Grammatik, iii. 249.

CHAPTER XXXIV.

§ 435. It is a common practice for languages to express by different modifications of the same root the three following ideas:—

This habit gives us three correlative adverbs—one of position, and two of direction.

§ 436. It is also a common practice of language to depart from the original expression of each particular idea, and to interchange the signs by which they are expressed.

§ 437. This may be seen in the following table, illustrative of the forms here, hither, hence, and taken from the Deutsche Grammatik, iii. 199.

§ 438. These local terminations were commoner in the earlier stages of language than at present. The following are from the Mœso-Gothic:—

Now a reason for the comparative frequency of these forms in Mœso-Gothic lies in the fact of the Gospel of Ulphilas being a translation from the Greek. The Greek forms in -θεν, ἔσωθεν, ἔξωθεν, ἄνωθεν, πόῤῥωθεν, πάντοθεν, were just the forms to encourage such a formation as that in -þro.—Deutsche Grammatik, iii. 199, &c.

§ 439. The -ce (=es) in hen-ce, when-ce, then-ce, has yet to be satisfactorily explained. The Old English is whenn-es, thenn-es. As far, therefore, as the spelling is concerned, they are in the same predicament with the word once, which is properly on-es, the genitive of one. This statement, however, explains only the peculiarity of their orthography; since it by no means follows, that, because the -s in ones and the -s in whennes, thennes are equally replaced by -ce in orthography, they must equally have the same origin in etymology.

§ 440. Yonder.—In the Mœso-Gothic we have the following forms: jáinar, jáina, jáinþrô=illic, illuc, illinc. They do not, however, explain the form yon-d-er. It is not clear whether the d=the -d in jâind, or the þ in jáinþro.

Anon, as used by Shakspeare, in the sense of presently.—The probable history of this word is as follows: the first syllable contains a root akin to the root yon, signifying distance in place. The second is a shortened form of the Old High German and Middle High German, -nt, a termination expressive, 1, of removal in space; 2, of removal in time; Old High German, ënont, ënnont; Middle High German, {376}ënentlig, jenunt=beyond. The transition from the idea of place to that of time is shown in the Old High German, nâhunt, and the Middle High German, vërnent=lately; the first from the root nigh, the latter from the root far.—See Deutsche Grammatik, iii. 215.

CHAPTER XXXV.

The masculine accusative cases of the relative and demonstrative pronoun are hwæne (hwone) and þæne (þone).

Notwithstanding the difference, the first form is a variety of the second; so that the adverbs when and then are pronominal in origin.

As to the word than, the conjunction of comparison, it is a variety of then; the notions of order, sequence, and comparison being allied.

This is good: then (or next in order) that is good, is an expression sufficiently similar to this is better than that to have given rise to it.

CHAPTER XXXVI.

§ 442. Prepositions.—Prepositions, as such, are wholly unsusceptible of inflection. Other parts of speech, in a state of inflection, may be used with a prepositional sense. This, however, is not an inflection of prepositions.

No word is ever made a preposition by the addition of a derivational^[59] element. If it were not for this, the practical classification of the prepositions, in respect to their form, would coincide with that of the adverbs. As it is, there are only the prepositions of deflection, and the absolute prepositions. On another principle of division there are the simple prepositions (in, on, &c.), and the complex prepositions (upon, roundabout, across).

The prepositions of deflection, when simple, originate chiefly in adverbs, as up, down, within, without, unless, indeed, we change the assertion, and say that the words in point (and the others like them) are adverbs originating in prepositions. The absence of characteristic terminations renders these decisions difficult.

The prepositions of deflection, when complex, originate chiefly in nouns, accompanied by an absolute preposition; as instead of of substantival, between of adjectival origin.

The absolute prepositions, in the English language, are in, on, of, at, up, by, to, for, from, till, with, through.

§ 443. Conjunctions.—Conjunctions, like prepositions, are wholly unsusceptible of inflection. Like prepositions they {379}are never made by means of a derivational element. Like prepositions they are either simple (as and, if), or complex (as also, nevertheless).

The conjunctions of deflection originate chiefly in imperative moods (as all save one, all except one); participles used like the ablative absolute in Latin (as all saving one, all excepting one); adverbs (as so); prepositions (as for); and relative neuters (as that).

§ 444. Yes, no.—Although not may be reduced to an adverb, nor to a conjunction, and none to a noun, these two words (the direct affirmative, and the direct negative) are referable to none of the current parts of speech. Accurate grammar places them in a class by themselves.

§ 445. Particles.—The word particle is a collective term for all those parts of speech that are naturally unsusceptible of inflection; comprising, 1, interjections; 2, direct affirmatives; 3, direct negatives; 4, absolute conjunctions; 5, absolute prepositions; 6, adverbs unsusceptible of degrees of comparison; 7, inseparable prefixes.

CHAPTER XXXVII.

§ 446. The inflection of pronouns has its natural peculiarities in language; it has also its natural difficulties in philology. These occur not in one language in particular, but in all generally. The most common peculiarity in the grammar of pronouns is the fact of what may be called their convertibility. Of this convertibility the following statements serve as illustration:—

1. Of case.—In our own language the words my and thy, although at present possessives, were previously datives, and, earlier still, accusatives. Again, the accusative you replaces the nominative ye, and vice versâ.

2. Of number.—The words thou and thee are, except in the mouths of Quakers, obsolete. The plural forms, ye and you, have replaced them.

3. Of person.—Laying aside the habit of the Germans and other nations, of using the third person plural for the second singular (as in expressions like wie befinden sie sich = how do they find themselves? instead of how do you find yourself?) the Greek language gives us examples of interchange in the way of persons in the promiscuous use of νιν, μιν, σφε, and ἑαυτοῦ; whilst sich and sik are used with a similar latitude in the Middle High German and Scandinavian.

These statements are made for the sake of illustrating, not of exhausting, the subject. It follows, however, as an inference from them, that the classification of pronouns is complicated. Even if we knew the original power and derivation of every form of every pronoun in a language, it would be far from an easy matter to determine therefrom the paradigm that they should take in grammar. To place a word according to its power in a late stage of language might confuse the study of an early stage. To say that because a word was once in a given class, it should always be so, would be to deny that in the present English they, these, and she are personal pronouns at all.

The two tests, then, of the grammatical place of a pronoun, its present power and its original power, are often conflicting.

In the English language the point of most importance in this department of grammar is the place of forms like mine and thine; in other words, of the forms in -n. Are they genitive cases of a personal pronoun, as mei and tui are supposed to be in Latin, or are they possessive pronouns like meus and tuus?

Now, if we take up the common grammars of the English language as it is, we find, that, whilst my and thy are dealt with as genitive cases, mine and thine are considered adjectives. In the Anglo-Saxon grammars, however, min and þin, the older forms of mine and thine, are treated as genitives; of which my and thy have been dealt with as abbreviated forms, and that by respectable scholars.

Now, to prove from the syntax of the older English that in many cases the two forms were convertible, and to answer that the words in question are either genitive cases or adjectives, is lax philology; since the real question is, which of the two is the primary, and which the secondary meaning?

§ 447. The à priori view of the likelihood of words like mine and thine being genitive cases, must be determined by the comparison of three series of facts.

1. The ideas expressed by the genitive case, with particular reference to the two preponderating notions of possession and partition. {382}

2. The circumstance of the particular notion of possession being, in the case of the personal pronouns of the two first persons singular, generally expressed by a form undoubtedly adjectival.

3. The extent to which the idea of partition becomes merged in that of possession, and vice versâ.

§ 448. The ideas of possession and partition as expressed by genitive forms.—If we take a hundred genitive cases, and observe their construction, we shall find, that, with a vast majority of them, the meaning is reducible to one of two heads; viz., the idea of possession or the idea of partition.

Compared with these two powers all the others are inconsiderable, both in number and importance; and if, as in the Greek and Latin languages, they take up a large space in the grammars, it is from their exceptional character rather than from their normal genitival signification.

Again, if both the ideas of possession and partition may, and in many cases must be, reduced to the more general idea of relation, this is a point of grammatical phraseology by no means affecting the practical and special bearings of the present division.

§ 449. The adjectival expression of the idea of possession.—All the world over, a property is a possession; and persons, at least, may be said to be the owners of their attributes. Whatever may be the nature of words like mine and thine, the adjectival character of their Latin equivalents, meus and tuus, is undoubted.

The ideas of partition and possession merge into one another.—A man's spade is the possession of a man; a man's hand is the part of a man. Nevertheless, when a man uses his hand as the instrument of his will, the idea which arises from the fact of its being part of his body is merged in the idea of the possessorship which arises from the feeling of ownership or mastery which is evinced in its subservience and application. Without following the refinements to which the further investigation of these questions would lead us, it is sufficient to suggest that the preponderance of the two allied ideas of partition and possession is often determined by the {383}personality or the non-personality of the subject, and that, when the subject is a person, the idea is chiefly possessive; when a thing, partitive—caput fluvii=the head, which is a part, of a river; caput Toli=the head, which is the possession, of Tolus.

But as persons may be degraded to the rank of things, and as things may, by personification, be elevated to the level of persons, this distinction, although real, may become apparently invalid. In phrases like a tributary to the Tiber—the criminal lost his eye—this field belongs to that parish—the ideas of possessorship and partition, as allied ideas subordinate to the idea of relationship in general, verify the interchange.

§ 450. These observations should bring us to the fact that there are two ideas which, more than any other, determine the evolution of a genitive case—the idea of partition and the idea of possession; and that genitive cases are likely to be evolved just in proportion as there is a necessity for the expression of these two ideas.—Let this be applied to the question of the à priori probability of the evolution of a genitive case to the pronouns of the first and second persons of the singular number.

§ 451. The idea of possession, and its likelihood of determining the evolution of a genitive form to the pronouns of the first and second person singular. —It is less likely to do so with such pronouns than with other words, inasmuch as it is less necessary. It has been before observed, that the practice of most languages shows a tendency to express the relation by adjectival forms—meus, tuus.

An objection against the conclusiveness of this argument will be mentioned in the sequel.

§ 452. The idea of partition, and its likelihood of determining the evolution of a genitive form, &c.—Less than with other words.

A personal pronoun of the singular number is the name of a unity, and, as such, the name of an object far less likely to be separated into parts than the name of a collection. Phrases like, some of them, one of you, many of us, any of them, few of us, &c., have no analogues in the singular number, such as one of me, a few of thee, &c. The partitive words that can {384}combine with singular pronouns are comparatively few; viz., half, quarter, part, &c.: and they can all combine equally with plurals—half of us, a quarter of them, a part of you, a portion of us. The partition of a singular object with a pronominal name is of rare occurrence in language.

This last statement proves something more than appears at first sight. It proves that no argument in favour of the so-called singular genitives, like mine and thine, can be drawn from the admission (if made) of the existence of the true plural genitives ou-r, you-r, thei-r. The two ideas are not in the same predicament. We can say, one of ten, or ten of twenty; but we cannot say one of one—Wæs hira Matheus sum=Matthew was one of them; Andreas—Your noither=neither of you; Amis and Ameloun—from Mr. Guest: Her eyder=either of them; Octavian.—Besides this, the form of the two numbers are neither identical, nor equally genitival; as may be seen by contrasting mi-n and thi-n with ou-r and you-r.

§ 453. Such are the chief à priori arguments against the genitival character of words like mine and thine.

Akin to these, and a point which precedes the à posteriori evidence as to the nature of the words in question, is the determination of the side on which lies the onus probandi. This question is material; inasmuch as, although the present writer believes, for his own part, that the forms under discussion are adjectival rather than genitival, this is not the point upon which he insists. What he insists upon is the fact of the genitival character of mine and thine requiring a particular proof; which particular proof no one has yet given: in other words, his position is that they are not to be thought genitive until proved to be such.

It has not been sufficiently considered that the primâ facie evidence is against them. They have not the form of a genitive case—indeed, they have a different one; and whoever assumes a second form for a given case has the burden of proof on his side.

§ 454. Against this circumstance of the -n in mine and thine being the sign of anything rather than of a genitive case, and against the primâ facie evidence afforded by it, the {385}following facts may, or have been, adduced as reasons on the other side. The appreciation of their value, either taken singly or in the way of cumulative evidence, is submitted to the reader. It will be seen that none of them are unexceptionable.

§ 455. The fact, that, if the words mine and thine are not genitive cases, there is not a genitive case at all.—It is not necessary that there should be one. Particular reasons in favour of the probability of personal pronouns of the singular number being destitute of such a case have been already adduced. It is more likely that a word should be defective than that it should have a separate form.

§ 456. The analogy of the forms mei and ἐμοῦ in Latin and Greek.—It cannot be denied that this has some value. Nevertheless, the argument deducible from it is anything but conclusive.

1. It is by no means an indubitable fact that mei and ἐμοῦ are really cases of the pronoun. The extension of a principle acknowledged in the Greek language might make them the genitive cases of adjectives used pronominally. Thus,

Assume the omission of the article and the extension of the Greek principle to the Latin language, and ἐμοῦ and mei may be cases, not of ἐμὲ and me, but of ἐμὸς and meus.

2. In the classical languages the partitive power was expressed by the genitive.

This is a reason for the evolution of a genitive power. Few such forms exist in the Gothic; part my is not English, nor was dæl min Anglo-Saxon,=part of me, or pars mei.

§ 457. The following differences of form, are found in the different Gothic languages, between the equivalents of mei and tui, the so-called genitives of ego and tu, and the equivalents of meus and tuus, the so-called possessive adjectives. {386}

In this list, those languages where the two forms are alike are not exhibited. This is the case with the Anglo-Saxon and Old Saxon.

In the above-noticed differences of form lie the best reasons for the assumption of a genitive case, as the origin of an adjectival form; and, undoubtedly, in those languages, where both forms occur, it is convenient to consider one as a case and one as an adjective.

§ 458. But this is not the present question. In Anglo-Saxon there is but one form, min and þin=mei and meus, tui and tuus, indifferently. Is this form an oblique case or an adjective?

§ 459. Etymological evidence.—Assuming two powers for the words min and þin, one genitive, and one adjectival, which is the original one? or, going beyond the Anglo-Saxon, assuming that of two forms like meina and meins, the one has been derived from the other, which is the primitive, radical, primary, or original one?

Men, from whom it is generally unsafe to differ, consider that the adjectival form is the derived one; and, as far as forms like mîner, as opposed to mîn, are concerned, the evidence of the foregoing list is in their favour. But what is the case with the Middle Dutch? The genitive mîns is evidently the derivative of mîn.

The reason why the forms like mîner seem derived is because they are longer and more complex than the others. Nevertheless, it is by no means an absolute rule in philology that the least compound form is the oldest. A word may be {387}adapted to a secondary meaning by a change in its parts in the way of omission, as well as by a change in the way of addition. Such is the general statement. Reasons for believing that in the particular cases of the words in question such is the fact, will be found hereafter.

As to the question whether it is most likely for an adjective to be derived from a case, or a case from an adjective, it may be said, that philology furnishes instances both ways. Ours is a case derived, in syntax at least, from an adjective. Cujus (as in cujum pecus) and sestertium are Latin instances of a nominative case being evolved from an oblique one.

§ 460. Syntactic evidence.—If in Anglo-Saxon we found such expressions as dæl min=pars mei, hælf þin=dimidium tui, we should have a reason, as far as it went, for believing in the existence of a genitive with a partitive power. Such instances, however, have yet to be quoted; whilst, even if quoted, they would not be conclusive. Expressions like σὸς πόθος=desiderium tui, σῆ προμηθίᾳ = providentiâ propter te, show the extent to which the possessive expression encroaches on the partitive.

1. The words min or þin, with a power anything rather than possessive, would not for that reason be proved (on the strength of their meaning) to be genitive cases rather than possessive pronouns; since such latitude in the power of the possessive pronoun is borne out by the comparison of languages—πατὲρ ἡμῶν (not ἡμέτερος) in Greek is pater noster (not nostrum) in Latin.

§ 461. Again—as min and þin are declined like adjectives, even as meus and tuus are so declined, we have means of ascertaining their nature from the form they take in certain constructions; thus, minra=meorum, and minre=meæ, are the genitive plural and the dative singular respectively. Thus, too, the Anglo-Saxon for of thy eyes should be eagena þinra, and the Anglo-Saxon for to my widow, should be wuduwan minre; just as in Latin, they would be oculorum tuorum, and viduæ meæ.

If, however, instead of this we find such expressions as eagena þin, or wuduwan min, we find evidence in favour of a {388}genitive case; for then the construction is not one of concord, but one of government, and the words þin and min must be construed as the Latin forms tui and mei would be in oculorum mei, and viduæ mei; viz.: as genitive cases. Now, whether a sufficient proportion of such constructions (real or apparent) exist or not, they have not yet been brought forward.

Such instances have yet to be quoted; whilst even if quoted, they would not be conclusive.

As early as the Mœso-Gothic stage of our language, we find rudiments of the omission of the inflection. The possessive pronouns in the neuter singular sometimes take the inflection, sometimes appear as crude forms, nim thata badi theinata=ἆρον σοῦ τὸν κράββατον (Mark ii. 9.) opposed to nim thata badi thein two verses afterwards. So also with mein and meinata.—Deutsche Grammatik, iv. 470. It is remarkable that this omission should begin with forms so marked as those of the neuter (-ata). It has, perhaps, its origin in the adverbial character of that gender.

Old High German.—Here the nominatives, both masculine and feminine, lose the inflection, whilst the neuter retains it—thin dohter, sîn quenâ, min dohter, sinaz lîb. In a few cases, when the pronoun comes after, even the oblique cases drop the inflection.—Deutsche Grammatik, 474-478.

Middle High German.—Preceding the noun, the nominative of all genders is destitute of inflection; sîn lîb, mîn ere, dîn lîb, &c. Following the nouns, the oblique cases do the same; ine herse sîn.—Deutsche Grammatik, 480. The influence of position should here be noticed. Undoubtedly a place after the substantive influences the omission of the inflection. This appears in its maximum in the Middle High German. In Mœso-Gothic we have mein leik and leik meinata.—Deutsche Grammatik, 470.

§ 463. Now by assuming (which is only a fair assumption) the extension of the Middle High German omission of the inflection to the Anglo-Saxon; and by supposing it to affect the words in question in all positions (i.e., both before and {389}after their nouns), we explain these constructions by a process which, in the mind of the present writer, is involved in fewer difficulties than the opposite doctrine of a genitive case, in words where it is not wanted, and with a termination which is foreign to it elsewhere.

To suppose two adjectival forms, one inflected (min, minre, &c.), and one uninflected, or common to all genders and both numbers (min), is to suppose no more than is the case with the uninflected þe, as compared with the inflected þæt.—See pp. 251-253.

CHAPTER XXXVIII.

§ 464. The remote origin of the weak præterite in -d or -t, has been considered by Grimm, in the Deutsche Grammatik. He maintains that it is the d in d-d, the reduplicate præterite of do. In all the Gothic languages the termination of the past tense is either -da, -ta, -de, -ði, -d, -t, or -ed, for the singular, and -don, -ton, -tûmês, or -ðum, for the plural; in other words, d, or an allied sound, appears once, if not oftener. In the plural præterite of the Mœso-Gothic we have something more, viz. the termination -dêdum; as nas-idêdum, nas-idêduþ, nas-idêdum, from nas-ja; sôk-idêdum, sôk-idêduþ, sôk-idêdum from sôk-ja; salb-ôdêdum, salb-ôdêduþ, sâlb-ôdêdun, from salbô. Here there is a second d. The same takes place with the dual form salb-ôdêduts; and with the subjunctive forms, salb-ôdêdjan, salb-ôdêduts, salb-ôdêdi, salb-ôdêdeits, salb-ôdêdeima, salb-ôdêdeiþ, salb-ôdêdeina. The English phrase, we did salve, as compared with salb-ôdêdum, is confirmatory of this.—Deutsche Grammatik, i. 1042.

§ 465. Some remarks of Dr. Trithen's on the Slavonic præterite, in the Transactions of the Philological Society, induce me to identify the d- in words like moved, &c., with the -t of the passive participles of the Latin language; as found in mon-it-us, voc-at-us, rap-t-us, and probably in Greek forms like τυφ-θ-είς.

l. The Slavonic præterite is commonly said to possess genders: in other words, there is one form for speaking of a past action when done by a male, and another for speaking of a past action when done by a female.

2. These forms are identical with those of the participles, masculine or feminine, as the case may be. Indeed the præterite is a participle; and the fact of its being so accounts for {391}the apparently remarkable fact of its inflection. If, instead of saying ille amavit, the Latins said ille amatus, whilst instead of saying illa amavit they said illa amata, they would exactly use the grammar of the Slavonians.

3. Hence, as one language, the Slavonic gives us the undoubted fact of an active præterite growing out of a passive participle (unless, indeed, we chose to say that both are derived from a common origin); and as the English participle and præterite, when weak, are nearly identical, we have reason for believing that the d, in the English active præterite, is the t in the Latin passive participle.

§ 466. The following extract exhibits Dr. Trithen's remarks on the Slavonic verb:—

PART V.

CHAPTER I.

§ 467. The word syntax is derived from the Greek syn (with or together), and taxis (arrangement). It relates to the arrangement, or putting together of words. Two or more words must be used before there can be any application of studied syntax.

Much that is considered by the generality of grammarians as syntax, can either be omitted altogether, or else be better studied under another name.

§ 468. To reduce a sentence to its elements, and to show that these elements are, 1, the subject, 2, the predicate, 3, the copula; to distinguish between simple terms and complex terms,—this is the department of logic.

To show the difference in force of expression, between such a sentence as great is Diana of the Ephesians, and Diana of the Ephesians is great, wherein the natural order of the subject and predicate is reversed, is a point of rhetoric.

I am moving.—To state that such a combination as I am moving is grammatical, is undoubtedly a point of syntax. Nevertheless it is a point better explained in a separate treatise, than in a work upon any particular language. The expression proves its correctness by the simple fact of its universal intelligibility.

I speaks.—To state that such a combination as I speaks, {393}admitting that I is exclusively the pronoun in the first person, and that speaks is exclusively the verb in the third, is undoubtedly a point of syntax. Nevertheless, it is a point which is better explained in a separate treatise, than in a work upon any particular language. An expression so ungrammatical, involves a contradiction in terms, which unassisted common sense can deal with. This position will again be reverted to.

There is to me a father.—Here we have a circumlocution equivalent to I have a father. In the English language the circumlocution is unnatural. In the Latin it is common. To determine this, is a matter of idiom rather than of syntax.

I am speaking, I was reading.—There was a stage in the Gothic languages when these forms were either inadmissible, or rare. Instead thereof, we had the present tense, I speak, and the past, I spoke. The same is the case with the classical languages in the classical stage. To determine the difference in idea between these pairs of forms is a matter of metaphysics. To determine at what period each idea came to have a separate mode of expression is a matter of the history of language. For example, vas láisands appears in Ulphilas (Matt. vii. 29). There, it appears as a rare form, and as a literal translation of the Greek ἦν διδασκών (was teaching). The Greek form itself was, however, an unclassical expression for ἐδίδασκε. In Anglo-Saxon this mode of speaking became common, and in English it is commoner still.—Deutsche Grammatik, iv. 5. This is a point of idiom involved with one of history.

Swear by your sword—swear on your sword.—Which of these two expressions is right? This depends on what the speaker means. If he mean make your oath in the full remembrance of the trust you put in your sword, and with the imprecation, therein implied, that it shall fail you, or turn against you if you speak falsely, the former expression is the right one. But, if he mean swear with your hand upon your sword, it is the latter which expresses his meaning. To take a different view of this question, and to write as a rule that {394}verbs of swearing are followed by the preposition on (or by) is to mistake the province of the grammar. Grammar tells no one what he should wish to say. It only tells him how what he wishes to say should be said.

Much of the criticism on the use of will and shall is faulty in this respect. Will expresses one idea of futurity, shall another. The syntax of the two words is very nearly that of any other two. That one of the words is oftenest used with a first person, and the other with a second, is a fact, as will be seen hereafter, connected with the nature of things, not of words.

§ 469. The following question now occurs. If the history of forms of speech be one thing, and the history of idioms another; if this question be a part of logic, and that question a part of rhetoric; and if such truly grammatical facts as government and concord are, as matters of common sense, to be left uninvestigated and unexplained, what remains as syntax? This is answered by the following distinction. There are two sorts of syntax; theoretical and practical, scientific and historical, pure and mixed. Of these, the first consists in the analysis and proof of those rules which common practice applies without investigation, and common sense appreciates, in a rough and gross manner, from an appreciation of the results. This is the syntax of government and concord, or of those points which find no place in the present work, for the following reason—they are either too easy or too hard for it. If explained scientifically they are matters of close and minute reasoning; if exhibited empirically they are mere rules for the memory. Besides this they are universal facts of languages in general, and not the particular facts of any one language. Like other universal facts they are capable of being expressed symbolically. That the verb (A) agrees with its pronoun (B) is an immutable fact: or, changing the mode of expression, we may say that language can only fulfil its great primary object of intelligibility when A = B. And so on throughout. A formal syntax thus exhibited, and even devised à priori, is a philological possibility. And it is also the measure of philological anomalies. {395}

§ 470. Pure syntax.—So much for one sort of syntax; viz., that portion of grammar which bears the same relation to the practice of language, that the investigation of the syllogism bears to the practice of reasoning. The positions concerning it are by no means invalidated by such phrases as I speaks (for I speak), &c. In cases like these there is no contradiction; since the peculiarity of the expression consists not in joining two incompatible persons, but in mistaking a third person for a first—and as far as the speaker is concerned, actually making it so. I must here anticipate some objections that may be raised to these views, by stating that I am perfectly aware that they lead to a conclusion which to most readers must appear startling and to some monstrous, viz., to the conclusion that there is no such thing as bad grammar at all; that everything is what the speaker chooses to make it; that a speaker may choose to make any expression whatever, provided it answer the purpose of language, and be intelligible; that, in short, whatever is is right. Notwithstanding this view of the consequence I still am satisfied with the truth of the premises. I may also add that the terms pure and mixed, themselves suggestive of much thought on the subject which they express, are not mine but Professor Sylvester's.

§ 471. Mixed syntax.—That, notwithstanding the previous limitations, there is still a considerable amount of syntax in the English, as in all other languages, may be seen from the sequel. If I undertook to indicate the essentials of mixed syntax, I should say that they consisted in the explanation of combinations apparently ungrammatical; in other words, that they ascertained the results of those causes which disturb the regularity of the pure syntax; that they measured the extent of the deviation; and that they referred it to some principle of the human mind—so accounting for it.

I have gone.—Pure syntax will not explain this. Nevertheless, the expression is good English. The power, however, of both have and gone is different from the usual power of those words. This difference mixed syntax explains. {396}

§ 472. Mixed syntax requires two sorts of knowledge—metaphysical, and historical.

1. To account for such a fact in language as the expression the man as rides to market, instead of the usual expression the man who rides to market, is a question of what is commonly called metaphysics. The idea of comparison is the idea common to the words as and who.

2. To account for such a fact in language as the expression I have ridden a horse is a question of history. We must know that when there was a sign of an accusative case in English the word horse had that sign; in other words that the expression was, originally, I have a horse as a ridden thing. These two views illustrate each other.

§ 473. In the English, as in all other languages, it is convenient to notice certain so-called figures of speech. They always furnish convenient modes of expression, and sometimes, as in the case of the one immediately about to be noticed, account for facts.

§ 474. Personification.—The ideas of apposition and collectiveness account for the apparent violations of the concord of number. The idea of personification applies to the concord of gender. A masculine or feminine gender, characteristic of persons, may be substituted for the neuter gender, characteristic of things. In this case the term is said to be personified.

The cities who aspired to liberty.—A personification of the idea expressed by cities is here necessary to justify the expression.

It, the sign of the neuter gender, as applied to a male or female child, is the reverse of the process.

§ 475. Ellipsis (from the Greek elleipein=to fall short), or a falling short, occurs in sentences like I sent to the bookseller's. Here the word shop or house is understood. Expressions like to go on all fours, and to eat of the fruit of the tree, are reducible to ellipses.

§ 476. Pleonasm (from the Greek pleonazein=to be in excess) occurs in sentences like the king, he reigns. Here the word he is superabundant. In many pleonastic {397}expressions we may suppose an interruption of the sentence, and afterwards an abrupt renewal of it; as the king—he reigns.

The fact of the word he neither qualifying nor explaining the word king, distinguishes pleonasm from apposition.

Pleonasm, as far as the view above is applicable, is reduced to what is, apparently, its opposite, viz., ellipsis.

My banks, they are furnished,—the most straitest sect,—these are pleonastic expressions. In the king, he reigns, the word king is in the same predicament as in the king, God bless him.

The double negative, allowed in Greek and Anglo-Saxon, but not admissible in English, is pleonastic.

The verb do, in I do speak, is not pleonastic. In respect to the sense it adds intensity. In respect to the construction it is not in apposition, but in the same predicament with verbs like must and should, as in I must go, &c.; i. e. it is a verb followed by an infinitive. This we know from its power in those languages where the infinitive has a characteristic sign; as, in German,

Besides this, make is similarly used in Old English.—But men make draw the branch thereof, and beren him to be graffed at Babyloyne.—Sir J. Mandeville.

§ 477. The figure zeugma.—They wear a garment like that of the Scythians, but a language peculiar to themselves.—The verb, naturally applying to garment only, is here used to govern language. This is called in Greek, zeugma (junction).

§ 478. My paternal home was made desolate, and he himself was sacrificed.—The sense of this is plain; he means my father. Yet no such substantive as father has gone before. It is supplied, however, from the word paternal. The sense indicated by paternal gives us a subject to which he can refer. In other words, the word he is understood, according to what is indicated, rather than according to what is expressed. This figure in Greek is called pros to semainomenon (according to the thing indicated). {398}

§ 479. Apposition.—Cæsar, the Roman emperor, invades Britain.—Here the words Roman emperor explain, or define, the word Cæsar; and the sentence, filled up, might stand, Cæsar, that is, the Roman emperor, &c. Again, the words Roman emperor might be wholly ejected; or, if not ejected, they might be thrown into a parenthesis. The practical bearing of this fact is exhibited by changing the form of the sentence, and inserting the conjunction and. In this case, instead of one person, two are spoken of, and the verb invades must be changed from the singular to the plural.

Now the words Roman emperor are said to be in apposition to Cæsar. They constitute, not an additional idea, but an explanation of the original one. They are, as it were, laid alongside (appositi) of the word Cæsar. Cases of doubtful number, wherein two substantives precede a verb, and wherein it is uncertain whether the verb should be singular or plural, are decided by determining whether the substantives be in apposition or the contrary. No matter how many nouns there may be, as long as it can be shown that they are in apposition, the verb is in the singular number.

§ 480. Collectiveness as opposed to plurality.—In sentences like the meeting was large, the multitude pursue pleasure, meeting, and multitude are each collective nouns; that is, although they present the idea of a single object, that object consists of a plurality of individuals. Hence, pursue is put in the plural number. To say, however, the meeting were large would sound improper. The number of the verb that shall accompany a collective noun depends upon whether the idea of the multiplicity of individuals, or that of the unity of the aggregate, shall predominate.

Sand and salt and a mass of iron is easier to bear than a man without understanding.—Let sand and salt and a mass of iron be dealt with as a series of things the aggregate of which forms a mixture, and the expression is allowable.

The king and the lords and commons forms an excellent frame of government.—Here the expression is doubtful. Substitute with for the first and, and there is no doubt as to the propriety of the singular form is. {399}

§ 481. The reduction of complex forms to simple ones.—Take, for instance, the current illustration, viz., the-king-of-Saxony's army.—Here the assertion is, not that the army belongs to Saxony, but that it belongs to the king of Saxony; which words must, for the sake of taking a true view of the construction, be dealt with as a single word in the possessive case. Here two cases are dealt with as one; and a complex term is treated as a single word.

The same reasoning applies to phrases like the two king Williams. If we say the two kings William, we must account for the phrase by apposition.

§ 482. True notion of the part of speech in use.—In he is gone, the word gone must be considered as equivalent to absent; that is, as an adjective. Otherwise the expression is as incorrect as the expression she is eloped. Strong participles are adjectival oftener than weak ones; their form being common to many adjectives.

True notion of the original form.—In the phrase I must speak, the word speak is an infinitive. In the phrase I am forced to speak, the word speak is (in the present English) an infinitive also. In one case, however, it is preceded by to; whilst in the other, the particle to is absent. The reason for this lies in the original difference of form. Speak - to=the Anglo-Saxon sprécan, a simple infinitive; to speak, or speak + to=the Anglo-Saxon to sprécanne, an infinitive in the dative case.

§ 483. Convertibility.—In the English language, the greater part of the words may, as far as their form is concerned, be one part of speech as well as another. Thus the combinations s-a-n-th, or f-r-e-n-k, if they existed at all, might exist as either nouns or verbs, as either substantives or adjectives, as conjunctions, adverbs, or prepositions. This is not the case in the Greek language. There, if a word be a substantive, it will probably end in -s, if an infinitive verb, in -ein, &c. The bearings of this difference between languages like the English and languages like the Greek will soon appear.

At present, it is sufficient to say that a word, {400}originally one part of speech (e.g. a noun), may become another (e.g. a verb). This may be called the convertibility of words.

There is an etymological convertibility, and a syntactic convertibility; and although, in some cases, the line of demarcation is not easily drawn between them, the distinction is intelligible and convenient.

§ 484. Etymological convertibility.—The words then and than, now adverbs or conjunctions, were once cases: in other words, they have been converted from one part of speech to another. Or, they may even be said to be cases, at the present moment; although only in an historical point of view. For the practice of language, they are not only adverbs or conjunctions, but they are adverbs or conjunctions exclusively.

§ 485. Syntactic convertibility.—The combination to err, is at this moment an infinitive verb. Nevertheless it can be used as the equivalent to the substantive error.

To err is human=error is human. Now this is an instance of syntactic conversion. Of the two meanings, there is no doubt as to which is the primary one; which primary meaning is part and parcel of the language at this moment.

The infinitive, when used as a substantive, can be used in a singular form only.

To err=error; but we have no such form as to errs=errors. Nor is it wanted. The infinitive, in a substantival sense, always conveys a general statement, so that even when singular, it has a plural power; just as man is mortal=men are mortal.

§ 486. The adjective used as a substantive.—Of these, we have examples in expressions like the blacks of Africa—the bitters and sweets of life—all fours were put to the ground. These are true instances of conversion, and are proved to be so by the fact of their taking a plural form.

Let the blind lead the blind is not an instance of conversion. The word blind in both instances remains an adjective, and is shown to remain so by its being uninflected.

§ 487. Uninflected parts of speech, used as substantive.—When King Richard III. says, none of your ifs, he uses the word if as a substantive=expressions of doubt. {401}

§ 488. The convertibility of words in English is very great; and it is so because the structure of the language favours it. As few words have any peculiar signs expressive of their being particular parts of speech, interchange is easy, and conversion follows the logical association of ideas unimpeded.

The convertibility of words is in the inverse ratio to the amount of their inflection.

CHAPTER II.

The remaining points connected with the syntax of substantives, are chiefly points of either ellipsis, or apposition.

Ellipsis of substantives.—The historical view of phrases, like Rundell and Bridge's, St. Pauls', &c., shows that this ellipsis is common to the English and the other Gothic languages. Furthermore, it shows that it is met with in languages not of the Gothic stock; and, finally, that the class of words to which it applies, is, there or thereabouts, the same generally.

A. 1. The words most commonly understood, are house and family, or words reducible to them. In Latin, Dianæ=ædem Dianæ.—Deutsche Grammatik, iv. 262.

3. Son, daughter, wife, widow.—Deutsche Grammatik, iv. 262.—Νηλεὺς Κόδρου, Greek.

1. Right and left—supply hand. This is, probably, a real ellipsis. The words right and left, have not yet become true substantives; inasmuch as they have no plural forms. In this respect, they stand in contrast with bitter and sweet; inasmuch as we can say he has tasted both the bitters and sweets of life. Nevertheless, the expression can be refined on.

2. All fours.—To go on all fours. No ellipsis. The word fours, is a true substantive, as proved by its existence as a plural.

From expressions like ποτήριον ψυχροῦ (Matt. xiv. 51), {403}from the Greek, and perfundit gelido (understand latice), from the Latin, we find that the present ellipsis was used with greater latitude in the classical languages than our own.

§ 490. Proper names can only be used in the singular number.—This is a rule of logic, rather than of grammar. When we say the four Georges, the Pitts and Camdens, &c., the words that thus take a plural form, have ceased to be proper names. They either mean—

§ 491. Collocation.—In the present English, the genitive case always precedes the noun by which it is governed—the man's hat=hominis pileus; never the hat man's=pileus hominis.

CHAPTER III.

§ 492. Pleonasm.—Pleonasm can take place with adjectives only in the expression of the degrees of comparison. Over and above the etymological signs of the comparative and superlative degrees, there may be used the superlative words more and most.

§ 493. Collocation.—As a general rule, the adjective precedes the substantive—a good man, not a man good.

When, however, the adjective is qualified by either the expression of its degree, or accompanied by another adjective, it may follow the substantive—

Single simple adjectives thus placed after their substantive, belong to the poetry of England, and especially to the ballad poetry—sighs profound—the leaves green.

§ 494. Government.—The only adjective that governs a case, is the word like. In the expression, this is like him, &c., the original power of the dative remains. This we infer—

1. From the fact that in most languages which have {405}inflections to a sufficient extent, the word meaning like governs a dative case.

2. That if ever we use in English any preposition at all to express similitude, it is the preposition to—like to me, like to death, &c.

Expressions like full of meat, good for John, are by no means instances of the government of adjectives; the really governing words being the prepositions to and for respectively.

The most that can be said, in cases like these, is that particular adjectives determine the use of particular prepositions—thus the preposition of, generally follows the adjective full, &c.

§ 495. The positive degree preceded by the adjective more, is equivalent to the comparative form—e. g., more wise=wiser.

The reasons for employing one expression in preference to the other, depend upon the nature of the particular word used.

When the word is, at one and the same time, of Anglo-Saxon origin and monosyllabic, there is no doubt about the preference to be given to the form in -er. Thus, wis-er is preferable to more wise.

When, however, the word is compound, or trisyllabic, the combination with the word more, is preferable.

Between these two extremes, there are several intermediate forms wherein the use of one rather than another, will depend upon the taste of the writer. The question, however, is a question of euphony, rather than of aught else. It is also illustrated by the principle of not multiplying secondary elements. In such a word as fruit-full-er, there are two additions to the root. The same is the case with the superlative, fruit-full-est.

§ 496. The 9th Chapter of Part IV., should be read carefully. There, there is indicated a refinement upon the current notions as to the power of the comparative degree, {406}and reasons are given for believing that the fundamental notion expressed by the comparative inflexion is the idea of comparison or contrast between two objects.

In this case, it is better in speaking of only two objects to use the comparative degree rather than the superlative—even when we use the definite article the. Thus—

This principle is capable of an application more extensive than our habits of speaking and writing will verify. Thus, to go to other parts of speech, we should logically say—

This statement may be refined on. It is chiefly made for the sake of giving fresh prominence to the idea of duality expressed by the terminations -er and -ter.

§ 497. The absence of inflection simplifies the syntax of adjectives. Violations of concord are impossible. We could not make an adjective disagree with its substantive if we wished.

CHAPTER IV.

§ 498. The syntax of substantives is, in English, simple, from the paucity of its inflections, a condition which is unfavourable towards the evolution of constructional complexities; the most remarkable exception being the phenomenon of convertibility noticed above.

The same is the case with adjectives. The want of inflexion simplifies their syntax equally with that of the substantives.

1. Signs of gender; 2. Signs of case; 3. Signs of number, to a greater extent, and with more peculiarities, than elsewhere.

Furthermore, the pronouns exhibit in a great degree the phenomena of conversion indicated in p. 400.

§ 499. Pleonasm in the syntax of pronouns.—In the following sentences the words in italics are pleonastic.

Of these forms, the first is more common than the second and third, and the fourth more common than the first.

§ 500. The fourth has another element of importance. It has given rise to the absurd notion that the genitive case in -s (father-s) is a contraction from his (father his).

To say nothing about the inapplicability of this rule to feminine genders, and plural numbers, the whole history of the Indo-Germanic languages is against it. {408}

4. The word his itself must be accounted for; and that cannot be done by assuming to be he + his.

5. The -s in father's is the -is in patris, and the -ος in πατέρος.

§ 501. The preceding examples illustrate an apparent paradox, viz., the fact of pleonasm and ellipsis being closely allied. The king he is just, dealt with as a single sentence, is undoubtedly pleonastic. But it is not necessary to be considered as a mere simple sentence. The king—may represent a first sentence incomplete, whilst he is just represents a second sentence in full. What is pleonasm in a single sentence, is ellipsis in a double one.

CHAPTER V.

§ 502. Personal pronouns.—The use of the second person plural instead of the second singular has been noticed in p. 246. This use of one number for another is current throughout the Gothic languages. A pronoun so used is conveniently called the pronomen reverentiæ.

§ 503. In English, however, there is a second change over and above the change of number, viz. that of case. We not only say ye instead of thou, but you instead of ye.—(See p. 245).

Mr. Guest remarks, "that at one time the two forms ye and you seem to have been nearly changing place in our language.

In German and the Danish the pronomen reverentiæ is got at by a change, not of number, but of person—in other words, the pronoun of the third person is used instead of that of the second; just as if, in the English, we said will they walk=will you walk, will ye walk, wilt thou walk.

the me is expletive, and is equivalent to for me. This expletive use of the dative is conveniently called the dativus ethicus. It occurs more frequently in the Latin than in the {410}English, and more frequently in the Greek than in the Latin.

§ 505. The reflected personal pronoun.—In the English language there is no equivalent to the Latin se, the German sich, and the Scandinavian sik, and sig.

It follows from this that the word self is used to a greater extent than would otherwise be the case.

He strikes him is ambiguous; inasmuch as him may mean either the person who strikes or some one else. In order to be clear we add the word self when the idea is reflective. He strikes himself is, at once, idiomatic, and unequivocal.

As a general rule, therefore, whenever we use a verb reflectively we use the word self in combination with the personal pronoun.

Yet this was not always the case. The use of the simple personal pronoun was current in Anglo-Saxon, and that, not only for the two first persons, but for the third as well.

The exceptions to this rule are either poetical expressions, or imperative moods.

§ 506. Reflective neuters.—In the phrase I strike me the verb strike is transitive; in other words, the word me expresses the object of an action, and the meaning is different from the meaning of the simple expression I strike.

In the phrase I fear me (used by Lord Campbell in his Lives of the Chancellors), the verb fear is intransitive or neuter; in other words, the word me (unless, indeed, fear mean terrify) {411}expresses no object of any action at all; whilst the meaning is the same as in the simple expression I fear.

Here the reflective pronoun appears out of place, i. e., after a neuter or intransitive verb.

Such a use, however, is but the fragment of an extensive system of reflective verbs thus formed, developed in different degrees in the different Gothic languages; but in all more than in the English.

§ 507. Equivocal reflectives.—The proper place of the reflective is after the verb.

The proper place of the governing pronoun is, in the indicative and subjunctive moods, before the verb.

Hence in expressions like the preceding there is no doubt as to the power of the pronoun.

The imperative mood, however, sometimes presents a complication. Here the governing person may follow the verb.

Mount ye=either be mounted, or mount yourselves. In phrases like this, and in phrases

the construction is ambiguous. Ye may either be a nominative case governing the verb busk, or an accusative case governed by it.

CHAPTER VI.

ON THE SYNTAX OF THE DEMONSTRATIVE PRONOUNS, AND THE PRONOUNS OF THE THIRD PERSON.

§ 508. Reasons have been given in p. 249, for considering the so-called pronouns of the third person (he, she, it, they) demonstrative rather than truly personal.

§ 509. As his, and her, are genitive cases (and not adjectives), there is no need of explaining such combinations as his mother, her father, inasmuch as no concord of gender is expected. The expressions are respectively equivalent to

§ 510. From p. 250, it may be seen that its is a secondary genitive, and it may be added, that it is of late origin in the language. The Anglo-Saxon form was his, the genitive of he for the neuter and masculine equally. Hence, when, in the old writers, we meet his, where we expect its, we must not suppose that any personification takes place, but simply that the old genitive common to the two genders is used in preference to the modern one limited to the neuter, and irregularly formed. This has been illustrated by Mr. Guest.

"The genitive its is of late introduction into our language. Though used by our dramatists and many of their cotemporaries, it does not occur in the versions of our Bible, the substitute being his or the compound term thereof."—Phil. Trans., No. 25.

§ 511. For the archaic and provincial use of him and he for it see ibid.; remembering that the two cases are different. His for its is an old form retained: him and he for it are really changes of gender.

§ 512. Take them things away.—Here we have them for those. The expression, although not to be imitated, is explained by the originally demonstrative power of them.

Sometimes the expression is still more anomalous, and we hear the so-called nominative case used instead of the accusative. In the expression take they things away, the use of they for them (itself for those) is similarly capable of being, down to a certain period of our language, explained as an archaism. The original accusative was þa, and þo: the form in -m being dative.

§ 513. This and that.—The remarks upon the use of these words in certain expressions is brought at once to the Latin scholar by the quotation of the two following lines from Ovid, and the suggestion of a well-known rule in the Eton Latin Grammar.

Here hic (=this or the one) refers to the antecedent last named (the air); whilst ille (=that or the other) refers to the antecedent first named (the sea).

Now on the strength of this example, combined with others, it is laid down as a rule in Latin that hic (this) refers to the last-named antecedent, ille to the first-named.

Suppose we say John's is a good sword and so is Charles's; this cut through a thick rope, the other cut through an iron rod. Or instead of saying this and that we may say the one and the other. It is clear that, in determining to which of the {414}two swords the respective demonstratives refer, the meaning will not help us at all, so that our only recourse is to the rules of grammar; and it is the opinion of the present writer that the rules of grammar will help us just as little. The Latin rule is adopted by scholars, but still it is a Latin rule rather than an English one.

The truth is, that it is a question which no authority can settle; and all that grammar can tell us is (what we know without it) that this refers to the name of the idea which is logically the most close at hand, and that to the idea which is logically the most distant.

What constitutes nearness or distance of ideas, in other words, what determines the sequence of ideas is another question. That the idea, however, of sequence, and, consequently of logical proximity and logical distance, is the fundamental idea in regard to the expressions in question is evident from the very use of the words this and that.

1. The idea to which the name was last given, or (changing the expression) the name of the last idea may be the nearest idea in the order of sequence, and, consequently, the idea referred to by the pronoun of proximity. In this case the idea closest at hand to the writer of the second line of the couplet quoted above was the idea of the atmosphere (aer), and it was, consequently, expressed by (this) hic.

2. Or the idea to which the name was first given, or (changing the expression) the name of the first idea may be the nearest idea in the order of sequence, and consequently the idea referred to it by the pronoun of proximity; inasmuch as the idea which occurs first is the most prominent one, and what is prominent appears near. In this case, the idea closest at hand to the writer of the second line of the couplet quoted above would have been the idea of the sea (pontus), and it would, consequently, have been the idea expressed by this (hic).

As Ovid, however, considered the idea at the end of the last half of one sentence to be the idea nearest to the {415}beginning of the next, we have him expressing himself as he does. On the other hand, it is easy to conceive a writer with whom the nearest idea is the idea that led the way to the others.

As I believe that one and the same individual may measure the sequence of his ideas sometimes according to one of these principles, and sometimes according to another, I believe that all rules about the relations of this and that are arbitrary.

It is just a matter of chance whether a thinker take up his line of ideas by the end or by the beginning. The analogies of such expressions as the following are in favour of this, in English, applying to the first subject, that to the second; since the word attorney takes the place of this, and applies to the first name of the two, i. e., to Thurlow.

CHAPTER VII.

§ 515. The undoubted constructions of the word self, in the present state of the cultivated English, are three-fold.

1. Government.—In my-self, thy-self, our-selves, and your-selves, the construction is that of a common substantive with an adjective or genitive case. My-self=my individuality, and is similarly construed—mea individualitas (or persona), or mei individualitas (or persona).

2. Apposition.—In him-self and them-selves, when accusative, the construction is that of a substantive in apposition with a pronoun. Him-self=him, the individual.

3. Composition.—It is only, however, when himself and themselves, are in the accusative case, that the construction is appositional. When they are used as nominatives, it must be explained on another principle. In phrases like

There is neither apposition nor government; him and them, being neither related to my and thy, so as to be governed, nor yet to he and they, so as to form an apposition. In order to come under one of these conditions, the phrases should be either he his self (they their selves), or else he he self (they they selves). In this difficulty, the only logical view that can be taken of the matter, is to consider the words himself and themselves, not as two words, but as a single word compounded; and even then, the compound will be of an irregular kind; inasmuch as the inflectional element -m, is dealt with as part and parcel of the root.

§ 516. Her-self.—The construction here is ambiguous. It is one of the preceding constructions. Which, however it is, {417}is uncertain; since her may be either a so-called genitive, like my, or an accusative like him.

Itself—is also ambiguous. The s may represent the -s in its, as well as the s- in self.

§ 517. In the exhibition of the second construction of the word self it was assumed that the case was a case of apposition, and that self was substantival in character. Nevertheless, this is by no means a necessary phenomenon. Self might, as far as its power is determined by its construction alone, in words like himself as easily be an adjective as a substantive. In which case the construction would be a matter, not of apposition, but of agreement. To illustrate this by the Latin language, himself, might equal either eum personam (him, the person), or eum personalem (him personal). The evidence, however, of the forms like myself, as well as other facts adduceable from comparative philology, prove the substantival character of self. On the other hand, it ought not to be concealed that another word, whereof the preponderance of the adjectival over the substantival power is undoubted, is found in the Old English, with just the same inconsistency as the word self; i.e., sometimes in government (like a substantive), and sometimes in either concord or apposition, like a word which may be either substantive or adjective. This word is one; the following illustrations of which are from Mr. Guest.—Phil. Trans. No. 22.

§ 518. As to the inflection of the word -self, all its compounds are substantives; inasmuch as they all take plural forms as far as certain logical limitations will allow them to do so—ourselves, yourselves, themselves.

Myself, thyself, himself, itself, and herself, are naturally singular, and under no circumstances can become plural.

Themselves is naturally plural, and under no circumstances can become singular.

Ourselves and yourselves are naturally plural; yet under certain circumstances they become singular.

In respect to the inflection in the way of case, there are no logical limitations whatever. There is nothing against the existence of a genitive form self's except the habit of the English language not to use one, founded on the little necessity for so doing.—Are you sure this is your own? Yes, I am sure it is my own self's. Such an expression is both logic and grammar.

When an adjective intervenes between self and its personal pronoun the construction is always in the way of government; in other words, the personal pronoun is always put in the genitive case.

§ 519. The construction of self and a personal pronoun with a verb may be noticed in this place. It is only in the case of the two pronouns of the singular number that any doubt can arise.

1. When myself or thyself stands alone, the verb that follows is in the third person—myself is (not am) weak, thyself is (not art) weak. Here the construction is just the same as in the proposition my body is weak.

2. When myself or thyself is preceded by I or thou, the verb that follows is in the first person—I, myself, am (not is) weak; thou, thyself, art (not is) weak.

CHAPTER VIII.

§ 520. The possessive pronouns fall into two classes. The first class contains the forms connected, partially in their etymology and wholly in their syntax, with my and thy, &c. The second class contains the forms connected, partially in their etymology and wholly in their syntax, with mine and thine, &c.

The first class is the class of what may be called the oblique possessives; the name being founded upon the etymological fact of their being connected with the oblique cases of the pronominal inflection.—My, thy, his (as in his book), her, its (as in its book), our, your, their. These are conveniently considered as the equivalents to the Latin forms mei, tui, ejus, nostrum, vestrum, eorum.

The second class is the class of what may be called the absolute possessives; the name being founded upon the syntactic fact of their being able to form the term of a proposition by themselves; as whose is this? Mine (not my).—Mine, thine, his (as in the book is his), hers, ours, yours, theirs are conveniently considered as the equivalents to the Latin forms meus, mea, meum; tuus, tua, tuum; suus, sua, suum; noster, nostra, nostrum; vester, vestra, vestrum. How far either or both of these two classes of pronouns are cases, or adjectives, is a point of etymology that has already been noticed (Part IV., chap. 37).

How far either or both are cases or adjectives is, in syntax, a matter of indifference.

§ 521. There is, however, a palpable difference between the construction of my and mine. We cannot say this is mine hat, and we cannot say this hat is my. Nevertheless, this {420}difference is not explained by any change of construction from that of adjectives to that of cases. As far as the syntax is concerned the construction of my and mine is equally that of an adjective agreeing with a substantive, and of a genitive (or possessive) case governed by a substantive.

Now a common genitive case can be used in two ways; either as part of a term, or as a whole term (i. e., absolutely).—1. As part of a term—this is John's hat. 2. As a whole term—this hat is John's.

And a common adjective can be used in two ways; either as part of a term, or as a whole term (i. e., absolutely).—1. As part of a term—these are good hats. 2. As a whole term—these hats are good.

Now whether we consider my, and the words like it, as adjectives or cases, they possess only one of the properties just illustrated, i. e., they can only be used as part of a term—this is my hat; not this hat is my.

And whether we consider mine, and the words like it, as adjectives or cases, they possess only one of the properties just illustrated, i. e., they can only be used as whole terms, or absolutely—this hat is mine; not this is mine hat.

For a full and perfect construction whether of an adjective or a genitive case, the possessive pronouns present the phenomenon of being, singly, incomplete, but, nevertheless, complimentary to each other when taken in their two forms.

In the absolute construction of a genitive case, the term is formed by the single word only so far as the expression is concerned. A substantive is always understood from what has preceded.—This discovery is Newton's=this discovery is Newton's discovery.

The same with adjectives.—This weather is fine=this weather is fine weather.

And the same with absolute pronouns.—This hat is mine=this hat is my hat; and this is a hat of mine=this is a hat of my hats.

In respect to all matters of syntax considered exclusively, it is so thoroughly a matter of indifference whether a word be an adjective or a genitive case that Wallis considers the {421}forms in -'s like father's, not as genitive cases but as adjectives. Looking to the logic of the question alone he is right, and looking to the practical syntax of the question he is right, also. He is only wrong on the etymological side of the question.

CHAPTER IX.

§ 522. The word that, although originally, when a demonstrative pronoun, a neuter singular, is now used as a relative for all genders, and both numbers.

§ 523. Etymologically, which is no true neuter of who, but a compound word. It is used, however, with less latitude than that. The beginning of the Lord's Prayer exhibits it in combination with a masculine noun. Generally, however, it is confined to the neuter gender; in which it is common to both numbers.

§ 524. Which has so nearly replaced what that the general use of this last word with its proper power, as a neuter relative, is, in the present English, vulgar, e.g.,

In one case, however, what is used as a true relative, viz., when the antecedent is either this or that.

§ 525. The word as, properly a conjunction, is occasionally used as a relative—the man as rides to market.

This expression is not to be imitated. It ought, however, to be explained. As is a conjunction denoting comparison. The ideas of comparison and equivalence are allied. The relative is ex vi termini the equivalent, in one part of a sentence, to the antecedent in another.

§ 526. It is necessary that the relative be in the same gender as the antecedent—the man who—the woman who—the thing which.

§ 527. It is necessary that the relative be in the same number with the antecedent. As, however, who, which, whom, are equally singular and plural, and as what, which is really singular, is not used as a relative, the application of this law is limited to the word whose. Now whose is, etymologically, a genitive case, and a genitive case of the singular number. Hence the expression the men whose daggers stabbed Cæsar can only be justified by considering that the word whose is plural as well as singular. Such is the case. If not the expression is as illogical as homines cujus sicæ, &c. would be in Latin.

§ 528. It is not necessary for the relative to be in the same case with its antecedent.

§ 529. The reason why the relative must agree with its antecedent in both number and gender, whilst it need not agree with it in case, is found in the following observations.

1. All sentences containing a relative contain two verbs—John who (1) trusts me (2) comes here.

3. Whilst, however, the actions are two in number, the {424}person or thing which does, or suffers them is single—John.

4. He (she or it) is single ex vi termini. The relative expresses the identity between the subjects (or objects) of the two actions. Thus who=John, or is another name for John.

5. Things and persons that are one and the same, are of one and the same gender. The John who trusts is necessarily of the same gender with the John who comes.

6. Things and persons that are one and the same, are of one and the same number. The number of Johns who trust, is the same as the number of Johns who come. Both these elements of concord are immutable.

7. But a third element of concord is not immutable. The person or thing that is an agent in the one part of the sentence, may be the object of an action in the other. The John whom I trust may trust me also. Hence

As the relative is only the antecedent in another form, it may change its case according to the construction.

§ 530. The books I want are here.—This is a specimen of a true ellipsis. In all such phrases in full, there are three essential elements.

3. The word which connects the two propositions, and without which, they naturally make separate, independent, unconnected statements.

Now, although true and unequivocal ellipses are scarce, {425}the preceding is one of the most unequivocal kind—the word which connects the two propositions being wanting.

§ 531. One or two points connected with the construction of those sentences wherein relative pronouns occur, are necessary to be familiarly understood in order for us to see our way clearly to certain real and apparent anomalies in the syntax of this class of words.

1. Every sentence wherein a relative occurs, is complex, i.e., it consists of two propositions—the man who rides is come=(1) the man is come; (2) who rides. Here the relative who has no meaning in itself, but takes a meaning from the noun of the preceding clause.

2. The relative is the demonstrative or personal pronoun under another form.—The two propositions (1) the man is come; (2) who rides=(1) the man is come; (2) he rides.

3. The demonstrative or personal pronoun is the substantive in another form.—The two propositions (1) the man is come; (2) he rides=(1) the man is come; (2) the man rides.

4. Hence the relative is the equivalent to a demonstrative pronoun, or to a substantive, indifferently.

5. But the relative is the equivalent to the pronoun and substantive, and something more. In sentences like

The identity between the person mentioned in the two propositions is implied, not expressed. This the relative expresses; and hence its use in languages.

6. From these observations we get a practical rule for determining doubtful constructions.

a. Reduce the sentence to the several propositions (which are never less than two) which it contains.

b. Replace the relative by its equivalent personal or demonstrative pronoun, or by its equivalent substantive.

c. The case of the demonstrative or substantive, is the case of the relative also.

Observe.—The words, according to the current etymology, indicate an explanation which, rightly or wrongly, has been urged in favour of expressions like the one in question, and which will be noticed in a future chapter.

§ 532. Observe.—That three circumstances complicate the syntax of the relative pronoun.

1. The elliptic form of the generality of the sentences wherein it follows the word than.

When the relative and antecedent are in different cases, and the relative is omitted, the antecedent is sometimes put in the case of the relative.

And so (as shown by Mr. Guest, Philological Transactions), Shakspeare has really written,—

The reason of this is clear. The verb that determines {427}the case of the relative is brought in contact with the antecedent, and the case of the antecedent is accommodated to the case of the relative.

The Greek phrase, χρῶμαι βιβλίοις οἷς ἔχω, is an instance of the converse process.

§ 534. When there are two words in a clause, each capable of being an antecedent, the relative refers to the latter.

Nevertheless, it is defensible, on the supposition that Solomon-the-son-of-David is a single many-worded name.

CHAPTER X.

All difficulties about the cases of the interrogative pronoun may be determined by framing an answer, and observing the case of the word with which the interrogative coincides. Whatever be the case of this word will also be the case of the interrogative.

Note.—The answer should always be made by means of a pronoun, as, by so doing we distinguish the accusative case from the nominative.

Note.—And, if necessary, it should be made in full. Thus the full answer to whom do you say that they seek? is, I say that they seek him.

§ 536. Nevertheless, such expressions as whom do they say that it is? are common, especially in oblique questions. The following examples are Mr. Guest's.—Philological Transactions.

Two circumstances encourage this confusion. 1. The presence of a second verb, which takes the appearance of a governing verb. 2. The omission of a really oblique antecedent or relative. 3. The use of accusative for nominative forms in the case of personal pronouns.

§ 537. The presence of a second verb, &c.—Tell me whom she is. Here tell is made to govern whom, instead of whom being left, as who, to agree with she.

§ 538. The omission, &c.—Tell me whom she is you love. Here the full construction requires a second pronoun—tell me who she is whom you love; or else, tell me her whom you love.

§ 539. To the question, who is this? many would answer not I, but me. This confusion of the case in the answer favours a confusion of case in the question.

It is clear that much of this reasoning applies to the relative powers of who, as well as to the interrogative.

But, it is possible that there may be no incorrectness at all: insomuch as whom may have become a true nominative. Mr. Guest has truly remarked that such is the case in the Scandinavian language, where hve-m=who=qui.

This view, if true, justifies the use of whom after the conjunctions than and as; so that the expression,—

The reason of this is clear. Whom is supposed to be admissible, not because the sentence admits an accusative case; but because custom has converted it into a nominative. For my own part, I doubt the application of the Danish rule to the English language. Things may be going that way, but they have not, as yet, gone far enough.

CHAPTER XI.

§ 540. In all sentences containing the statement of a reciprocal or mutual action there are in reality two assertions, viz., the assertion that A. strikes (or loves) B., and the assertion that B. strikes (or loves) A.; the action forming one, the reaction another. Hence, if the expressions exactly coincided with the fact signified, there would always be two propositions. This, however, is not the habit of language. Hence arises a more compendious form of expression, giving origin to an ellipsis of a peculiar kind. Phrases like Eteocles and Polynices killed each other are elliptical, for Eteocles and Polynices killed—each the other. Here the second proposition expands and explains the first, whilst the first supplies the verb to the second. Each, however, is elliptic. The first is without the object, the second without the verb. That the verb must be in the plural (or dual) number, that one of the nouns must be in the nominative case, and that the other must be objective, is self-evident from the structure of the sentence; such being the conditions of the expression of the idea. An aposiopesis takes place after a plural verb, and then there follows a clause wherein the verb is supplied from what went before.

§ 541. This is the syntax. As to the power of the words each and one in the expression (each other and one another), I am not prepared to say that in the common practice of the English language there is any distinction between them. A distinction, however, if it existed would give strength to our language. Where two persons performed a reciprocal action on another, the expression might be one another; as Eteocles and Polynices killed one another. Where more than two {432}persons were engaged on each side of a reciprocal action the expression might be each other; as, the ten champions praised each other.

This amount of perspicuity is attained, by different processes, in the French, Spanish, and Scandinavian languages.

1. French.—Ils (i.e., A. and B.) se battaient—l'un l'autre. Ils (A. B. C.) se battaient—les uns les autres. In Spanish, uno otro=l'un l'autre, and unos otros=les uns les autres.

2. Danish.—Hinander=the French l'un l'autre; whilst hverandre=les uns les autres.

The Lapplandic, and, probably other languages, have the same elements of perspicuity.

CHAPTER XII.

§ 542. Different nations have different methods of expressing indeterminate propositions.

Sometimes it is by the use of the passive voice. This is the common method in Latin and Greek, and is also current in English—dicitur, λέγεται, it is said.

Sometimes the plural pronoun of the third person is used. This also is an English locution—they say=the world at large says.

The word man has an indeterminate sense in the Modern German; as, man sagt=they say.

The word man was also used indeterminately in the Old English, although it is not so used in the Modern.—Deutsche Grammatik.

In the Old English, the form man often lost the -n, and became me.—Deutsche Grammatik. This form is also extinct.

The present indeterminate pronoun is one; as, one says=they say=it is said=man sagt, German=on dit, French=si dice, Italian.

It has been stated in p. 257, that the indeterminate pronoun one has no etymological connection with the numeral one; but that it is derived from the French on=homme=homo=man; and that it has replaced the Old English, man or me.

§ 543. Two other pronouns, or, to speak more in accordance with the present habit of the English language, one {434}pronoun, and one adverb of pronominal origin are also used indeterminately viz., it and there.

§ 544. It can be either the subject or the predicate of a sentence,—it is this, this is it, I am it, it is I. When it is the subject of a proposition, the verb necessarily agrees with it, and can be of the singular number only; no matter what be the number of the predicate—it is this, it is these.

When it is the predicate of a proposition, the number of the verb depends upon the number of the subject. These points of universal syntax are mentioned here for the sake of illustrating some anomalous forms.

§ 545. There can only be the predicate of a subject. It differs from it in this respect. It follows also that it must differ from it in never affecting the number of the verb. This is determined by the nature of the subject—there is this, there are these.

When we say there is these, the analogy between the words there and it misleads us; the expression being illogical.

Furthermore, although a predicate, there always stands in the beginning of propositions, i.e., in the place of the subject. This also misleads.

§ 546. Although it, when the subject, being itself singular, absolutely requires that its verb should be singular also, there is a tendency to use it incorrectly, and to treat it as a plural. Thus, in German, when the predicate is plural, the verb joined to the singular form es (=it) is plural—es sind menschen, literally translated=it are men; which, though bad English, is good German.

CHAPTER XIII.

§ 547. The rule of most practical importance about the articles is the rule that determines when the article shall be repeated as often as there is a fresh substantive, and when it shall not.

When two or more substantives following each other denote the same object, the article precedes the first only. We say the secretary and treasurer (or, a secretary and treasurer), when the two offices are held by one person.

When two or more substantives following each other denote different objects, the article is repeated, and precedes each. We say the (or a) secretary and the (or a) treasurer, when the two offices are held by different persons.

CHAPTER XIV.

§ 548. The numeral one is naturally single. All the rest are naturally plural.

Nevertheless such expressions—one two (=one collection of two), two threes (=two collections of three), are legitimate. These are so because the sense of the word is changed. We may talk of several ones just as we may talk of several aces; and of one two just as of one pair.

Expressions like the thousandth-and-first are incorrect. They mean neither one thing nor another: 1001st being expressed by the thousand-and-first, and 1000th + 1st being expressed by the thousandth and the first.

Here it may be noticed that, although I never found it to do so, the word odd is capable of taking an ordinal form. The thousand-and-odd-th is as good an expression as the thousand-and-eight-th.

The construction of phrases like the thousand-and-first is the same construction as we find in the king-of-Saxony's army.

§ 549. It is by no means a matter of indifference whether we say the two first or the first two.

The captains of two different classes at school should be called the two first boys. The first and second boys of the same class should be called the first two boys. I believe that when this rule is attended to, more is due to the printer than to the author: such, at least, is the case with myself.

CHAPTER XV.

§ 550. For the purposes of syntax it is necessary to divide verbs into the five following divisions: transitive, intransitive, auxiliary, substantive, and impersonal.

Transitive verbs.—In transitive verbs the action is never a simple action. It always affects some object or other,—I move my limbs; I strike my enemy. The presence of a transitive verb implies also the presence of a noun; which noun is the name of the object affected. A transitive verb, unaccompanied by a noun, either expressed or understood, is a contradiction in terms. The absence of the nouns, in and of itself, makes it intransitive. I move means, simply, I am in a state of moving. I strike means, simply, I am in the act of striking. Verbs like move and strike are naturally transitive.

Intransitive verbs.—An act may take place, and yet no object be affected by it. To hunger, to thirst, to sleep, to wake, are verbs that indicate states of being, rather than actions affecting objects. Verbs like hunger, and sleep, are naturally intransitive.

Many verbs, naturally transitive, may be used as intransitive,—e.g., I move, I strike, &c.

Many verbs, naturally intransitive, may be used as transitives,—e.g., I walked the horse=I made the horse walk.

This variation in the use of one and the same verb is of much importance in the question of the government of verbs.

A. Transitive verbs are naturally followed by some noun or other; and that noun is always the name of something affected by them as an object. {438}

B. Intransitive verbs are not naturally followed by any noun at all; and when they are so followed, the noun is never the name of anything affected by them as an object.

Nevertheless, intransitive verbs may be followed by nouns denoting the manner, degree, or instrumentality of their action,—I walk with my feet=incedo pedibus.

§ 552. The verb substantive has this peculiarity, viz. that for all purposes of syntax it is no verb at all. I speak may, logically, be reduced to I am speaking; in which case it is only the part of a verb. Etymologically, indeed, the verb substantive is a verb; inasmuch as it is inflected as such: but for the purposes of construction, it is a copula only, i.e., it merely denotes the agreement or disagreement between the subject and the predicate.

This does not apply to the infinitive mood. The infinitive mood of the so-called verb substantive is a noun; not, however, because it is a verb substantive, but because it is an infinitive mood.

CHAPTER XVI.

§ 553. The verb must agree with its subject in person, I walk, not I walks: he walks, not he walk.

It must also agree with it in number,—we walk, not we walks: he walks, not he walk.

Clear as these rules are, they require some expansion before they become sufficient to solve all the doubtful points of English syntax connected with the concord of the verb.

A. It is I, your master, who command you. Query? would it is I, your master, who commands you, be correct? This is an example of a disputed point of concord in respect to the person of the verb.

B. The wages of sin is death. Query? would the wages of sin are death be correct? This is an example of a disputed point of concord in respect to the number of the verb.

§ 554. In respect to the concord of person the following rules will carry us through a portion of the difficulties.

Rule.—In sentences, where there is but one proposition, when a noun and a pronoun of different persons are in apposition, the verb agrees with the first of them,—I, your master, command you (not commands): your master, I, commands you (not command).

To understand the nature of the difficulty, it is necessary to remember that subjects may be extremely complex as well as perfectly simple; and that a complex subject may contain, at one and the same time, a noun substantive and a pronoun,—I, the keeper; he, the merchant, &c.

Now all noun-substantives are naturally of the third person—John speaks, the men run, the commander gives orders. Consequently the verb is of the third person also. {440}

But, the pronoun with which such a noun-substantive may be placed in apposition, may be a pronoun of either person, the first or second: I or thou—I the commander—thou the commander.—In this case the construction requires consideration. With which does the verb agree? with the substantive which requires a third person? or with the pronoun which requires a first or second?

Undoubtedly the idea which comes first is the leading idea; and, undoubtedly, the idea which explains, qualifies, or defines it, is the subordinate idea: and, undoubtedly, it is the leading idea which determines the construction of the verb. We may illustrate this from the analogy of a similar construction in respect to number—a man with a horse and a gig meets me on the road. Here the ideas are three; nevertheless the verb is singular. No addition of subordinate elements interferes with the construction that is determined by the leading idea. In the expression I, your master, the ideas are two; viz. the idea expressed by I, and the idea expressed by master. Nevertheless, as the one only explains or defines the other, the construction is the same as if the idea were single. Your master, I, is in the same condition. The general statement is made concerning the master, and it is intended to say what he does. The word I merely defines the expression by stating who the master is. Of the two expressions the latter is the awkwardest. The construction, however, is the same for both.

From the analysis of the structure of complex subjects of the kind in question, combined with a rule concerning the position of the subject, which will soon be laid down, I believe that, for all single propositions, the foregoing rule is absolute.

Rule.—In all single propositions the verb agrees in person with the noun (whether substantive or pronoun) which comes first.

§ 555. But the expression it is I, your master, who command (or commands) you, is not a single proposition. It is a sentence containing two propositions.

Here, the word master is, so to say, undistributed. It may belong to either clause of the sentence, i.e., the whole sentence may be divided into

This is the first point to observe. The next is that the verb in the second clause (command or commands) is governed, not by either the personal pronoun or the substantive, but by the relative, i.e., in the particular case before us, not by either I or master, but by who.

And this brings us to the following question—with which of the two antecedents does the relative agree? with I or with master?

Rule 1.—When the two antecedents are in the same proposition, the relative agrees with the first. Thus—

Rule 2.—When the two antecedents are in different propositions, the relative agrees with the second. Thus—

This, however, is not all. What determines whether the two antecedents shall be in the same or in different propositions? I believe that the following rules for what may be called the distribution of the substantive antecedent will bear criticism.

Rule 1. That when there is any natural connection between the substantive antecedent and the verb governed by the relative, the antecedent belongs to the second clause. Thus, in the expression just quoted, the word master is logically connected with the word command; and this fact makes the expression, It is I your master who commands you the better of the two.

Rule 2. That when there is no natural connection between the substantive antecedent and the verb governed by the {442}relative, the antecedent belongs to the first clause. It is I, John, who command (not commands) you.

3. Of such two antecedents the relative agrees with the one which stands in the same proposition with itself.

4. Which position is determined by the connection or want of connection between the substantive antecedent and the verb governed by the relative.

Respecting the person of the verb in the first proposition of a complex sentence there is no doubt. I, your master, who commands you to make haste, am (not is) in a hurry. Here, I am in a hurry is the first proposition; who commands you to make haste, the second.

It is not difficult to see why the construction of sentences consisting of two propositions is open to an amount of latitude which is not admissible in the construction of single propositions. As long as the different parts of a complex idea are contained within the limits of a single proposition, their subordinate character is easily discerned. When, however, they amount to whole propositions, they take the appearance of being independent members of the sentence.

§ 556. The concord of number.—It is believed that the following three rules will carry us through all difficulties of the kind just exhibited.

Rule 1. That the verb agrees with the subject, and with nothing but the subject. The only way to justify such an expression as the wages of sin is death, is to consider death not as the subject, but as the predicate; in other words, to consider the construction to be, death is the wages of sin.

Rule 2. That, except in the case of the word there (p. 434), the word which comes first is always the subject, until the contrary be proved. {443}

Rule 3. That no number of connected singular nouns can govern a plural verb, unless they be connected by a copulative conjunction. The sun and moon shine,—the sun in conjunction with the moon shines.

§ 557. Plural subjects with singular predicates.—The wages of sin are death.—Honest men are the salt of the earth.

Singular subjects with plural predicates.—These constructions are rarer than the preceding: inasmuch as two or more persons (or things) are oftener spoken of as being equivalent to one, than one person (or thing) is spoken of as being equivalent to two or more.

CHAPTER XVII.

§ 558. The government of verbs is of two sorts, (1.) objective, and (2.) modal.

It is objective where the noun which follows the verb is the name of some object affected by the action of the verb,—as he strikes me; he wounds the enemy.

It is modal when the noun which follows the verb is not the name of any object affected by the verb, but the name of some object explaining the manner in which the action of the verb takes place, the instrument with which it is done, the end for which it is done, &c.

The government of all transitive verbs is necessarily objective. It may also be modal,—I strike the enemy with the sword=ferio hostem gladio.

The government of all intransitive verbs can only be modal,—I walk with the stick. When we say, I walk the horse, the word walk has changed its meaning, and signifies make to walk, and is, by the very fact of its being followed by the name of an object, converted from an intransitive into a transitive verb.

The modal construction may also be called the adverbial construction; because the effect of the noun is akin to that of an adverb,—I fight with bravery=I fight bravely: he walks a king=he walks regally. The modal (or adverbial) construction (or government) sometimes takes the appearance of the objective: inasmuch as intransitive verbs are frequently followed by a substantive; which substantive is in the objective case. Nevertheless, this is no proof of government. For a verb to be capable of governing an objective case, it must be a verb signifying an action affecting an object: and {445}if there be no such object, there is no room for any objective government. To break the sleep of the righteous, is to affect, by breaking, the sleep of the righteous: but, to sleep the sleep of the righteous, is not to affect by sleeping the sleep of the righteous; since the act of sleeping is an act that affects no object whatever. It is a state. We may, indeed, give it the appearance of a transitive verb, as we do when we say, the opiate slept the patient, meaning thereby, lulled to sleep; but the transitive character is only apparent.

To sleep the sleep of the righteous is to sleep in agreement with—or according to—or after the manner of—the sleep of the righteous, and the construction is adverbial.

In the grammars of the classical languages, the following rule is exceptionable—Quodvis verbum admittit accusativum nominis sibi cognati. It does so; but it governs the accusative case not objectively but modally.

§ 559. Modal verbs may be divided into a multiplicity of divisions. Of such, it is not necessary in English to give more than the following four:—

1. Appositional.—As, she walks a queen: you consider me safe. The appositional construction is, in reality, a matter of concord rather than of gender. It will be considered more fully in the following section.

2. Traditive.—As, I give the book to you=do librum tibi. I teach you the lesson=διδάσκω σὲ τὴν διδασκάλιαν. In all traditive expressions there are three ideas; (1.) an agent, (2.) an object, (3.) a person, or thing, to which the object is made over, or transferred, by the agent. For this idea the term dative is too restricted: since in Greek and some other languages, both the name of the object conveyed, and the name of the person to whom it is conveyed are, frequently, put in the accusative case.

3. Instrumental.—As, I fight with a sword=pugno ense=feohte sweorde,—Anglo-Saxon.

§ 560. Verb and nominative case.—No verb governs a nominative case. The appositional construction seems to require such a form of government; but the form is only apparent. {446}

Here, although the word is is followed by a nominative case, it by no means governs one—at least not as a verb.

It has been stated above that the so-called verb substantive is only a verb for the purposes of etymology. In syntax, it is only a part of a verb, i. e., the copula.

Now this fact changes the question of the construction in expressions like it is I, &c., from a point of government to one of concord. In the previous examples the words it, is, and I, were, respectively, subject, copula, and predicate; and, as it is the function of the copula to denote the agreement between the predicate and the subject, the real point to investigate is the nature of the concord between these two parts of a proposition.

1. It has no necessary concord in gender—she is a man in courage—he is a woman in effeminacy—it is a girl.

2. It has no necessary concord in number—sin is the wages of death—it is these that do the mischief.

4. It has, however, a necessary concord in case. Nothing but a nominative case can, by itself, constitute a term of either kind—subject or predicate. Hence, both terms must be in the nominative, and, consequently, both in the same case. Expressions like this is for me are elliptic. The logical expression is this is a thing for me.

Rule 1.—All words connected with a nominative case by the copula (i.e., the so-called verb-substantive) must be nominative.—It is I; I am safe.

Rule 2.—All words in apposition with a word so connected must be nominative.—It is difficult to illustrate this from the English language from our want of inflexions. In Latin, however, we say vocor Johannes=I am called John, not vocor Johannem. Here the logical equivalent is ego sum vocatus Johannes—where—

2. Vocatus is nominative because it is the predicate agreeing with the subject.

3. Johannes, is nominative because it is part of the predicate, and in apposition with vocatus.

N.B. Although in precise language Johannes is said to agree with vocatus rather than to be in apposition with it, the expression, as it stands, is correct. Apposition is the agreement of substantives, agreement the apposition of adjectives.

Rule 3.—All verbs which, when resolved into a copula and participle, have their participle in apposition (or agreeing) with the noun, are in the same condition as simple copulas—she walks a queen=she is walking a queen=illa est incedens regina.

Rule 4.—The construction of a subject and copula preceded by the conjunction that, is the same in respect to the predicate by which they are followed as if the sentence were an isolated proposition.

This rule determines the propriety of the expression—I believe that it is he as opposed to the expression I believe that it is him.

Now, as the relation between the subject and predicate of a proposition cannot be affected by a word which does not belong to it, the construction is the same as if the propositions were wholly separate.

N.B. The question (in cases where the conjunction that is not used), as to the greater propriety of the two expressions—I believe it to be him—I believe it to be he—has yet to be considered.

§ 561. The verb and genitive case.—No verb in the present English governs a genitive case. In Anglo-Saxon certain verbs did: e.g., verbs of ruling and others—weolde thises middangeardes=he ruled (wealded) this earth's. Genitive cases, too, governed by a verb are common both in Latin and Greek. To eat of the fruit of the tree is no genitive construction, however much it may be equivalent to one. Fruit is in the objective case, and is governed not by the verb but by the preposition of.

§ 562. The verb and accusative.—All transitive verbs govern an accusative case,—he strikes me, thee, him, her, it, us, you, them.

The verb and dative case.—The word give, and a few others, govern a dative case. Phrases like give it him, whom shall I give it, are perfectly correct, and have been explained above. The prepositional construction give it to him,—to whom shall I give it? is unnecessary. The evidence of this is the same as in the construction of the adjective like.

§ 563. The partitive construction.—Certain transitive verbs, the action whereof is extended not to the whole, but only to a part of their object, are followed by the preposition of and an objective case. To eat of the fruit of the tree=to eat a part (or some) of the fruit of the tree: to drink of the water of the well=to drink a part (or some) of the water of the well. It is not necessary, here, to suppose the ellipsis of the words part (or some). The construction is a construction that has grown out of the partitive power of the genitive case; for which case the preposition of, followed by the objective, serves as an equivalent.

§ 564. It has been already stated that forms like I believe {449}it to be him, and forms like I believe it to be he, had not been investigated. Of these, the former is, logically, correct.

With this accusative infinitive, it agrees, as being part of the same complex idea. And him does the same.

In English we have two methods of expressing one idea; the method in question, and the method by means of the conjunction, that.

In the first example, it is the object; and it-to-be-him forms one complex term.

In the second, he agrees with it; and it is the subject of a separate, though connected, proposition.

Of these two forms the Latin language adopts but one, viz., the former,—credo eum esse, not credo quod illud est ille.

§ 565. The expression ob differentiam.—The classical languages, although having but one of the two previous forms, are enabled to effect a variation in the application of it, which, although perhaps illogical, is convenient. When the speaker means himself, the noun that follows, esse, or εἶναι, is nominative,—φημὶ εἶναι δεσπότης=I say that I am the master: ait fuisse celerrimus=he says that he himself was the swiftest—but, φημὶ εἶναι δεσπότην=I say that he (some one else) is the master; and ait fuisse celerrimum=he says that he (some one else) is the swiftest. This, though not adopted in English, is capable of being adopted,—He believes it to be he (i.e., the speaker) who invented the machine; but, he believes it to be him (that is, another person) who invented it.

§ 566. When the substantive infinitive, to be, is preceded by a passive participle, combined with the verb substantive, the construction is nominative,—it is believed to be he who spoke, not it is believed to be him.—Here there are two propositions:

Now, here, it is the subject, and, as such, nominative. But it is also the equivalent to to be he, which must be nominative as well. To be he is believed=esse-ille creditur,—or, changing the mode of proof,—

3. To be he is a subordinate part of the predicate, in apposition with believed—est creditum, nempe entitas ejus. Or, to be he is believed=esse-ille est creditum.

As a general expression for the syntax of copulas and appositional constructions, the current rule, that copulas and appositional verbs must be followed by the same case by which they are preceded, stands good.

CHAPTER XVIII.

§ 567. The present participle, or the participle in -ing, must be considered in respect to its relations with the substantive in -ing. Dying-day is, probably, no more a participle than morning-walk. In respect to the syntax of such expressions as the forthcoming, I consider that they are either participles or substantives.

1. When substantives, they are in regimen, and govern a genitive case—What is the meaning of the lady's holding up her train? Here the word holding=the act of holding.—Quid est significatio elevationis pallæ de parte fœminæ.

2. When participles, they are in apposition or concord, and would, if inflected, appear in the same case with the substantive, or pronoun, preceding them—What is the meaning of the lady holding up her train? Here the word holding=in the act of holding, and answers to the Latin fœminæ elevantis.—Quid est significatio fœminæ elevantis pallam?

For the extent to which the view differs from that of Priestley, and still more with that of Mr. Guest, see Phil. Trans., 25.

§ 568. The past participle corresponds not with the Greek form τυπτόμενος, but with the form τετυμμένος. I am beaten is essentially a combination, expressive not of present but of past time, just like the Latin sum verberatus. Its Greek equivalent is not εἰμὶ τυπτόμενος=I am a man in the act of being beaten, but εἰμὶ τετυμμένος=I am a man who has been beaten. It is past in respect to the action, though present in respect to the state brought about by the action. This essentially past element in the so-called present expression, I am beaten, will be again referred to.

CHAPTER XIX.

§ 569. The infinitive mood is a noun. The current rule that when two verbs come together the latter is placed in the infinitive mood means that one verb can govern another only by converting it into a noun—I begin to move=I begin the act of moving. Verbs, as verbs, can only come together in the way of apposition—I irritate, I beat, I talk at him, I call him names, &c.

§ 570. The construction, however, of English infinitives is twofold. (1.) Objective. (2.) Gerundial.

When one verb is followed by another without the preposition to, the construction must be considered to have grown out of the objective case, or from the form in -an.

When a verb is followed by another, preceded by the preposition to, the construction must be considered to have grown out of the so-called gerund, i.e., the form in -nne, i.e., the dative case—I begin to move. This is the case with the great majority of English verbs.

The following examples, from the Old English, of the gerundial construction where we have, at present, the objective, are Mr. Guest's.

§ 571. Imperatives have three peculiarities. (1.) They can only, in English, be used in the second person: (2.) They take pronouns after, instead of before, them: (3.) They often omit the pronoun altogether.

CHAPTER XX.

§ 573. Notwithstanding its name, the present tense in English, does not express a strictly present action. It rather expresses an habitual one. He speaks well=he is a good speaker. If a man means to say that he is in the act of speaking, he says I am speaking.

It has also, especially when combined with a subjunctive mood, a future power—I beat you (=I will beat you) if you don't leave off.

§ 574. The English præterite is the equivalent, not to the Greek perfect but the Greek aorist. I beat=ἔτυψα not τέτυφα. The true perfect is expressed, in English, by the auxiliary have + the past participle.

CHAPTER XXI.

§ 576. The concord of persons.—A difficulty that occurs frequently in the Latin language is rare in English. In expressions like ego et ille followed by a verb, there arises a question as to the person in which that verb should be used. Is it to be in the first person in order to agree with ego, or in the third in order to agree with ille? For the sake of laying down a rule upon these and similar points, the classical grammarians arrange the persons (as they do the genders) according to their dignity, making the verb (or adjective if it be a question of gender) agree with the most worthy. In respect to persons, the first is more worthy than the second, and the second more worthy than the third. Hence, the Latins said—

Now, in English, the plural form is the same for all three persons. Hence we say I and you are friends, you and I are friends, I and he are friends, &c., so that, for the practice of language, the question as to the relative dignity of the three persons is a matter of indifference.

Nevertheless, it may occur even in English. Whenever two or more pronouns of different persons, and of the singular number, follow each other disjunctively, the question of concord arises. I or you,—you or he,—he or I. I believe that, in these cases, the rule is as follows:—

1. Whenever the words either or neither precede the {457}pronouns, the verb is in the third person. Either you or I is in the wrong; neither you nor I is in the wrong.

2. Whenever the disjunctive is simple (i. e. unaccompanied with the word either or neither) the verb agrees with the first of the two pronouns.

Now, provided that they are correct, it is clear that the English language knows nothing about the relative degrees of dignity between these three pronouns; since its habit is to make the verb agree with the one which is placed first—whatever may be the person. I am strongly inclined to believe that the same is the case in Latin; in which case (in the sentence ego et Balbus sustulimus manus) sustulimus agrees, in person, with ego, not because the first person is the worthiest, but because it comes first in the proposition. That the greater supposed worth of the first person may be a reason for putting it first in the proposition is likely enough.

CHAPTER XXII.

The following couplet from Dryden's "Mac Flecnoe" exhibits a construction which requires explanation:—

Here the word hight=was called, and seems to present an instance of the participle being used in a passive sense without the so-called verb substantive. Yet it does no such thing. The word is no participle at all; but a simple preterite. Certain verbs are naturally either passive or active, as one of two allied meanings may predominate. To be called is passive; so is, to be beaten. But, to bear as a name is active; so is, to take a beating. The word, hight, is of the same class of verbs with the Latin vapulo; and it is the same as the Latin word, cluo.—Barbican cluit=Barbican audivit=Barbican it hight.

CHAPTER XXIII.

§ 578. The auxiliary verbs, in English, play a most important part in the syntax of the language. They may be classified upon a variety of principles. The following, however, are all that need here be applied.

A. Classification of auxiliaries according to their inflectional or non-inflectional powers.—Inflectional auxiliaries are those that may either replace or be replaced by an inflection. Thus—I am struck=the Latin ferior, and the Greek τύπτομαι. These auxiliaries are in the same relation to verbs that prepositions are to nouns. The inflectional auxiliaries are,—

1. Have; equivalent to an inflection in the way of tense—I have bitten=mo-mordi.

4. May; equivalent to an inflection in the way of mood. I am come that I may see=venio ut vid-eam.

5. Be; equivalent to an inflection in the way of voice. To be beaten=verberari, τύπτεσθαι.

6. Am, art, is, are; ditto. Also equivalent to an inflection in the way of tense. I am moving=move-o.

7. Was, were; ditto, ditto. I was beaten=ἐ-τύφθην. I was moving=move-bam.

B. Classification of auxiliaries according to their non-auxiliary significations.—The power of the word have in the combination of I have a horse is clear enough. It means possession. The power of the same word in the combination I have been is not so clear; nevertheless it is a power which has grown out of the idea of possession. This shows that {460}the power of a verb as an auxiliary may be a modification of its original power; i. e., of the power it has in non-auxiliary constructions. Sometimes the difference is very little: the word let, in let us go, has its natural sense of permission unimpaired. Sometimes it is lost altogether. Can and may exist only as auxiliaries.

3. Auxiliary derived from the idea of future destination, dependent upon circumstances external to the agent—shall. There are etymological reasons for believing that shall is no present tense, but a perfect.

4. Auxiliary derived from the idea of future destination, dependent upon the volition of the agent—will. Shall is simply predictive; will is predictive and promissive as well.

5. Auxiliary derived from the idea of power, dependent upon circumstances external to the agent—may.

6. Auxiliary derived from the idea of power, dependent upon circumstances internal to the agent—can. May is simply permissive; can is potential. In respect to the idea of power residing in the agent being the cause which determines a contingent action, can is in the same relation to may as will is to shall.

C. Classification of auxiliary verbs in respect to their mode {461}of construction.—Auxiliary verbs combine with others in three ways.

1. With participles.—a) With the present, or active, participle—I am speaking: b) With the past, or passive, participle—I am beaten, I have beaten.

2. With infinitives.—a) With the objective infinitive—I can speak: b) With the gerundial infinitive—I have to speak.

3. With both infinitives and participles.—I shall have done, I mean to have done.

D. Auxiliary verbs may be classified according to their effect.—Thus—have makes the combination in which it appears equivalent to a tense; be to a passive form; may to a sign of mood, &c.

This sketch of the different lights under which auxiliary verbs may be viewed, has been written for the sake of illustrating, rather than exhausting, the subject.

§ 579. The following is an exhibition of some of the times in which an action may take place, as found in either the English or other languages, expressed by the use of either an inflection or a combination.

1. Present.—An action taking place at the time of speaking, and incomplete.—I am beating, I am being beaten. Not expressed, in English, by the simple present tense; since I beat means I am in the habit of beating.

2. Aorist.—An action that took place in past time, or previous to the time of speaking, and which has no connection with the time of speaking.—I struck, I was stricken. Expressed, in English, by the præterite, in Greek by the aorist. The term aorist, from the Greek ἀ-όριστος=undefined, is a convenient name for this sort of time.

3. Future.—An action that has neither taken place, nor is taking place at the time of speaking, but which is stated as one which will take place.—Expressed, in English, by the combination of will or shall with an infinitive mood. In Latin and Greek by an inflection. I shall (or will) speak, λέκ-σω, dica-m. {462}

None of these expressions imply more than a single action; in other words, they have no relation to any second action occurring simultaneously with them, before them, or after them.—I am speaking now, I spoke yesterday, I shall speak to-morrow. Of course, the act of mentioning them is not considered as an action related to them in the sense here meant.

By considering past, present, or future actions not only by themselves, but as related to other past, present, or future actions, we get fresh varieties of expression. Thus, an act may have been going on, when some other act, itself an act of past time, interrupted it. Here the action agrees with a present action, in being incomplete; but it differs from it in having been rendered incomplete by an action that has past. This is exactly the case with the—

4. Imperfect.—I was reading when he entered. Here we have two acts; the act of reading and the act of entering. Both are past as regards the time of speaking, but both are present as regards each other. This is expressed, in English, by the past tense of the verb substantive and the present participle, I was speaking; and in Latin and Greek by the imperfect tense, dicebam, ἔτυπτον.

5. Perfect.—Action past, but connected with the present by its effects or consequences.—I have written, and here is the letter. Expressed in English by the auxiliary verb have, followed by the participle passive in the accusative case and neuter gender of the singular number. The Greek expresses this by the reduplicate perfect: τέ-τυφα=I have beaten.

6. Pluperfect.—Action past, but connected with a second action, subsequent to it, which is also past.—I had written when he came in.

7. Future present.—Action future as regards the time of speaking, present as regards some future time.—I shall be speaking about this time to-morrow.

8. Future præterite.—Action future as regards the time of speaking, past as regards some future time.—I shall have spoken by this time to-morrow. {463}

These are the chief expressions which are simply determined by the relations of actions to each other, and to the time of speaking, either in the English or any other language. But over and above the simple idea of time, there may be others superadded: thus, the phrase, I do speak means, not only that I am in the habit of speaking, but that I also insist upon it being understood that I am so.

Again, an action that is mentioned as either taking place, or as having taken place at a given time, may take place again and again. Hence the idea of habit may arise out of the idea of either present time or aorist time.

Again, one tense, or one combination, may be used for another. I was speaking when he enters.

1. The emphatic present and præterite.—Expressed by do (or did), as stated above. A man says I do (or did) speak, read, &c., when, either directly or by implication, it is asserted or implied that he does not. As a question implies doubt, do is used in interrogations.

2. The predictive future.—I shall be there to-morrow. This means simply that the speaker will be present. It gives no clue to the circumstances that will determine his being so.

3. The promissive future.—I will be there to-morrow.—This means not only that the speaker will be present, but that he intends being so. For further observations on shall and will, see pp. 471-474.

4. That the power of the present tense is, in English, not present, but habitual, has already been twice stated.

§ 580. The representative expression of past and future time.—An action may be past; yet, for the sake of bringing it more vividly before the hearers, we may make it present. {464}He walks (for walked) up to him, and knocks (for knocked) him down. This denotes a single action; and is by no means the natural habitual power of the English present. So, in respect to a future, I beat you if you don't leave off, for I will beat you. This use of the present tense is sometimes called the historic use of the present tense. I find it more convenient to call it the representative use; inasmuch as it is used more after the principles of painting than of history; the former of which, necessarily, represents things as present, the latter, more naturally, describes them as past.

The use of the representative present to express simple actions is unequivocally correct. To the expression, however, of complex actions it gives an illogical character,—As I was doing this he enters (for entered). Nevertheless, such a use of the present is a fact in language, and we must take it as it occurs.

§ 581. The present tense can be used instead of the future; and that on the principle of representation. Can a future be used for a present? No.

The present tense can be used instead of the aorist; and that on the principle of representation. Can a past tense, or combination, be used for a present?

In respect to the perfect tense there is no doubt. The answer is in the affirmative. For all purposes of syntax a perfect tense, or a combination equivalent to one, is a present tense. Contrast the expression, I come that I may see; with the expression, I came that I might see; i.e., the present construction with the aorist. Then, bring in the perfect construction, I have come. It differs with the aorist, and agrees with the present. I have come that I may see. The reason for this is clear. There is not only a present element in all perfects, but for the purposes of syntax, the present element predominates. Hence expressions like I shall go, need give us no trouble; even though shall be considered as a perfect tense. Suppose the root, sk-ll to mean to be destined (or fated). Provided we consider the effects of the action to be continued up to the time of speaking, we may say I have been destined to go, just as well as we can say I am destined to go. {465}

The use of the aorist as a present (except so far as both the tenses agree in their power of expressing habitual actions) is a more difficult investigation. It bears upon such expressions as I ought to go, &c., and will be taken up in p. 475.

§ 582. Certain adverbs, i.e., those of time, require certain tenses. I am then, I was now, I was hereafter, &c., are contradictory expressions. They are not so much bad grammar as impossible nonsense. Nevertheless, we have in Latin such expressions as

Here the connection of the present and perfect ideas explains the apparent contradiction. The present state may be the result of a previous one; so that a preterite element may be involved in a present expression. Ut sumus=since I have been where I am.

It is hardly necessary to remark that such expressions as since I am here (where since=inasmuch as) do not come under this class.

§ 583. Two fresh varieties in the use of tenses and auxiliary verbs may be arrived at by considering the following ideas, which may be superadded to that of simple time.

1. Continuance in the case of future actions.—A future action may not only take place, but continue: thus, a man may, on a given day, not only be called by a particular name, but may keep that name. When Hesiod says that, notwithstanding certain changes which shall have taken place, good shall continue to be mixed with bad, he does not say, ἐσθλὰ μιχθήσεται κακοῖσιν, but,

Here μετεγγραφήσεται means change from one class to another, ἠγγεγράψεται continuance in the same.—See Mathiæ, ii. § 498.

Seidler remarks that κληθήσεται, est nomen accipiet; κεκλήσεται, nomen geret.

Now it is quite true that this Greek tense, the so-called paulo-post-futurum, "bears the same relation to the other futures as, among the tenses of past time, the perfectum does to the aorist."—(Mathiæ.) And it is also true that it by no means answers to the English shall have been. Yet the logical elements of both are the same. In the English expression, the past power of the perfect predominates, in the Greek its present power.

2. Habit in the case of past actions.—I had dined when I rode out. This may apply to a particular dinner, followed by a particular ride. But it may also mean that when the speaker had dined, according to habit, he rode out, according to habit also. This gives us a variety of pluperfect; which is, in the French language, represented by separate combination—j'avais diné, j'eus diné.

§ 584. It is necessary to remember that the connection between the present and the past time, which is involved in the idea of a perfect tense (τέτυφα), or perfect combination (I have beaten), is of several sorts.

It may consist in the present proof of the past fact,—I have written, and here is the evidence.

It may consist in the present effects of the past fact,—I have written, and here is the answer.

Without either enumerating or classifying these different kinds of connexion, it is necessary to indicate two sorts of inference to which they may give origin.

1. The inference of continuance.—When a person says, I have learned my lesson, we presume that he can say it, i. e., that, he has a present knowledge of it. Upon this principle {467}κέκτημαι=I have earned=I possess. The past action is assumed to be continued in its effects.

2. The inference of contrast.—When a person says, I have been young, we presume that he is so no longer. The action is past, but it is continued up to the time of speaking by the contrast which it supplies. Upon this principle, fuit Ilium means Ilium is no more.

In speaking, this difference can be expressed by a difference of accent. I have learned my lesson, implies that I don't mean to learn it again. I have learned my lesson, implies that I can say it.

§ 585. The construction of the auxiliary, may, will be considered in the Chapter on Conjunctions; that of can, must, and let, offer nothing remarkable. The combination of the auxiliary, have, with the past participle requires notice. It is, here, advisable to make the following classifications.

1. The combination with the participle of a transitive verb.—I have ridden the horse; thou hast broken the sword; he has smitten the enemy.

2. The combination with the participle of an intransitive verb,—I have waited; thou hast hungered; he has slept.

3. The combination with the participle of the verb substantive,—I have been; thou hast been; he has been.

It is by examples of the first of these three divisions that the true construction is to be shown.

For an object of any sort to be in the possession of a person, it must previously have existed. If I possess a horse, that horse must have had a previous existence.

Hence, in all expressions like I have ridden a horse, there are two ideas, a past idea in the participle, and a present idea in the word denoting possession.

For an object of any sort, affected in a particular manner, to be in the possession of a person, it must previously have been affected in the manner required. If I possess a horse that has been ridden, the riding must have taken place before I mention the fact of the ridden horse being in my possession; inasmuch as I speak of it as a thing already done,—the participle, ridden, being in the past tense. {468}

I have ridden a horse=I have a horse ridden=I have a horse as a ridden horse, or (changing the gender and dealing with the word horse as a thing)=I have a horse as a ridden thing.

In this case the syntax is of the usual sort. (1) Have=own=habeo=teneo; (2) horse is the accusative case=equum; (3) ridden is a past participle agreeing either with horse, or with a word in apposition with it understood.

Mark the words in italics. The word ridden does not agree with horse, since it is of the neuter gender. Neither if we said I have ridden the horses, would it agree with horses; since it is of the singular number.

The true construction is arrived at by supplying the word thing. I have a horse as a ridden thing=habeo equum equitatum (neuter). Here the construction is the same as triste lupus stabulis.

I have horses as a ridden thing=habeo equos equitatam (singular, neuter). Here the construction is—

The classical writers supply instances of this use of have. Compertum habeo, milites, verba viris virtutem non addere=I have discovered=I am in possession of the discovery. Quæ cum ita sint, satis de Cæsare hoc dictum habeo.

2. The combination of have with an intransitive verb is irreducible to the idea of possession: indeed, it is illogical. In I have waited, we cannot make the idea expressed by the word waited the object of the verb have or possess. The expression has become a part of language by means of the extension of a false analogy. It is an instance of an illegitimate imitation.

3. The combination of have with been is more illogical still, and is a stronger instance of the influence of an illegitimate imitation. In German and Italian, where even intransitive verbs are combined with the equivalents to the English have {469}(haben and avere), the verb substantive is not so combined; on the contrary, the combinations are

1. The idea of direction towards an object conveyed by the dative case, and by combinations equivalent to it.

2. The extent to which the ideas of necessity, obligation, or intention are connected with the idea of something that has to be done, or something towards which some action has a tendency.

3. The fact that expressions like the one in question historically represent an original dative case, or its equivalent; since to speak grows out of the Anglo-Saxon form to sprecanne, which, although called a gerund, is really a dative case of the infinitive mood.

When Johnson (see Mr. Guest, Phil. Trans. No. 44) thought that, in the phrase he is to blame, the word blame was a noun, if he meant a noun in the way that culpa is a noun, his view was wrong. But if he meant a noun in the way that culpare, ad culpandum, are nouns, it was right.

§ 587. I am to blame.—This idiom is one degree more complex than the previous one; since I am to blame=I am to be blamed. As early, however, as the Anglo-Saxon period the gerunds were liable to be used in a passive sense: he is to lufigenne=not he is to love, but he is to be loved.

The principle of this confusion may be discovered by considering that an object to be blamed, is an object for some one to blame, an object to be loved is an object for some one to love.

§ 588. Shall and will.—The simply predictive future verb is shall. Nevertheless, it is only used in the first person. The second and third persons are expressed by the promissive verb will.

The promissive future verb is will. Nevertheless, it is only used in the first person. The second and third persons are expressed by the predictive verb shall. {470}

"In primis personis shall simpliciter prædicentis est; will, quasi promittentis aut minantis.

"In secundis et tertiis personis, shall promittentis est aut minantis: will simpliciter prædicentis.

Again—"would et should illud indicant quod erat vel esset futurum: cum hoc tantum discrimine: would voluntatem innuit, seu agentis propensionem: should simpliciter futuritionem."—Wallis, p. 107.

§ 589. Archdeacon Hare explains this by a usus ethicus. "In fact, this was one of the artifices to which the genius of the Greek language had recourse, to avoid speaking presumptuously of the future: for there is an awful, irrepressible, and almost instinctive consciousness of the uncertainty of the future, and of our own powerlessness over it, which, in all cultivated languages, has silently and imperceptibly modified the modes of expression with regard to it: and from a double kind of litotes, the one belonging to human nature generally, the other imposed by good-breeding on the individual, and urging him to veil the manifestations of his will, we are induced to frame all sorts of shifts for the sake of speaking with becoming modesty. Another method, as we know, frequently adopted by the Greeks was the use of the conditional moods: and as sentiments of this kind always imply some degree of intellectual refinement, and strengthen with its increase, this is called an Attic usage. The same name too has often been given to the above-mentioned middle forms of the future; not that in either case the practice was peculiar to the Attic dialect, but that it was more general where the feelings which produced it were {471}strong and more distinct. Here again our own language supplies us with an exact parallel: indeed this is the only way of accounting for the singular mixture of the two verbs shall and will, by which, as we have no auxiliary answering to the German werde, we express the future tense. Our future, or at least what answers to it, is, I shall, thou wilt, he will. When speaking in the first person, we speak submissively: when speaking to or of another, we speak courteously. In our older writers, for instance in our translation of the Bible, shall is applied to all three persons: we had not then reacht that stage of politeness which shrinks from the appearance even of speaking compulsorily of another. On the other hand the Scotch use will in the first person: that is, as a nation they have not acquired that particular shade of good-breeding which shrinks from thrusting itself^[61] forward."

§ 590. Notice of the use of will and shall, by Professor De Morgan.—"The matter to be explained is the synonymous character of will in the first person with shall in the second and third; and of shall in the first person with will in the second and third: shall (1) and will (2, 3) are called predictive: shall (2, 3) and will (1) promissive. The suggestion now proposed will require four distinctive names.

"Archdeacon Hare's usus ethicus is taken from the brighter side of human nature:—'When speaking in the first person we speak submissively; when speaking to or of another, we speak courteously.' This explains I shall, thou wilt; but I cannot think it explains I will, thou shalt. It often happens {473}that you will, with a persuasive tone, is used courteously for something next to, if not quite, you shall. The present explanation is taken from the darker side; and it is to be feared that the à priori probabilities are in its favour.

"In introducing the common mode of stating the future tenses, grammar has proceeded as if she were more than a formal science. She has no more business to collect together I shall, thou wilt, he will, than to do the same with I rule, thou art ruled, he is ruled.

"It seems to be the natural disposition of man to think of his own volition in two of the following catagories, and of another man's in the other two:

"The ego, with reference to the non-ego, is apt, thinking of himself, to propound the alternative, 'Shall I compel, or shall I leave him to do as he likes?' so that, thinking of the other, the alternative is, 'shall he be restrained, or shall he be left to his own will?' Accordingly, the express introduction of his own will is likely to have reference to compulsion, in case of opposition: the express introduction of the will of another, is likely to mean no more than the gracious permission of the ego to let non-ego do as he likes. Correlatively, the suppression of reference to his own will, and the adoption of a simply predictive form on the part of the ego, is likely to be the mode with which, when the person is changed, he will associate the idea of another having his own way; while the suppression of reference to the will of the non-ego is likely to infer restraint produced by the predominant will of the ego.

"Occasionally, the will of the non-ego is referred to as under restraint in modern times. To I will not, the answer is sometimes you shall, meaning, in spite of the will—sometimes you will, meaning that the will will be changed by fear or sense of the inutility of resistance."^[62]

§ 591. I am beaten.—This is a present combination, and it is present on the strength of the verb am, not on the strength of the participle beaten, which is præterite.

The following table exhibits the expedients on the part of the different languages of the Gothic stock, since the loss of the proper passive form of the Mœso-Gothic.

§ 592. Ought, would, &c., used as presents.—These words are not in the predicament of shall.

But they are not, like shall, perfect forms; i. e., they have no natural present element in them.

So had their equivalents in Greek: ἐχρῆν=χρὴ, ἔδει=δεῖ, προσῆκεν=προσήκει.

In Latin, too, would was often not represented by either volo or volebam, but by velim.

The assertion of duty or obligation is one of those assertions which men like to soften in the expression: should, ought.

Very often when we say you should (or ought to) do this, we leave to be added by implication—but you do not.

Very often when we say I could (or might) do this, we leave to be added by implication—but I do not exert my power.

Now, if what is left undone be the present element in this assertion, the duty to do it, or the power of doing it, constitutes a past element in it; since the power (or duty) is, in relation to the performance, a cause—insufficient, indeed, but still antecedent. This hypothesis is suggested rather than asserted.

§ 593. By substituting the words I am bound for I ought, {476}we may see the expedients to which this present use of the præterite forces us.

I am bound to do this now = I owe to do this now. However, we do not say owe, but ought.

Hence, when we wish to say I was bound to do this two years ago, we cannot say I ought (owed) to do this, &c., since ought is already used in a present sense.

We therefore say, instead, I ought to have done this two years ago; which has a similar, but by no means an identical meaning.

I was bound to pay two years ago, means two years ago I was under an obligation to make a payment, either then or at some future time.

I was bound to have paid, &c., means I was under an obligation to have made a payment.

Common people sometimes say, you had not ought to do so and so; and they have a reason for saying it.

The Latin language is more logical. It says not debet factum fuisse, but debuit fieri.

CHAPTER XXIV.

§ 594. The syntax of the adverb is simpler than that of any other part of speech, excepting, perhaps, that of the adjective.

Neither have they any government. They seem, indeed, to have it, when they are in the comparative or superlative degree; but it is merely apparent. In this is better than that, the word that is governed neither by better nor by than. It is not governed at all. It is a nominative case; the subject of a separate proposition. This is better (i. e., more good) than that is good. Even if we admit such an expression as he is stronger than me to be good English, there is no adverbial government. Than, if it govern me at all, governs it as a preposition.

The position of an adverb is, in respect to matters of syntax, pre-eminently parenthetic; i. e., it may be omitted without injuring the construction. He is fighting—now; he was fighting—then; he fights—bravely; I am—almost—tired, &c.

§ 595. By referring to the Chapter on the Adverbs, we shall find that the neuter adjective is frequently converted into an adverb by deflection. As any neuter adjective may be so deflected, we may justify such expressions as full (for fully) as conspicuous, and peculiar (for peculiarly) bad grace, &c. We are not, however, bound to imitate everything that we can justify.

§ 596. The termination -ly was originally adjectival. At present it is a derivational syllable by which we can convert an adjective into an adverb: brave, brave-ly. {478}When, however, the adjective ends in -ly already, the formation is awkward. I eat my daily bread is unexceptionable English; I eat my bread daily is exceptionable. One of two things must here take place: the two syllables -ly are packed into one (the full expression being dai-li-ly), or else the construction is that of a neuter adjective deflected.

Adverbs are convertible. The then men=οἱ νῦν βρότοι, &c. This will be seen more clearly in the Chapter on Conjunctions.

§ 597. It has been remarked that in expressions like he sleeps the sleep of the righteous, the construction is adverbial. So it is in expressions like he walked a mile, it weighs a pound. The ideas expressed by mile and pound are not the names of anything that serves as either object or instrument to the verb. They only denote the manner of the action, and define the meaning of the verb.

§ 598. From whence, from thence.—This is an expression which, if it have not taken root in our language, is likely to do so. It is an instance of excess of expression in the way of syntax; the -ce denoting direction from a place, and the preposition doing the same. It is not so important to determine what this construction is, as to suggest what it is not. It is not an instance of an adverb governed by a preposition. If the two words be dealt with as logically separate, whence (or thence) must be a noun=which place (or that place); just as from then till now=from that time to this. But if (which is the better view) the two words be dealt with as one (i. e., as an improper compound) the preposition from has lost its natural power, and become the element of an adverb.

CHAPTER XXV.

§ 599. All prepositions govern an oblique case. If a word cease to do this, it ceases to be a preposition. In the first of the two following sentences the word up is a preposition, in the second an adverb.

All prepositions in English precede the noun which they govern. I climbed up the tree, never I climbed the tree up. This is a matter not of government, but of collocation. It is the case in most languages; and, from the frequency of its occurrence, the term pre-position (or prefix) has originated. Nevertheless, it is by no means a philological necessity. In many languages the prepositions are post-positive, following their noun.

§ 600. No preposition, in the present English, governs a genitive case. This remark is made, because expressions like the part of the body=pars corporis,—a piece of the bread=portio panis, make it appear as if the preposition of did so. The true expression is, that the preposition of followed by an objective case, is equivalent, in many instances, to the genitive case of the classical languages.

§ 601. The writer, however, of a paper on English preterites and genitives, in the Philological Museum (II. 261) objects to the current doctrine concerning such constructions as, this is a picture of the king's. Instead of considering the sentence elliptic, and equivalent to this is a picture of or (from) the king's pictures, he entertains the following view,—"I confess, however, that I feel some doubt whether this phrase is {480}indeed to be regarded as elliptical, that is, whether the phrase in room of which it is said to stand, was ever actually in use. It has sometimes struck me that this may be a relict of the old practice of using the genitive after nouns as well as before them, only with the insertion of the preposition of. One of the passages quoted above from 'Arnold's Chronicle,' supplies an instance of a genitive so situated; and one cannot help thinking that it was the notion that of governed the genitive, that led the old translators of Virgil to call his poem The Booke of Eneidos, as it is termed by Phaer, and Gawin Douglas, and in the translation printed by Caxton. Hence it may be that we put the genitive after the noun in such cases, in order to express those relations which are most appropriately expressed by the genitive preceding it. A picture of the king's is something very different from the king's picture: and so many other relations are designated by of with the objective noun, that if we wish to denote possession thereby, it leaves an ambiguity: so, for this purpose, when we want to subjoin the name of the possessor to the thing possest, we have recourse to the genitive, by prefixing which we are wont to express the same idea. At all events as, if we were askt whose castle Alnwick is, we should answer, The Duke of Northumberland's; so we should also say, What a grand castle that is of the Duke of Northumberland's! without at all taking into account whether he had other castles besides: and our expression would be equally appropriate, whether he had or not."

Ours does not necessarily mean of us. It may also mean of our hills, i. e., of the hills of our choice. Nightes may mean of the night's hours. In the expression, what a grand castle, &c., it is submitted to the reader that we do take into our account other castles, which the Duke of Northumberland {481}may or may not have. The Booke of Eneidos is a mistaken Latinism. As it does not seem to have been sufficiently considered that the real case governed by of (as by de in Latin) is the ablative, it is the opinion of the present writer that no instance has yet been produced of of either governing, or having governed a genitive case.

§ 602. It is not so safe to say in the present English that no preposition governs a dative. The expression give it him is good English; and it is also equivalent to the Latin da ei. But we may also say give it to him. Now the German zu=to governs a dative case, and in Anglo-Saxon, the preposition to, when prefixed to the infinitive mood, required the case that followed it to be a dative.

§ 603. When the infinitive mood is used as the subject of a proposition, i.e., as a nominative case, it is impossible to allow to the preposition to, by which it is preceded, any separate existence whatever,—to rise=rising; to err=error. Here the preposition must, for the purposes of syntax, be considered as incorporated with the noun, just like an inseparable inflection. As such it may be preceded by another preposition. The following example, although a Grecism, illustrates this:—

§ 604. Akin to this, but not the same, is the so-called vulgarism, consisting of the use of the preposition for. I am ready to go=I am ready for going=the so-called vulgarism, I am ready for to go. Now, this expression differs from the last in exhibiting, not only a verbal accumulation of prepositions, but a logical accumulation as well: inasmuch as for and to express like ideas.

§ 605. Composition converts prepositions into adverbs. Whether we say upstanding or standing-up, we express the manner in which an action takes place, and not the relation between two substantives. The so-called prepositional compounds in Greek (ἀναβαίνω, ἀποθνήσκω, &c.) are all adverbial.

CHAPTER XXVI.

§ 606. A CONJUNCTION is a part of speech which connects propositions,—the day is bright, is one proposition. The sun shines, is another. The day is bright because the sun shines is a pair of propositions connected by the conjunction, because.

From this it follows, that whenever there is a conjunction, there are two subjects, two copulas, and two predicates: i.e., two propositions in all their parts.

But this may be expressed compendiously. The sun shines, and the moon shines, may be expressed by the sun and moon shine.

Nevertheless, however compendious may be the expression, there are always two propositions wherever there is one conjunction. A part of speech that merely combines two words is a preposition—the sun along with the moon shines.

It is also highly important to remember that many double propositions may be expressed so compendiously as to look like one. When this takes place, and any question arises as to the construction, they must be exhibited in their fully expanded form; i.e., the second subject, the second predicate, and the second copula must be supplied. This can always be done from the first proposition,—he likes you better than me=he likes you better than he likes me. The compendious expression of the second proposition is the first point of note in the syntax of conjunctions.

§ 607. The second point in the syntax of conjunctions is the fact of their great convertibility. Most conjunctions have been developed out of some other part of speech. {483}

The conjunction of comparison, than, is derived from the adverb of time, then; which is derived from the accusative singular of the demonstrative pronoun.

One and the same word, in one and the same sentence, may be a conjunction or preposition, as the case may be.

All fled but John.—If this mean all fled except John, the word but is a preposition, the word John is an accusative case, and the proposition is single. If, instead of John, we had a personal pronoun, we should say all fled but him.

All fled but John.—If this mean all fled, but John did not fly, the word but is a conjunction, the word John is a nominative case, and the propositions are two in number. If, instead of John, we had a personal pronoun, we should say, all fled but he.

From the fact of the great convertibility of conjunctions it is often necessary to determine whether a word be a conjunction or not. If it be a conjunction, it cannot govern a case. If it govern a case, it is no conjunction but a preposition. A conjunction cannot govern a case, for the following reason,—the word that follows it must be the subject of the second proposition, and, as such, a nominative case.

§ 608. The third point to determine in the syntax of conjunctions is the certainty or uncertainty in the mind of the speaker as to the facts expressed by the propositions which they serve to connect.

1. Each proposition may contain a certain, definite, absolute fact—the day is clear because the sun shines. Here, there is neither doubt nor contingency of either the day being clear, or of the sun shining.

2. Of two propositions one may be the condition of the other—the day will be clear if the sun shine. Here, although it is certain that if the sun shine the day will be clear, there is {484}no certainty of the sun shining. Of the two propositions one only embodies a certain fact, and that is certain only conditionally.

Now an action, wherein there enters any notion of uncertainty, or indefinitude, and is at the same time connected with another action, is expressed, not by the indicative mood, but by the subjunctive. If the sun shine (not shines) the day will be clear.

Simple uncertainty will not constitute a subjunctive construction,—I am, perhaps, in the wrong.

But, the two combined constitute the construction in question,—if I be wrong, you are right.

Now, a conjunction that connects two certain propositions may be said to govern an indicative mood.

And a conjunction that connects an uncertain proposition with a certain one, may be said to govern a subjunctive mood.

The government of mood is the only form of government of which conjunctions are capable.

§ 609. Previous to the question of the government of conjunctions in the way of mood, it is necessary to notice certain points of agreement between them and the relative pronouns; inasmuch as, in many cases, the relative pronoun exerts the same government, in the way of determining the mood of the verb, as the conjunction.

Between the relative pronouns and conjunctions in general there is this point of connection,—both join propositions. Wherever there is a relative, there is a second proposition. So there is wherever there is a conjunction.

Between certain relative pronouns and those particular conjunctions that govern a subjunctive mood there is also a point of connection. Both suggest an element of uncertainty or indefinitude. This the relative pronouns do, through the logical elements common to them and to the interrogatives: these latter essentially suggesting the idea of doubt. Wherever the person, or thing, connected with an action, and expressed by a relative be indefinite, there is room for the use {485}a subjunctive mood. Thus—he that troubled you shall bear his judgment, whosoever he be.

§ 610. By considering the nature of such words as when, their origin as relatives on the one hand, and their conjunctional character on the other hand, we are prepared for finding a relative element in words like till, until, before, as long as, &c. These can all be expanded into expressions like until the time when, during the time when, &c. Hence, in an expression like seek out his wickedness till thou find (not findest) none, the principle of the construction is nearly the same as in he that troubled you, &c., or vice versâ.^[63]

§ 611. In most conditional expressions the subjunctive mood should follow the conjunction. All the following expressions are conditional.

Expressions like except and unless are equally conditional with words like if and provided that, since they are equivalent to if—not.

Expressions like though and although are peculiar. They join propositions, of which the one is a primâ facie reason against the existence of the other: and this is the conditional element. In the sentence, if the children be so badly brought up, they are not to be trusted, the bad bringing-up is the reason {486}for their being unfit to be trusted; and, as far as the expression is concerned, is admitted to be so. The only uncertainty lies in the question as to the degree of the badness of the education. The inference from it is unequivocal.

But if, instead of saying if, we say although, and omit the word not, so that the sentence run although the children be so badly brought up they are to be trusted, we do two things: we indicate the general relation of cause and effect that exists between bad bringing-up and unfitness for being trusted, but we also, at the same time, take an exception to it in the particular instance before us. These remarks have been made for the sake of showing the extent to which words like though, &c., are conditional.

It must be remembered, however, that conjunctions, like the ones lately quoted, do not govern subjunctive moods because they are conditional, but because, in the particular condition which they accompany, there is an element of uncertainty.

§ 612. This introduces a fresh question. Conditional conjunctions are of two sorts:—

1. Those which express a condition as an actual fact, and one admitted as such by the speaker.

2. Those which express a condition as a possible fact, and one which the speaker either does not admit, or admits only in a qualified manner.

Since the children are so badly brought up, &c.—This is an instance of the first construction. The speaker admits as an actual fact the bad bringing-up of the children.

If the children be so badly brought-up, &c.—This is an instance of the second construction. The speaker admits as a possible (perhaps, as a probable) fact the bad bringing-up of the children: but he does not adopt it as an indubitable one.

§ 613. Now, if every conjunction had a fixed unvariable meaning, there would be no difficulty in determining whether a condition was absolute, and beyond doubt, or possible, and liable to doubt. But such is not the case.

Although may precede a proposition which is admitted as well as one which is doubted. {487}

If, too, may precede propositions wherein there is no doubt whatever implied: in other words it may be used instead of since.

In some languages this interchange goes farther than in others; in the Greek, for instance, such is the case with εἰ, to a very great extent indeed.

Hence we must look to the meaning of the sentence in general, rather than to the particular conjunction used.

It is a philological fact (probably referable to the usus ethicus) that if may stand instead of since.

It is also a philological fact that when it does so it should be followed by the indicative mood.

This is written in the way of illustration. What applies to if applies to other conjunctions as well.

§ 614. As a point of practice, the following method of determining the amount of doubt expressed in a conditional proposition is useful:—

Insert, immediately after the conjunction, one of the two following phrases,—(1.) as is the case; (2.) as may or may not be the case. By ascertaining which of these two supplements expresses the meaning of the speaker, we ascertain the mood of the verb which follows.

When the first formula is one required, there is no element of doubt, and the verb should be in the indicative mood. If (as is the case), he is gone, I must follow him.

When the second formula is the one required, there is an element of doubt, and the verb should be in the subjunctive mood. If (as may or may not be the case) he be gone, I must follow him.

§ 615. The use of the word that in expressions like I eat that I may live, &c., is a modification of the subjunctive construction, that is conveniently called potential. It denotes that one act is done for the sake of supplying the power or opportunity for the performance of another.

In English the word that, so used, cannot be said to govern a mood, although generally followed by either may or might. {488}It should rather be said to require a certain combination to follow it. The most important point connected with the powers of that is the so-called succession of tenses.

§ 616. The succession of tenses.—Whenever the conjunction that expresses intention, and consequently connects two verbs, the second of which takes place after the first, the verbs in question must be in the same tense.

In the Greek language this is expressed by a difference of mood; the subjunctive being the construction equivalent to may, the optative to might. The Latin idiom coincides with the English.

A little consideration will show that this rule is absolute. For a man to be doing one action (in present time) in order that some other action may follow it (in past time) is to reverse the order of cause and effect. To do anything in A.D. 1851, that something may result from it in 1850 is a contradiction; and so it is to say I do this that I might gain by it.

The reasons against the converse construction are nearly, if not equally cogent. To have done anything at any previous time in order that a present effect may follow, is, ipso facto, to convert a past act into a present one, or, to speak in the language of the grammarian, to convert an aorist into a perfect. To say I did this that I may gain by it, is to make, by the very effect of the expression, either may equivalent to might, or did equivalent to have done.

A clear perception of the logical necessity of the law of the succession of tenses, is necessary for understanding the nature of several anomalous passages in the classical writers. In the following, an aorist is followed not by an optative, but by a subjunctive.

Here it is necessary to construe ἔδωκε, has given and continues to allow, which is to construe it like a perfect^[64] tense. Upon similar passages Mathiæ writes, "but frequently the conjunctive is used, although the preceding word be in the time past, viz., when the verb which depends upon the conjunction shows an action continued to the present time." That means when the verb is really a perfect.

In Latin, where the same form is both aorist and perfect, the succession of tenses is a means of determining which of the two meanings it conveys. Veni ut videam=I have come that I may see. Veni ut viderem=I came that I might see.

Arnold states, from Krüger and Zumpt, that even where the præterite was clearly a perfect (i. e., =to have with the participle), the Roman ear was so accustomed to the imperfect subjunctive, that it preferred such an expression as diu dubitavi num melius esset to diu dubitavi num melius sit. The latter part of the statement is sure enough; but it is by no means so sure that dubitavi, and similar forms in similar constructions are perfects. There is no reason for considering this to be the case in the present instance. It seems to be so, because it is connected with diu; but an action may last a long time, and yet not last up to the time of speaking. Diu dubitavi probably expresses, I doubted a long time, and leaves it to be inferred that now I do not doubt.

§ 617. It has been stated above that whilst the Latin and English have a succession of tenses, the Greek language {490}exhibits what may be called a succession of moods. This suggests inquiry. Is the difference real? If so, how is it explained? If not, which of the two grammatical systems is right?—the English and Latin on the one side, or the Greek on the other? Should τύπτοιμι be reduced to a past tense, or verberarem be considered an optative mood.

The present writer has no hesitation in stating his belief, that all the phænomena explicable by the assumption of an optative mood are equally explicable by an expansion of the subjunctive, and a different distribution of its tenses.

1. Let τύψω be considered a subjunctive future instead of a subjunctive aorist.

1. The double forms τύψαιμι and τύψοιμι, one of which would remain unplaced.

The first reason I am not prepared to impugn. Valeat quantum, &c. The second indicates a class of expressions which tense will not explain, and which mood will. Yet this is not conclusive. Would that thou wert is thoroughly optative: yet it is expressed by a tense.

The form of the so-called optatives proves nothing. Neither the subjunctive nor the optative has any signs of mood at all, except the negative one of the absence of the augment. Their signs are the signs of tense.

1. The analogy of other languages. The imperfect has a subjunctive in Latin. So has the future.

2. The undoubtedly future character of the so-called aorist imperative. To give an order to do a thing in past time is a philological contradiction. Forms like βλέψον must be future. Though θὲς and τίθει differ in power, they both mean an {491}action subsequent to, or, at any rate, simultaneous with the order given; certainly not one anterior to it.

§ 618. Be may stand for may be. In this case the preterite is not were but might be. The sentence, what care I how fair the lady be, if she be not fair to her admirer? is accurate. Here be = may be. But, what cared I how fair the lady were, if she were not fair to her admirer? is inaccurate. It ought to run thus,—what cared I how fair the lady might be, if she were not fair to her admirer?^[65]

§ 619. Disjunctives.—Disjunctives (or, nor) are of two sorts, real, and nominal.

A king or queen always rules in England. Here the disjunction is real; king or queen being different names for different objects. In all real disjunctions the inference is, that if one out of two (or more) individuals (or classes) do not perform a certain action, the other does.

A sovereign or supreme ruler always rules in England. Here the disjunction is nominal; sovereign and supreme governor being different names for the same object. In all nominal disjunctives the inference is, that if an agent (or agents) do not perform a certain action under one name, he does (or they do) it under another.

In the English language there is no separate word to distinguish the nominal from the real disjunctive. In Latin, {492}vel is considered by Harris to be disjunctive, sive subdisjunctive. As a periphrasis the combination in other words is subdisjunctive.

Both nominal and real disjunctives agree in this,—whatever may be the number of nouns which they connect, the construction of the verb is the same as if there were but one—Henry or John, or Thomas, walks (not walk); the sun, or solar luminary, shines (not shine). The disjunctive isolates the subject however much it may be placed in juxtaposition with other nouns.

§ 620. Either, neither.—Many disjunctives imply an alternative. If it be not this person (or thing) that performs a certain action (or exists in a certain state) it is some other. If a person (or thing) do not perform a certain action (or exist in a certain state), under one name, he (or it) does so under another. This alternative is expressed by the word either.

When the word either is connected immediately with the copula of a proposition, it is, if not a true conjunction, at least a part of a conjunctional periphrasis.—This either is or is not so.

When it belongs more to one of the terms of a proposition than to the copula, it is a pronoun,—Either I or you is in the wrong. It is either you or I.

I use the words, part of a conjunctional periphrasis, because the full conjunction is either + or (or neither + nor); the essential conjunctions being the latter words. To these, either (or neither) is superadded, indicating the manner in which the disjunction expressed by or (or nor) takes place; i. e., they show that it takes place in the manner of an alternative. Now, this superadded power is rather adverbial than conjunctional.

§ 621. From the pronominal character of the word either, when it forms part of a term, and from the power of the disjunctive, or, in isolating the subject of the verb, combined with an assumption which will be explained hereafter, we get at the principle of certain rules for doubtful constructions.

In expressions like either you or I is in the wrong, we must {493}consider either not only as a pronoun, but as the leading pronoun of the proposition; a pronoun of which or I is an explanation; and, finally, as the pronoun which determines the person of the verb. Either you or I is wrong=one of us (you or I) is wrong.

Then, as to expressions like I, or you, am in the wrong. Here, I is the leading pronoun, which determines the person of the verbs; the words, or you, being parenthetic, and subordinate. These statements bear upon the rules of p. 457.

§ 622. Will this principle justify such expressions as either they or we is in the wrong?

Or will it justify such expressions as I or they am in the wrong? In all which sentences one pronoun is plural.

Perhaps not. The assumption that has been just alluded to, as helping to explain certain doubtful constructions, is the following, viz., that in cases of apposition, disjunction, and complex terms, the first word is the one which determines the character of the sentence wherein it occurs. This is a practice of the English language, which, in the opinion of the present writer, nothing but a very decided preponderance of a difference in person, gender, or number, can overrule. Such may fairly be considered to be the case in the three examples just adduced; especially as there is also the secondary influence of the conjunctional character of the word either. Thus, although we say,—

As for the other two expressions, they are in the same predicament, with an additional reason for the use of the plural. It contains the singular. The chief object of the present remarks has been less to explain details than to give due prominence to the following leading principles.

4. That when it is the first word of the proposition it determines the concord of the verb, unless its character of a noun of the singular number and third person be disguised by the prominence of some plural form, or some pronoun of the first or second person in the latter part of the term.

5. That in a simple disjunctive proposition (i.e., one where either does not occur) all nouns are subordinate to the first.

§ 623. I believe that the use of either is limited to real disjunctives; in other words, that we can say either a king or a queen always reigns in England, but that we cannot say either a sovereign or a supreme ruler always reigns in England.

CHAPTER XXVII.

§ 624. When the verb is in the infinitive mood, the negative precedes it.—Not to advance is to retreat.

When the verb is not in the infinitive mood, the negative follows it.—He advanced not. I cannot.

This rule is absolute. It only seems to precede the verb in such expressions as I do not advance, I cannot advance, I have not advanced, &c. However, the words do, can, and have, are no infinitives; and it consequently follows them. The word advance is an infinitive, and it consequently precedes it. Wallis's rule makes an equivalent statement, although differently. "Adverbium negandi not (non) verbo postponitur (nempe auxiliari primo si adsit; aut si non adsit auxiliare, verbo principali): aliis tamen orationis partibus præfigi solet."—P. 113.

That the negative is rarely used, except with an auxiliary, in other words, that the presence of a negative converts a simple form like it burneth not into the circumlocution it does not burn, is a fact in the practice of the English language. The syntax is the same in either expression.

§ 625. What may be called the distribution of the negative is pretty regular in English. Thus, when the word not comes between an indicative, imperative, or subjunctive mood and an infinitive verb, it almost always is taken with the word which it follows—I can not eat may mean either I can—not eat (i.e., I can abstain), or I can not—eat (i.e., I am unable to eat); but, as stated above, it almost always has the latter signification.

But not always. In Byron's "Deformed Transformed" we find the following lines:— {496}

Here not to refuse=to accept; and is probably a Grecism. To not refuse would, perhaps, be better.

§ 626. In the present English, two negatives make an affirmative. I have not not seen him=I have seen him. In Greek this was not the case. Duæ aut plures negativæ apud Græcos vehementius negant is a well-known rule. The Anglo-Saxon idiom differed from the English and coincided with the Greek. The French negative is only apparently double; words like point, pas, mean not not, but at all. Je ne parle pas = I not speak at all, not I not speak no.

§ 627. Questions of appeal.—All questions imply want of information; want of information may then imply doubt; doubt, perplexity; and perplexity the absence of an alternative. In this way, what are called, by Mr. Arnold,^[67] questions of appeal, are, practically speaking, negatives. What should I do? when asked in extreme perplexity, means that nothing can well be done. In the following passage we have the presence of a question instead of a negative:—

§ 628. The following extract from the Philological Museum (vol. ii.) illustrates a curious and minute distinction, which the author shows to have been current when Wicliffe wrote, but which was becoming obsolete when Sir Thomas More wrote. It is an extract from that writer against Tyndall.

"I would not here note by the way that Tyndall here translateth no for nay, for it is but a trifle and mistaking of the Englishe worde: saving that ye shoulde see that he whych in two so plain Englishe wordes, and so common as in naye and no can not tell when he should take the one and when the tother, is not for translating into Englishe a man very mete. For the use of these two wordes in aunswering a question is this. No aunswereth the question framed by the affirmative. As for ensample if a manne should aske Tindall himselfe: ys an heretike meete to translate Holy Scripture into Englishe? lo to thys question if he will aunswere trew Englishe, he must aunswere nay and not no. But and if the question be asked hym thus lo: is not an heretike mete to translate Holy Scripture into Englishe? To this question if he will aunswere trewe Englishe, he must aunswere no and not nay. And a lyke difference is there betwene these two adverbs ye and yes. For if the question bee framed unto Tindall by the affirmative in thys fashion. If an heretique falsely translate the New Testament into Englishe, to make his false heresyes seem the word of Godde, be his bokes worthy to be burned? To this questyon asked in thys wyse, yf he will aunswere true Englishe, he must aunswere ye and not yes. But now if the question be asked him thus lo; by the negative. If an heretike falsely translate the Newe Testament into Englishe to make his false heresyes seme the word of God, be not hys bokes well worthy to be burned? To thys question in thys fashion framed if he will aunswere trewe Englishe he may not aunswere ye but he must answere yes, and say yes marry be they, bothe the translation and the translatour, and al that wyll hold wyth them."

CHAPTER XXVIII.

§ 629. Broadly speaking, all adverbial constructions are absolute. The term, however, is conveniently limited to a particular combination of the noun, verb, and participle. When two actions are connected with each other either by the fact of their simultaneous occurrence, or as cause and effect, they may be expressed within the limits of a single proposition, by expressing the one by means of a verb, and the other by means of a noun and participle agreeing with each other. The door being open, the horse was stolen.

Considering the nature of the connection between the two actions, we find good grounds for expecting à priori that the participle will be in the instrumental case, when such exists in the language; and when not, in some case allied to it, i.e., the ablative or dative.

In Latin the ablative is the case that is used absolutely. Sole orto, claruit dies.

In the present English, however, the nominative is the absolute case. He made the best proverbs, him alone excepted, is an expression of Tillotson's. We should now write he alone excepted. The present mode of expression is only to be justified by considering the nominative form to be a dative one, just as in the expression you are here, the word you, although an accusative, is considered as a nominative. A real nominative absolute is as illogical as a real accusative case governing a verb.

PART VI.

§ 630. Prosody deals with metre; and with accent, quantity and the articulate sounds, as subordinate to metre. For these the reader is referred to Part III. Chapters 1. 6. 7.

Metre is a general term for the recurrence, within certain intervals, of syllables similarly affected.

Syllables may be similarly affected: 1. in respect to their quantities; 2. in respect to their accents; 3. in respect to their articulations.

Here, besides the recurrence of similar accents, there is a recurrence of the same articulate sounds; viz. of o + ld.

§ 631. Metres founded upon the periodic recurrence of similar articulations are of two sorts.

1. Alliterative metres.—In alliterative metres a certain {500}number of words, within a certain period, must begin with a similar articulation.

Alliteration is the general character of all the early Gothic metres. (See Rask's Anglo-Saxon Grammar, Rask, On the Icelandic Prosody, and Conybeare, On Anglo-Saxon Poetry.)

2. Assonant metres.—In assonant metres a certain number of words, within a certain period, must end with a similar articulation. All rhymes and all approaches to rhyme, form the assonant metres. The word assonant has a limited as well as a general sense.

§ 632. All metre goes by the name of poetry, although all poetry is not metrical. The Hebrew poetry (see Lowth, De Sacra Poesi Hebræorum) is characterized by the recurrence of similar ideas.

§ 633. The metres of the classical languages consist essentially in the recurrence of similar quantities; accent also playing a part. The incompatibility of the classical metres with the English prosody lies in the fact (stated at p. 166), that the classic writer measures quantity by the length of the syllable taken altogether, while the Englishman measures it by the length of the vowel alone.

§ 634. The English metres consist essentially of the recurrence of similar accents; the recurrence of similar articulations being sometimes (as in all rhyming poetry) superadded.

§ 635. In the specimen of alliteration lately quoted the only articulation that occurred was the letter c. It is very evident that the two, the three, or the four first letters, or even the whole syllable, might have coincided. Such is the case with the following lines from Lord Byron:

Alliteration, as an ornament, must be distinguished from alliteration as the essential character of metre. Alliteration, as an ornament, is liable to many varieties. {501}

§ 636. Rhyme.—In English versification, rhyme is, next to accent, the most important element. The true nature of a rhyme may best be exhibited after the analysis of a syllable, and the exhibition of certain recurrent combinations, that look like rhyme without being so.

Let the syllable told be taken to pieces. For metrical purposes it consists of three parts or elements: 1, the vowel (o); 2, the part preceding the vowel (t); 3, the part following the vowel (ld). The same may be done with the word bold. The two words can now be compared with each other. The comparison shows that the vowel is in each the same (o); that the part following the vowel (ld) is the same; and, finally, that the part preceding the vowel is different (t and b). This difference between the part preceding the vowel is essential.

Told, compared with itself (told), is no rhyme, but an homœoteleuton (ὁμοῖος, homoios=like, and τελεύτη, teleutæ=end) or like-ending. It differs from a rhyme in having the parts preceding the vowel alike. Absolute identity of termination is not recognized in English poetry, except so far as it is mistaken for rhyme.

Here the difference in spelling simulates a difference in sound, and a homœoteleuton takes the appearance of a rhyme.

Bold and note.—As compared with each other, these words have two of the elements of a rhyme: viz. the identity of the vowel, and the difference of the parts preceding it. They want, however, the third essential, or the identity of the parts following; ld being different from t. The coincidence, however, as far as it goes, constitutes a point in metre. The words in question are assonances in the limited sense of the term; and because the identity lies in the vowels, they may be named vowel assonances. Vowel assonances are recognized in (amongst others) the Spanish and Scandinavian metrical systems. In English they occur only when they pass as rhymes. {502}

Bold and mild.—Here also are two of the elements of a rhyme, viz., the identity of the parts following the vowel (ld), and the difference of the parts preceding (b and m). The identity of the vowel (o being different from i) is, however, wanting. The words in question are assonances in the limited sense of the term, and consonantal assonances. Recognized in the Scandinavian, they occur in English only when they pass as rhymes.

Rhymes may consist of a single syllable, as told, bold, of two syllables, as water, daughter; of three, as cheerily, wearily. Now, the rhyme begins where the dissimilarity of parts immediately before the main vowel begins. Then follows the vowel; and, lastly, the parts after the vowel. All the parts after the vowel must be absolutely identical. Mere similarity is insufficient.

In the lines just quoted there is no rhyme, but an assonance. The identity of the parts after the main syllable is destroyed by the single sound of g in gone.

A rhyme, to be perfect, must fall on syllables equally accented.—To make sky and the last syllable of merrily serve as rhymes, is to couple an accented syllable with an unaccented one.

A rhyme, to be perfect, must fall upon syllables absolutely accented.—To make the last syllables of words like flighty and merrily serve as rhymes, is to couple together two unaccented syllables.

Hence there may be (as in the case of blank verse) accent without rhyme; but there cannot be rhyme without accent.

A rhyme consists in the combination of like and unlike sounds.—Words like I and eye (homœoteleuta), ease and cease (vowel assonances), love and grove (consonantal assonances), are printers' rhymes; or mere combinations of like and unlike letters.

A rhyme, moreover, consists in the combination of like and unlike articulate sounds. Hit and it are not rhymes, but identical endings; the h being no articulation. To my ear, at least, the pair of words, hit and it, comes under a different class from the pair hit (or it) and pit.

§ 637. A full and perfect rhyme (the term being stringently defined) consists in the recurrence of one or more final syllables equally and absolutely accented, wherein the vowel and the part following the vowel shall be identical, whilst the part preceding the vowel shall be different. It is also necessary that the part preceding the vowel be articulate.^[69]

The deviations from the above-given rule, so common in the poetry of all languages, constitute not rhymes, but assonances, &c., that, by poetic licence, are recognized as equivalents to rhymes.

§ 638. Measure.—In lines like the following, the accent occurs on every second syllable; in other words, every accented syllable is accompanied by an unaccented one.

This accented syllable and its accompanying unaccented one constitute a measure. The number of the syllables being two, the measure in question is dissyllabic.

§ 639. In lines like the following the accent falls on every third syllable, so that the number of syllables to the measure is three, and the measure is trisyllabic.

The primary division of the English measures is into the dissyllabic and the trisyllabic.

§ 640. Dissyllabic measures.—The words týrant and presúme are equally dissyllabic measures; in one, however, the accent falls on the first, in the other on the second syllable. This leads us to a farther division of the English measures.

A measure like presúme (where the accent lies on the second syllable) may be repeated throughout a whole verse, or a whole series of verses; as,

A measure like týrant (where the accent lies on the first syllable) may be repeated throughout a whole verse, or a whole series of verses; as,

§ 641. Trisyllabic measures.—The words mérrily, disáble, cavaliér, are equally trisyllabic, but not similarly accented. Each constitutes a separate measure, which may be continued through a whole verse, or a whole series of verses; as,

§ 642. The nature of measures may, as we have already seen, be determined by the proportion of the accented and unaccented syllables. It may also be determined by the proportion of the long and short syllables.—In the one case we measure by the accent, in the other by the quantity. Measures determined by the quantity are called feet. The word foot being thus defined, we have no feet in the English metres; since in English we determine our measures by accent only.

The classical grammarians express their feet by symbols; [ˉ] denoting length, [˘] shortness. Forms like [˘ˉ ˉ˘ ˉ˘˘ ˘ˉ˘ ˘˘ˉ] &c., are the symbolical representations of the classical feet.

The classical grammarians have names for their feet; e.g., iambic is the name of [˘ˉ], trochee of [ˉ˘], dactyle of [ˉ˘˘], amphibrachys of [˘ˉ˘], Anapæst of [˘˘ˉ], &c.

The English grammarians have no symbols for their feet: since they have no form for expressing the absence of the accent. Sometimes they borrow the classical forms [˘] and [ˉ]. These, however, being originally meant for the expression of quantity, confusion arises from the use of them.

Neither have the English grammarians names for their measures. Sometimes, they borrow the classical terms iambic, trochee, &c. These, however, being meant for the expression of quantity, confusion arises from the use of them.

As symbols for the English measures, I indicate the use of a as denoting an accented, x an unaccented syllable; or else that of + as denoting an accented, - an unaccented syllable. Finally, ´ may denote the accent, ¨ the absence of it.

As names for the English measures I have nothing to offer. At times it is convenient to suppose that they have a definite order of arrangement, and to call words like týrant the first measure, and words like presúme the second measure. In like manner, mérrily is measure 3; disáble, 4; and cavaliér, 5. As the number of measures is (from the necessity of the case) limited, this can be done conveniently. The classical {506}names are never used with impunity. Their adoption invariably engenders confusion. It is very true that, mutatis mutandis (i. e., accent being substituted for quantity), words like týrant and presúme are trochees and iambics; but it is also true that, with the common nomenclature, the full extent of the change is rarely appreciated.

That however much one measure may predominate in a series of verses, it is rarely unmixed with others. In

the measure a x appears in the place of x a. This is but a single example of a very general fact, and of a subject liable to a multiplicity of rules.

§ 643. Grouped together according to certain rules, measures constitute lines or verses; and grouped together according to certain rules, lines constitute couplets, triplets, stanzas, &c.

The succession, or periodic return, of rhymes constitutes stanzas, or continuous metre as the case may be.

The quantity of accents in a line constitutes the nature of the verse, taken by itself. {507}

The succession, or periodic return, of verses of the same length has the same effect with the succession, or periodic return, of rhymes; viz., it constitutes stanzas, or continuous metre, as the case may be.

This leads to the nomenclature of the English metres. Of these, none in any of the trisyllabic measures have recognized and technical names; neither have any that are referable to the measure a x.

§ 644. Taking, however, those that are named, we have the following list of terms.

1. Octosyllabics.—Four measures x a, and (unless the rhyme be double) eight syllables. Common in Sir W. Scott's poetry.

2. Heroics.—Five measures x a. This is the common measure in narrative and didactic poetry.

3. Alexandrines.—Six measures x a. This name is said to be taken from the early romances on the deeds of Alexander the Great.

4. Service metre.—Seven measures x a. This is the common metre of the psalm-versions. Thence its name.

§ 645. Such are the names of certain lines or verses taken by themselves. Combined or divided they form—

The heroic couplet is called also riding rhyme; it being the metre wherein Chaucer's Canterbury Tales (told by a party riding to Canterbury) are chiefly written. {508}

2. Heroic triplets.—Same as the preceding, except that three rhymes come in succession.

4. Elegiacs.—The metre of Gray's Elegy. Heroics in four-line stanzas with alternate rhymes.

5. Rhyme royal.—Seven lines of heroics, with the last two rhymes successive, and the first five recurring at intervals. Sometimes the last line is an Alexandrine. There are varieties in this metre according to the intervals of the first five rhymes:—

6. Ottava rima.—The metre in Italian for narrative poetry. Eight lines of heroics; the first six rhyming alternately, the last two in succession.—Byron's Don Juan in English, Orlando Furioso, &c., in Italian.

7. Spenserian stanza.—Eight lines of heroics closed by an Alexandrine. There are varieties of this metre according to the interval of the rhymes.

8. Terza rima.—Taken from the Italian, where it is the metre of Dante's Divina Commedia. Heroics with three rhymes recurring at intervals.—Lord Byron's Prophecy of Dante.

9. Poulterer's measure.—Alexandrines and service measures alternately. Found in the poetry of Henry the Eighth's time.

10. Ballad metre.—Stanzas of four lines; the first and third having four, the second and fourth having three measures each. Rhymes alternate.

§ 646. Scansion.—Let the stanza just quoted be read as two lines, and it will be seen that a couplet of ballad metre is equivalent to a line of service metre. Such, indeed, was the origin of the ballad metre. Observe also the pause (marked |) both in the Alexandrine and the service metres. This indicates a question as to where lines end; in other words, how can we distinguish one long line from two short ones.

It may, perhaps, partake of the nature of a metrical fiction to consider that (in all rhyming poetry) the length of the verse is determined by the occurrence of the rhyme. Nevertheless, as the matter cannot be left to the printer only, and as some definition is requisite, the one in point is attended by as few inconveniences as any other. It must not, however, be concealed that lines as short as

§ 647. In metres of measure a x, the number of syllables is double the number of accents, unless the final rhyme be single; in which case the syllables are the fewest.

In metres of measure x a the number of syllables is double the number of accents, unless the rhyme be double (or treble); in which case the syllables are the most numerous.

Now this view (which may be carried throughout the whole five measures) of the proportion between the accents and the syllables, taken with the fact that it is determined by the nature of the final syllable, indicates a division of our metres into symmetrical (where the number of the syllables is the multiple of the number of accents), and unsymmetrical (where it is not so).

For practical purposes, however, the length of the last measure may be considered as indifferent, and the terms indicated may be reserved for the forthcoming class of metres. {510}

§ 648. Of the metres in question, Coleridge's Christabel and Byron's Siege of Corinth are the current specimens. In the latter we have the couplet:

In the second of these lines, the accents and the syllables are symmetrical; which is not the case with the first. Now to every, or any, accent in the second line an additional unaccented syllable may be added, and the movement be still preserved. It is the fact of the accents and syllables (irrespective of the latitude allowed to the final measure) being here unsymmetrical (or, if symmetrical, only so by accident) that gives to the metres in question their peculiar character. Added to this, the change from x x a, to x a x, and a x x, is more frequent than elsewhere. One point respecting them must be borne in mind; viz., that they are essentially trisyllabic metres from which unaccented syllables are withdrawn, rather than dissyllabic ones wherein unaccented syllables are inserted.

§ 649. Of measures of one, and of measures of four syllables the occurrence is rare, and perhaps equivocal.

§ 650. The majority of English words are of the form a x; that is, words like týrant are commoner than words like presúme.

The multitude of unaccentuated words like the, from, &c., taken along with the fact that they precede the words with which they agree, or which they govern, accounts for the apparent antagonism between the formulæ of our words and the formulæ of our metres. The contrast between a Swedish line of the form a x, and its literal English version (x a), {511}shows this. In Swedish, the secondary part of the construction follows, in English it precedes, the main word:—

This is quoted for the sake of showing the bearing of the etymology and syntax of a language upon its prosody.

§ 651. The classical metres as read by Englishmen.—In p. 500 it is stated that "the metres of the classical languages consist essentially in the recurrence of similar quantities; accent playing a part." Now there are reasons for investigating the facts involved in this statement more closely than has hitherto been done; since the following circumstances make some inquiry into the extent of the differences between the English and the classical systems of metre, an appropriate element of a work upon the English language.

1. The classical poets are authors preeminently familiarized to the educated English reader.

2. The notions imbibed from a study of the classical prosodies have been unduly mixed up with those which should have been derived more especially from the poetry of the Gothic nations.

3. The attempt to introduce (so-called) Latin and Greek metres into the Gothic tongues, has been partially successful on the Continent, and not unattempted in Great Britain.

The second, viz., "that the notions imbibed, &c." will bear some illustration; an illustration which verifies the assertion made in p. 505, that the English grammarians "sometimes borrow the classical terms iambic, trochee," &c., and apply them to their own metres.

How is this done? In two ways, one of which is wholly incorrect, the other partially correct, but inconvenient.

To imagine that we have in English, for the practical purposes of prosody, syllables long in quantity or short in quantity, syllables capable of being arranged in groups {512}constituting feet, and feet adapted for the construction of hexametres, pentametres, sapphics, and alcaics, just as the Latins and Greeks had, is wholly incorrect. The English system of versification is founded, not upon the periodic recurrence of similar quantities, but upon the periodic recurrence of similar accents.

The less incorrect method consists in giving up all ideas of the existence of quantity, in the proper sense of the word, as an essential element in English metre; whilst we admit accent as its equivalent; in which case the presence of an accent is supposed to have the same import as the lengthening and the absence of one, as the shortening of a syllable; so that, mutatis mutandis, a is the equivalent to [ˉ], and x to [˘].

§ 653. With this view there are a certain number of classical feet, with their syllables affected in the way of quantity, to which they are equivalent English measures with their syllables affected in the way of accent. Thus if the formula

And so on in respect to the larger groups of similarly affected syllables which constitute whole lines and stanzas; verses like

are (A), trochaic; (B), iambic; (C), dactylic; (D), amphibrachych; and (E), anapæstic, respectively.

And so, with the exception of the word amphibrachych (which I do not remember to have seen) the terms have been used. And so, with the same exception, systems of versification have been classified.

§ 654. Reasons against the classical nomenclature as applied to English metres.—These lie in the two following facts:—

1. Certain English metres have often a very different character from their supposed classical analogues.

§ 655. Certain English metres have often a very different metrical character, &c.—Compare such a so-called English anapæst as—

For the latter line to have the same movement as the former, it must be read thus—

Now we well know that, whatever may be any English scholar's notions of the Greek accents, this is not the way in which he reads Greek anapæsts.

Again the trochaic movement of the iambic senarius is a point upon which the most exclusive Greek metrists have insisted; urging the necessity of reading (for example) the first line in the Hecuba—

§ 656. I have said that certain English metres have often a very different metrical character, &c. I can strengthen the reasons against the use of classical terms in English prosody, by enlarging upon the word often. The frequency of the occurrence of a difference of character between classical and English metres similarly named is not a matter of accident, but is, in many cases, a necessity arising out of the structure of the English language as compared with that of the Greek and Latin—especially the Greek.

With the exception of the so-called second futures, there is no word in Greek whereof the last syllable is accented. Hence, no English line ending with an accented syllable can have a Greek equivalent. Accent for accent—

but no Greek word (with the exception of the so-called second futures like νεμῶ=nemô) and (probably) no Latin word at all, is accented like presúme and cavalíer.

From this it follows that although the first three measures of such so-called English anapæsts as—

may be represented by Greek equivalents (i. e., equivalents in the way of accent)—

a parallel to the last measure (-ery stréet) can only be got at by one of two methods; i. e., by making the verse end in a so-called second future, or else in a vowel preceded by an accented syllable, and cut off—

Now it is clear that when, over and above the fact of certain Greek metres having a different movement from their supposed English equivalents, there is the additional circumstance of such an incompatibility being less an accident than a necessary effect of difference of character in the two languages, the use of terms suggestive of a closer likeness than either does or ever can exist is to be condemned; and this is the case with the words, dactylic, trochaic, iambic, anapæstic, as applied to English versification.

§ 657. Certain classical feet have no English equivalents.—Whoever has considered the principles of English prosody, must have realized the important fact that, ex vi termini, no English measure can have either more or less than one accented syllable.

On the other hand, the classical metrists have several measures in both predicaments. Thus to go no farther than the trisyllabic feet, we have the pyrrhic ([˘ ˘]) and tribrach ([˘ ˘ ˘]) without a long syllable at all, and the spondee ([ˉ ˉ]), amphimacer ([ˉ ˘ ˉ]), and molossus ([ˉ ˉ ˉ]) with more than one long syllable. It follows, then that (even mutatis mutandis, i.e., with the accent considered as the equivalent to the long syllable) English pyrrhics, English tribrachs, English amphimacers, English spondees, and English molossi are, each and all, prosodial impossibilities.

It is submitted to the reader that the latter reason (based wholly upon the limitations that arise out of the structure of language) strengthens the objections of the previous section.

§ 658. The classical metres metrical even to English readers. The attention of the reader is directed to the difficulty involved in the following (apparently or partially) contradictory facts.

1. Accent and quantity differ; and the metrical systems founded upon them differ also.

4. Nevertheless, notwithstanding the difference of the principle upon which they are constructed, the classical metres, even as read by Englishmen, and read accentually, are metrical to English ears.

§ 659. Preliminary to the investigation of the problem in question it is necessary to remark—

1. That, the correctness or incorrectness of the English pronunciation of the dead languages has nothing to do with the matter. Whether we read Homer exactly, as Homer would read his own immortal poems, or whether we read them in such a way as would be unintelligible to Homer reappearing upon earth, is perfectly indifferent.

2. That whether, as was indicated by the author of Μέτρον ἄριστον, we pronounce the anapæst pătŭlæ, precisely as we pronounce the dactyle Tītўrĕ, or draw a distinction between them is also indifferent. However much, as is done in some of the schools, we may say scri-bere rather than scrib-ere, or am-or, rather than a-mor, under the notion that we are lengthening or shortening certain syllables, one unsurmountable dilemma still remains, viz., that the shorter we pronounce the vowel, the more we suggest the notion of the consonant which follows it being doubled; whilst double consonants lengthen the vowel which precedes them. Hence, whilst it is certain that patulæ and Tityre may be pronounced (and that without hurting the metre) so as to be both of the same quantity, it is doubtful what that quantity is. Sound for sound Tĭtyre may be as short as pătulæ. Sound for sound pāttulæ may be as long as Tīttyre.

a. That Englishmen do not read the classical metres according to their quantities.

§ 660. Why are the classical metres metrical to English readers?—Notwithstanding the extent to which quantity differs from accent, there is no metre so exclusively founded upon the former as to be without a certain amount of the {517}latter; and in the majority (at least) of the classical (and probably other) metres there is a sufficient amount of accentual elements to constitute metre; even independent of the quantitative ones.

§ 661. Latitude in respect to the periodicity of the recurrence of similarly accented syllables in English.—Metre (as stated in p. 499), "is the recurrence, within certain intervals, of syllables similarly affected."

The particular way in which syllables are affected in English metre is that of accent.

The more regular the period at which similar accents recur the more typical the metre.

§ 662. Symmetrical metres.—Allowing for indifference of the number of syllables in the last measure, it is evident that in all lines where the measures are dissyllabic the syllables will be a multiple of the accents, i. e., they will be twice as numerous. Hence, with three accents there are six syllables; with four accents, eight syllables, &c.

Similarly, in all lines where the measures are trisyllabic the syllables will also be multiples of the accents, i. e., they will be thrice as numerous. Hence, with three accents there will be nine syllables, with four accents, twelve syllables, and with seven accents, twenty-one syllables.

§ 663. Unsymmetrical metres.—Lines, where the syllables are not a multiple of the accents, may be called unsymmetrical. Occasional specimens of such lines occur interspersed amongst others of symmetrical character. Where this occurs the general character of the versification may be considered as symmetrical also.

The case, however, is different where the whole character of the versification is unsymmetrical, as it is in the greater part of Coleridge's Christabel, and Byron's Siege of Corinth. {518}

§ 664. Many (perhaps all) classical metres on a level with the unsymmetrical English ones.—The following is the notation of the extract in the preceding section.

Now many Latin metres present a recurrence of accent little more irregular than the quotation just analysed. The following is the accentual formula of the first two stanzas of the second ode of the first Book of Horace. {519}

A longer list of examples would show us that, throughout the whole of the classical metres the same accents recur, sometimes with less, and sometimes with but very little more irregularity than they recur in the unsymmetrical metres of our own language.

§ 665. Conversion of English into classical metres.—In the preface to his Translation of Aristophanes, Mr. Walsh has shown (and, I believe, for the first time), that, by a different distribution of lines, very fair hexameters may be made out of the well-known lines on the Burial of Sir John Moore:—

§ 667. The rationale of so remarkable a phænomenon as regularity of accent in verses considered to have been composed with a view to quantity only has yet to be investigated. That it was necessary to the structure of the metres in question is certain.

§ 668. Cæsura.—The cæsura of the classical metrists is the result of—

1. The necessity in the classical metres (as just indicated) of an accented syllable in certain parts of the verses.

2. The nearly total absence in the classical languages of words with an accent on the last syllable.

From the joint effect of these two causes, it follows that in certain parts of a verse no final syllable can occur, or (changing the expression) no word can terminate.

Thus, in a language consisting chiefly of dissyllables, of which the first alone was accented, and in a metre which required the sixth syllable to be accented, the fifth and seventh would each be at end of words, and that simply because the sixth was not.

Whilst in a language consisting chiefly of either dissyllables or trisyllables, and in a metre of the same sort as before, {521}if the fifth were not final, the seventh would be so, or vice versa.

§ 669. Cæsura means cutting. In a language destitute of words accented on the last syllable, and in a metre requiring the sixth syllable to be accented, a measure (foot) of either the formula x a, or x x a (i. e., a measure with the accent at the end), except in the case of words of four or more syllables, must always be either itself divided, or else cause the division of the following measures—division meaning the distribution of the syllables of the measure (foot) over two or more words. Thus—

a. If the accented syllable (the sixth) be the first of a word of any length, the preceding one (the fifth) must be the final one of the word which went before; in which case the first and last parts belong to different words, and the measure (foot) is divided or cut.

b. If the accented syllable (the sixth) be the second of a word of three syllables, the succeeding one which is at the end of the word, is the first part of the measure which follows; in which case the first and last parts of the measure (foot) which follows the accented syllable is divided or cut.

As the cæsura, or the necessity for dividing certain measures between two words, arises out of the structure of language, it only occurs in tongues where there is a notable absence of words accented on the last syllable. Consequently there is no cæsura^[71] in the English.

§ 670. As far as accent is concerned, the classical poets write in measures rather than feet. See p. 505.

§ 671. Although the idea of writing English hexameters, &c., on the principle of an accent in a measure taking the place of the long syllables in a foot, is chimerical; it is perfectly practicable to write English verses upon the same {523}principle which the classics themselves have written on, i.e., with accents recurring within certain limits; in which case the so-called classical metre is merely an unsymmetrical verse of a new kind. This may be either blank verse or rhyme.

§ 672. The chief reason against the naturalization of metres of the sort in question (over and above the practical one of our having another kind in use already), lies in the fact of their being perplexing to the readers who have not been {525}trained to classical cadences, whilst they suggest and violate the idea of quantity to those who have.

may be read in two ways. We may either lay full stress upon the word ere, and read—

or we may lay little or no stress upon either ere or her, reserving the full accentuation for the syllable faith- in faithless, in which case the reading would be

Lines of this sort may be called examples of convertible metres, since by changing the accent a dissyllabic line may be converted into one partially trisyllabic, and vice versâ.

This property of convertibility is explained by the fact of accentuation being a relative quality. In the example before us ere is sufficiently strongly accented to stand in contrast to her, but it is not sufficiently strongly accented to stand upon a par with the faith- in faithless if decidedly pronounced.

The real character of convertible lines is determined from the character of the lines with which they are associated. {527}That the second mode of reading the line in question is the proper one, may be shown by reference to the stanza wherein it occurs.

would be read improperly. The stanza wherein it occurs is essentially dissyllabic (a x).

§ 674. Metrical and grammatical combinations.—Words, or parts of words, that are combined as measures, are words, or parts of words, combined metrically, or in metrical combination.

Syllables combined as words, or words combined as portions of a sentence, are syllables and words grammatically combined, or in grammatical combination.

When the syllables contained in the same measure (or connected metrically) are also contained in the same construction (or connected grammatically), the metrical and the grammatical combinations coincide. Such is the case with the line

where the same division separates both the measure and the subdivisions of the sense, inasmuch as the word the is connected with the word glories equally in grammar and in metre, in syntax and in prosody. So is of with Brian, and the with Brave.

Here the metrical division is one thing, the grammatical division another, and there is no coincidence.

In the following stanza the coincidence of the metrical and grammatical combination is nearly complete:—

§ 675. Rhythm.—The character of a metre is marked and prominent in proportion as the metrical and the grammatical {529}combinations coincide. The extent to which the measure a x x is the basis of the stanza last quoted is concealed by the antagonism of the metre and the construction. If it were not for the axiom, that every metre is to be considered uniform until there is proof to the contrary, the lines might be divided thus:—

The variety which arises in versification from the different degrees of the coincidence and non-coincidence between the metrical and grammatical combinations may be called rhythm.

§ 676. Constant and inconstant parts of a rhythm.—See § 636. Of the three parts or elements of a rhyme, the vowel and the part which follows the vowel are constant, i.e., they cannot be changed without changing or destroying the rhyme. In told and bold, plunder, blunder, both the o or u on one side, and the -ld or -nder on the other are immutable.

Of the three parts, or elements, of a rhyme the part which precedes the vowel is inconstant, i.e, it must be changed in order to effect the rhyme. Thus, old and old, told and told, bold and bold, do not rhyme with each other; although old, bold, told, scold, &c. do.

Rule 1. In two or more syllables that rhyme with each other, neither the vowel nor the sounds which follow it can be different.

Rule 2. In two or more syllables that rhyme with each other, the sounds which precede the vowel cannot be alike.

Now the number of sounds which can precede a vowel is limited: it is that of the consonants and consonantal combinations; of which a list can be made a priori.

1. Divide the word to which a rhyme is required, into its constant and inconstant elements.

2. Make up the inconstant element by the different consonants and consonantal combinations until they are exhausted.

3. In the list of words so formed, mark off those which have an existence in the language; these will all rhyme with each other; and if the list of combinations be exhaustive, there are no other words which will do so.

Example.—From the word told, separate the o and -ld, which are constant.

Instead of the inconstant element t, write successively, p, pl, pr, b, bl, br, &c.: so that you have the following list:—t-old, p-old, pl-old, pr-old, b-old, bl-old, br-old, &c.

Of these plold, blold, and brold, have no existence in the language; the rest, however, are rhymes.

§ 677. All words have the same number of possible, but not the same number of actual rhymes. Thus, silver is a word amenable to the same process as told—pilver, plilver, prilver, bilver, &c.; yet silver is a word without a corresponding rhyme. This is because the combinations which answer to it do not constitute words, or combinations of words in the English language.

This has been written, not for the sake of showing poets how to manufacture rhymes, but in order to prove that a result which apparently depends on the ingenuity of writers, is reducible to a very humble mechanical process, founded upon the nature of rhyme and the limits to the combinations of consonants.

PART VII.

§ 678. The consideration of the dialects of the English language is best taken in hand after the historical investigation of the elements of the English population. For this, see Part I.

It is also best taken in hand after the analysis of the grammatical structure of the language. For this, see Part IV.

This is because both the last-named subjects are necessary as preliminaries. The structure of the language supplies us with the points in which one dialect may differ from another, whilst the history of the immigrant populations may furnish an ethnological reason for such differences as are found to occur.

§ 679. By putting together the history of the migrations into a country, and the grammatical structure of the language which they introduced, we find that there are two methods of classifying the dialects. These may be called the ethnological, and the structural methods.

According to the former, we place in the same class those dialects which were introduced by the same section of immigrants. Thus, a body of Germans, starting from the same part of Germany, and belonging to the same section of the Germanic population, even if, whilst at sea, they separated into two, three, or more divisions, and landed upon widely separated portions of Great Britain, would introduce dialects which were allied ethnologically; even though, by one of them changing rapidly, and the others not changing at all, they might, in their external characters, differ from each other, and agree with dialects of a different introduction. Hence, the ethnological principle is essentially historical, and {532}is based upon the idea of affiliation or affinity in the way of descent.

The structural principle is different. Two dialects introduced by different sections (perhaps it would be better to say sub-sections) of an immigrant population may suffer similar changes; e. g., they may lose the same inflexions, adopt similar euphonic processes, or incorporate the same words. In this case, their external characters become mutually alike. Hence, if we take two (or move) such dialects, and place them in the same class, we do so simply because they are alike; not because they are affiliated.

Such are the two chief principles of classification. Generally, they coincide; in other words, similarity of external characters is primâ facie evidence of affinity in the way of affiliation, identity of origin being the safest assumption in the way of cause; whilst identity of origin is generally a sufficient ground for calculating upon similarity of external form; such being, a priori, its probable effect.

Still, the evidence of one in favour of the other is only primâ facie evidence. Dialects of the same origin may grow unlike; dialects of different origins alike.

§ 680. The causes, then, which determine those minute differences of language, which go by the name of dialects are twofold.—1. Original difference; 2. Subsequent change.

§ 681. The original difference between the two sections (or sub-sections) of an immigrant population are referable to either—1. Difference of locality in respect to the portion of the country from which they originated; or 2. Difference in the date of the invasion.

Two bodies of immigrants, one from the Eyder, and the other from the Scheldt, even if they left their respective localities on the same day of the same month, would most probably differ from one another; and that in the same way that a Yorkshireman differs from a Hampshire man.

On the other hand, two bodies of immigrants, each leaving the very same locality, but one in 200 A.D., and the other in 500 A.D., would also, most probably, differ; and that as a Yorkshireman of 1850 A.D. differs from one of 1550 A.D. {533}

§ 682. The subsequent changes which may affect the dialect of an immigrant population are chiefly referable to either, 1. Influences exerted by the dialects of the aborigines of the invaded country; 2. Influences of simple growth, or development. A dialect introduced from Germany to a portion of Great Britain, where the aborigines spoke Gaelic, would (if affected at all by the indigenous dialect) be differently affected from a dialect similarly circumstanced in a British, Welsh, and Cambrian district.

A language which changes rapidly, will, at the end of a certain period, wear a different aspect from one which changes slowly.

§ 683. A full and perfect apparatus for the minute philology of the dialects of a country like Great Britain, would consist in—

3. The exact details of the provincialisms of the whole of that part of Germany which contributed, or is supposed to have contributed, to the Anglo-Saxon immigration.

4. The details of the original languages or dialects of the Aboriginal Britons at the time of the different invasions.

§ 684. Such are the preliminaries which are wanted for the purposes of investigation. Others are requisite for the proper understanding of the facts already ascertained, and the doctrines generally admitted; the present writer believing that these two classes are by no means coextensive.

Of such preliminaries, the most important are those connected with 1. the structure of language, and 2. the history of individual documents; in other words, certain points of philology, and certain points of bibliography.

§ 685. Philological preliminaries.—These are points of pronunciation, points of grammatical structure, and glossarial peculiarities. It is only the first two which will be noticed. They occur in 1. the modern, 2. the ancient local forms of speech. {534}

§ 686. Present provincial dialects.—In the way of grammar we find, in the present provincial dialects (amongst many others), the following old forms—

1. A plural in en—we call-en, ye call-en, they call-en. Respecting this, the writer in the Quarterly Review, has the following doctrine:—

"It appears to have been popularly known, if not in East Anglia proper, at all events in the district immediately to the westward, since we find it in Orm, in an Eastern-Midland copy of the Rule of Nuns, sæc. XIII., and in process of time in Suffolk. Various conjectures have been advanced as to the origin of this form, of which we have no certain examples before the thirteenth century.^[72] We believe the true state of the case to have been as follows. It is well known that the Saxon dialects differ from the Gothic, Old-German, &c. in the form of the present indicative plural—making all three persons to end in -aþ or -ad;—we—ȝe—hi—lufi-aþ (-ad). Schmeller and other German philologists observe that a nasal has been here elided, the true ancient form being -and, -ant, or -ent. Traces of this termination are found in the Cotton MS. of the Old Saxon Evangelical Harmony, and still more abundantly in the popular dialects of the Middle-Rhenish district from Cologne to the borders of Switzerland. These not only exhibit the full termination -ent, but also two modifications of it, one dropping the nasal and the other the dental. E.g.:—

—the last exactly corresponding with the Mercian. It is remarkable that none of the above forms appear in classical German compositions, while they abound in the Miracle-plays, vernacular sermons, and similar productions of the thirteenth and fourteenth centuries, specially addressed to the uneducated classes. We may, therefore, reasonably conclude from analogy that similar forms were popularly current in our midland counties, gradually insinuating themselves into the {535}written language. We have plenty of examples of similar phenomena. It would be difficult to find written instances of the pronouns scho, or she, their, you, the auxiliaries sal, suld, &c., before the twelfth century; but their extensive prevalence in the thirteenth proves that they must have been popularly employed somewhere even in times which have left us no documentary evidence of their existence."

I prefer to consider this termination as -en, a mere extension of the subjunctive form to the indicative.

2. An infinitive form in -ie; as to sowie, to reapie,—Wiltshire. (Mr. Guest).

3. The participial form in -and; as goand, slepand,—Lincolnshire (?), Northumberland, Scotland.

4. The common use of the termination -th in the third person present; goeth, hath, speaketh,—Devonshire.

The following changes (a few out of many) are matters not of grammar, but of pronunciation:—

Oy for i—foyne, twoyne, for fine, twine,—Cheshire, Cambridgeshire, Suffolk.

Oy for o—noite, foil, coil, hoil, for note, foal, coal, hole,—Halifax.

Ooy for oo—nooin, gooise, fooil, tooil, for noon, goose, fool, tool,—Halifax. {536}

W inserted (with or without a modification)—as spwort, scworn, whoam, for sport, scorn, home,—Cumberland, West Riding of Yorkshire.

Iv for oo, or yoo when a vowel follows—as Samivel for Samuel; Emmanivel for Emmanuel. In all these we have seen a tendency to diphthongal sounds.

In the following instances the practice is reversed, and instead of the vowel being made a diphthong, the diphthong becomes a vowel, as,

Of these the substitution of oo for ow, and of ee for i, are of importance in the questions of the Appendix.

Ā for ō—sair, mair, baith, for sore, more, both,—Cumberland, Scotland.

O for ă—mon for man,—Cheshire. Lond for land,—East-Anglian Semi-Saxon.

Y inserted before a vowel—styake, ryape, for stake, rope,—Borrowdale; especially after g (a point to be noticed), gyarden, gyown, for garden, gown,—Warwickshire, &c.; and at the beginning of a word, as yat, yan, for ate, one (ane),—Westmoreland, Bilsdale.

P for b—poat for boat.—Welsh pronunciation of many English words. See the speeches of Sir Hugh Evans in Merry Wives of Windsor.

D for th (þ)—wid=with; tudder=the other,—Borrowdale, Westmoreland. Initial (especially before a consonant)—drash, droo=thrash, through,—Devonshire, Wilts.

K for ch (tsh)—thack, pick, for thatch, pitch,—Westmoreland, Lincolnshire, Halifax.

G preserved from the Anglo-Saxon—lig, lie. Anglo-Saxon, licgan,—Lincolnshire, North of England.

K before s, the preceding vowel being lengthened by way of compensation—neist for next, seist for sixth,—Halifax.

D and v after a consonant—gol for gold, siller for silver,—Suffolk. The ejection of f is rarer; mysel for myself, however, occurs in most dialects.

L final, after a short vowel,—in which case the vowel is lengthened—poo for pull,—Cheshire, Scotland.

Al changed to a open—hawf for half, saumon for salmon,—Cumberland, Scotland.

Transpositions of the liquid r are common in all our provincial dialects; as gars, brid, perty, for grass, bird, pretty. Here the provincial forms are the oldest, gærs, brid, &c., being the Anglo-Saxon forms. Again; acsian, Anglo-Saxon=ask, English.

1. The ge- (see § 409), prefixed to the past participle (ge-boren=borne) is, in certain localities,^[73] omitted.

2. The present^[74] plural form -s, encroaches upon the form in -n. Thus, munuces=munucan=monks.

3. The infinitive ends in -a, instead of -an. This is Scandinavian, but it is also Frisian.

5. The article^[74] the is used instead of se, seo, þæt=ὁ, ἡ, τὸ, for both the numbers, and all the cases and genders.

1. Forms with the slenderer, or more vocalic^[74] sounds, replace forms which in the West-Saxon are broad or diphthongal.^[75] Beda mentions that Cœlin is the Northumbrian form of Ceawlin.

2. The simple^[74] sound of k replaces the combination out of which the modern sound of ch has been evolved.

3. The sound of sk replaces either the sh, or the sound out of which it has been evolved.

The meaning of these last two statements is explained by the following extract: "Another characteristic is the infusion of Scandinavian words, of which there are slight traces in monuments of the tenth century, and strong and unequivocal ones in those of the thirteenth and fourteenth. Some of the above criteria may be verified by a simple and obvious process, namely, a reference to the topographical nomenclature of our provinces. Whoever takes the trouble to consult the Gazetteer of England will find, that of our numerous 'Carltons' not one is to be met with south of the Mersey, west of the Staffordshire Tame, or south of the Thames; and that 'Fiskertons,' 'Skiptons,' 'Skelbrookes,' and a whole host of similar names are equally introuvables in the same district. They are, with scarcely a single exception, northern or eastern; and we know from Ælfric's Glossary, from Domesday and the Chartularies, that this distinction of pronunciation was established as early as the eleventh century. 'Kirby' or 'Kirkby,' is a specimen of joint Anglian and {539}Scandinavian influence, furnishing a clue to the ethnology of the district wherever it occurs. The converse of this rule does not hold with equal universality, various causes having gradually introduced soft palatal sounds into districts to which they did not properly belong. Such are, however, of very partial occurrence, and form the exception rather than the rule."—Quarterly Review, No. CLXIV.

Bibliographical preliminaries.—The leading facts here are the difference between 1. the locality of the authorship, and 2, the locality of the transcription of a book.

Thus: the composition of a Devonshire poet may find readers in Northumberland, and his work be transcribed by Northumbrian copyist. Now this Northumbrian copyist may do one of two things: he may transcribe the Devonian production verbatim et literatim; in which case his countrymen read the MS. just as a Londoner reads Burns, i.e., in the dialect of the writer, and not in the dialect of the reader. On the other hand, he may accommodate as well as transcribe, i.e., he may change the non-Northumbrian into Northumbrian expressions, in which case his countrymen read the MS. in their own rather than the writer's dialect.

Now it is clear, that in a literature where transcription, combined with accommodation, is as common as simple transcription, we are never sure of knowing the dialect of an author unless we also know the dialect of his transcriber. In no literature is there more of this semi-translation than in the Anglo-Saxon and the early English; a fact which sometimes raises difficulties, by disconnecting the evidence of authorship with the otherwise natural inferences as to the dialect employed; whilst, at others, it smoothes them away by supplying as many specimens of fresh dialects, as there are extant MSS. of an often copied composition.

Inquiring whether certain peculiarities of dialect in Layamon's Brut, really emanated from the author, a writer in the Quarterly Review, (No. clxiv.) remarks, that to decide this it "would be necessary to have access either to the priest's autograph, or to a more faithful copy of it than it was the practice to make either in his age or the succeeding {540}ones. A transcriber of an early English composition followed his own ideas of language, grammar, and orthography; and if he did not entirely obliterate the characteristic peculiarities of his original, he was pretty sure, like the Conde de Olivares, 'd'y meter beaucour du sein.' The practical proof of this is to be found in the existing copies of those works, almost every one of which exhibits some peculiarity of features. We have 'Trevisa' and 'Robert of Gloucester,' in two distinct forms—'Pier's Ploughman,' in at least three, and 'Hampole's Pricke of Conscience,' in half a dozen, without any absolute certainty which approximates most to what the authors wrote. With regard to Layamon, it might be supposed that the older copy is the more likely to represent the original; but we have internal evidence that it is not the priest's autograph; and it is impossible to know what alterations it may have undergone in the course of one or more transcriptions."

Again, in noticing the orthography of the Ormulum (alluded to in the present volume, § 266), he writes: "It is true that in this instance we have the rare advantage of possessing the author's autograph, a circumstance which cannot with confidence be predicated of any other considerable work of the same period. The author was, moreover, as Mr. Thorpe observes, a kind of critic in his own language; and we therefore find in his work, a regularity of orthography, grammar, and metre, hardly to be paralleled in the same age. All this might, in a great measure, disappear in the very next copy; for fidelity of transcription was no virtue of the thirteenth or the fourteenth century; at least with respect to vernacular works. It becomes, therefore, in many cases a problem of no small complication, to decide with certainty respecting the original metre, or language, of a given mediæval composition, with such data as we now possess."

From all this it follows, that the inquirer must talk of copies rather than of authors.

§ 688. Caution.—Differences of spelling do not always imply differences of pronunciation; perhaps they may be primâ facie of such. Still it is uncritical to be over-hasty in {541}separating, as specimens of dialect, works which, perhaps, only differ in being specimens of separate orthographies.

§ 689. Caution.—The accommodation of a transcribed work is susceptible of degrees. It may go so far as absolutely to replace one dialect by another, or it may go no farther than the omission of the more unintelligible expressions, and the substitution of others more familiar. I again quote the Quarterly Review,—"There are very few matters more difficult than to determine à priori, in what precise form a vernacular composition of the thirteenth century might be written, or what form it might assume in a very short period. Among the Anglo-Saxon charters of the eleventh and twelfth centuries, many are modelled upon the literary Anglo-Saxon, with a few slight changes of orthography and inflection; while others abound with dialectical peculiarities of various sorts. Those peculiarities may generally be accounted for from local causes. An East-Anglian scribe does not employ broad western forms, nor a West of England man East-Anglian ones; though each might keep his provincial peculiarities out of sight, and produce something not materially different from the language of Ælfric."

§ 690. Caution.—In the Reeve's Tale, Chaucer puts into the mouth of one of his north-country clerks, a native of the Strother, in the north-west part of the deanery of Craven, where the Northumbrian dialect rather preponderates over the Anglian, certain Yorkshire glosses. "Chaucer^[76] undoubtedly copied the language of some native; and the general accuracy, with which he gives it, shows that he was an attentive observer of all that passed around him.

"We subjoin an extract from the poem, in order to give our readers an opportunity of comparing southern and northern English, as they co-existed in the fifteenth century. It is from a MS. that has never been collated; but which we believe to be well worthy the attention of any future editor of the Canterbury Tales. The italics denote variations from the printed text:—

"Excepting the obsolete forms hethen (hence), swa, lorn, whilke, alswa, capel—all the above provincialisms are still, more or less, current in the north-west part of Yorkshire. Na, ham(e), fra, banes, attanes, ra, bathe, are pure Northumbrian. Wang (cheek or temple) is seldom heard, except in the phrase wang tooth, dens molaris. Ill, adj., for bad—lathe (barn)—and fond (foolish)—are most frequently and familiarly used in the West Riding, or its immediate borders."

Now this indicates a class of writings which, in the critical history of our local dialect, must be used with great caution and address. An imitation of dialect may be so lax as to let its only merit consist in a deviation from the standard idiom.

In the Lear of Shakspeare we have speeches from a Kentish clown. Is this the dialect of the character, the dialect of the writer, or is it some conventional dialect appropriated to theatrical purposes? I think the latter.

In Ben Jonson's Tale of a Tub, one (and more than one of the characters) speaks thus. His residence is the neighbourhood of London, Tottenham Court.

I consider that this represents the dialect of the neighbourhood of London, not on the strength of its being put in the mouth of a man of Tottenham, but from other and independent circumstances.

Not so, however, with the provincialisms of another of Ben Jonson's plays, the Sad Shepherd:—

The scene of the play is Sherwood Forest: the language, however, as far as I may venture an opinion, is not the language from which the present Nottinghamshire dialect has come down.

§ 691. Caution.—Again, the word old, as applied to language, has a double meaning.

The language of the United States was imported from England into America in the reign of Queen Elizabeth. The language of South Australia has been introduced within the present generation. In one sense, the American English is older than the Australian. It was earliest separated from the mother-tongue.

The language, however, of America may (I speak only in the way of illustration, and consequently hypothetically), in the course of time, become the least old of the two; the word old being taken in another sense. It may change with greater rapidity. It may lose its inflections. It may depart more from the structure of the mother-tongue, and preserve fewer of its old elements. In this sense the Australian (provided that it has altered least, and that it retain the greatest number of the old inflections) will be the older tongue of the two.

Now what may be said of the language of two countries, may be said of the dialects of two districts. The one dialect may run its changes apace; the other alter but by degrees. {545}Hence, of two works in two such dialects, the one would appear older than the other, although in reality the two were cotemporary.

Hence, also, it is a lax expression to say that it is the old forms (the archaisms) that the provincial dialects retain. The provincial forms are archaic only when the current language changes more rapidly than the local idiom. When the local idiom changes fastest, the archaic forms belong to the standard mode of speech.

The provincial forms, goand, slepand, for going and sleeping, are archaic. Here the archaism is with the provincial form.

The forms almost, horses, nought but, contrasted with the provincialisms ommost, hosses, nobbot, are archaic. They have not been changed so much as they will be. Here the archaism (that is, the nearer approach to the older form) is with the standard idiom. A sequestered locality is preservative of old forms. But writing and education are preservatives of them also.

§ 692. With these preliminaries a brief notice of the English dialects, in their different stages, may begin.

The districts north of the Humber.—There is so large an amount of specimens of the dialects of this area in the Anglo-Saxon stage of our language, the area itself so closely coincides with the political division of the kingdom of Northumberland, whilst the present arrangement (more or less provisional) of the Anglo-Saxon dialects consists of the divisions of them into the, 1, West-Saxon; 2, Mercian; and 3, Northumbrian, that it is best to give a general view of the whole tract before the minuter details of the different counties which compose them are noticed. The data for the Northumbrian division of the Anglo-Saxon dialects are as follows:—

1. Wanley's Fragment of Cædmon.—The north-east of Yorkshire was the birth-place of the Anglo-Saxon monk Cædmon. Nevertheless, the form in which his poems in full have come down to us is that of a West-Saxon composition. This indicates the probability of the original work having first been re-cast, and afterwards lost. Be this as it may, the {546}following short fragment has been printed by Wanley, from an ancient MS., and by Hickes from Bede, Hist. Eccl., 4, 24, and it is considered, in the first form, to approach or, perhaps, to represent the Northumbrian of the original poem.

3. The Ruthwell Runes.—The inscription in Anglo-Saxon Runic letters, on the Ruthwell Cross, is thus deciphered and translated by Mr. Kemble:—

"The dialect of these lines is that of Northumberland in the seventh, eighth, and even ninth centuries. The first peculiarity is in the æ for e in the oblique cases, and which I have observed in the cotemporary MS. of Cuðberht's letter at St. Gallen. This, which is strictly organic, and represents the uncorrupted Gothic genitive in -as, and dative in -a, as well as the Old Saxon forms of the substantive, is evidence of great antiquity. But that which is, perhaps, the most characteristic of the Northumbrian dialect is the formation of the infinitive in -a and -æ, instead of -an (hældæ, gistiga). The Durham Book has, I believe, throughout but one single verb, which makes the infinitive in -an, and that is the anomalous word bean=to be; even wosa and wiortha following the common rule. The word ungket is another incontrovertible proof of extreme antiquity, having, to the best of my knowledge, never been found but in this passage. It is the dual of the first personal pronoun Ic, and corresponds to the very rare dual of the second personal pronoun incit, which occurs twice in Cædmon."^[78]

4. The Cotton Psalter.—This is a Latin Psalter in the Cotton collection, accompanied by an Anglo-Saxon interlineation. Place uncertain. Time, ninth century or earlier. The following points of difference between this and the West-Saxon are indicated by Mr. Garnett, Phil. Soc. No. 27.

5. The Durham Gospels—Quatuor Evangelia Latine, ex translatione B. Hieronymi, cum glossâ interlineatâ Saxonica. Nero, D. 4.

6. The Rituale Ecclesiæ Dunhelmensis.—Edited for the Surtees Society by Mr. Stevenson. Place: neighbourhood of Durham. Time: A.D. 970. Differences between the Psalter and Ritual:—

a. The form for the first person is in the Psalter generally -u. In the Ritual it is generally -o. In West Saxon, -e.

Psalter.—Getreow-u, I believe; cleopi-u, I call; sell-u, I give; ondred-u, I fear; ageld-u, I pay; getimbr-u, I build. Forms in -o; sitt-o, I sit; drinc-o, I drink.

b. In the West Saxon the plural present of verbs ends in -að: we lufi-að, ge lufi-að, hi lufi-að. The Psalter also exhibits this West Saxon form. But the plurals of the Ritual {550}end in -s: as, bidd-as=we pray; giwoed-es=put on; wyrc-as=do.

c. The infinitives of verbs end in the West Saxon in -an, as cwed-an=to say. So they do in the Psalter. But in the Ritual the -n is omitted, and the infinitive ends simply in -a: cuoetha=to say; inngeonga=to enter.

d. The oblique cases and plurals of substantives in West Saxon end in -an: as heortan=heart's; heortan=hearts. So they do in the Psalter. But in the Ritual the -n is omitted, and the word ends simply in -a or -e; as nome=of a name (West Saxon nam-an); hearta=hearts.

7. The Rushworth Gospels.—Place, Harewood in Wharfdale, Yorkshire. Time, according to Wanley, the end of the ninth century.

1. That the Ruthwell inscription gives us a sample of the so-called Northumbrian Anglo-Saxon, and that as it is spoken in Scotland, i.e., in Galloway. For the bearings of this see Part II., c. 3.

2. That the Rushworth Gospels take us as far south as the West Riding of Yorkshire.

3. That there are no specimens from any Cumberland, Westmoreland, or North Lancashire localities, these being, most probably, exclusively Celtic.

§ 693. The most general statements concerning this great section of the Anglo-Saxon, is that—

1. It prefers the slenderer and more vocalic to the broader and more diphthongal forms.

3. The forms without the prefix ge-, to those with them. Nevertheless the form ge-cenned (=natus) occurs in the first line of the extract from the Durham Gospels.

§ 694. The Old and Middle English MSS. from this quarter are numerous; falling into two classes:

1. Transcriptions with accommodation from works composed southwards. Here the characteristics of the dialect are not absolute. {551}

2. Northern copies of northern compositions. Here the characteristics of the dialect are at the maximum. Sir Tristram is one of the most important works of this class; and in the wider sense of the term Northumbrian, it is a matter of indifference on which side of the Border it was composed. See § 190.

Northumberland.—Northern frontier, East Scotland; the direction of the influence being from South to North, rather than from North to South, i. e., Berwickshire and the Lothians being Northumbrian and English, rather than Northumberland Scotch.

West frontier Celtic—the Cumberland and Westmoreland Britons having been encroached upon by the Northumbrians of Northumberland.

Present dialect.—Believed to be nearly uniform over the counties of Northumberland and Durham; but changing in character in North Yorkshire, and in Cumberland and Westmoreland.

The Anglo-Saxon immigration considered to have been Angle (so-called) rather than Saxon.

Danish admixture—Very great. Possibly, as far as the marks that it has left on the language, greater than in any other part of England.^[79]—See § 152.

Cumberland, Westmoreland, North Lancashire.—Anglo-Saxon elements introduced from portions of Northumbria rather than directly from the Continent.

Celtic language persistent until a comparatively late though undetermined period.

Northern frontier, West-Scotland—the direction of the influence being from Scotland to England, rather than vice versâ; Carlisle being more of a Scotch town than Berwick.

Topographical nomenclature characterized by the preponderance of compounds of -thwaite; as Braithwaite, &c.

North Lancashire, Westmoreland, and Cumberland, "exhibit many Anglian^[80] peculiarities, which may have been occasioned in some degree by the colonies in the south, planted in that district by William Rufus (Saxon Chronicle, A.D. 1092.) A comparison of Anderson's ballads with Burns's songs, will show how like Cumbrian is to Scottish, but how different. We believe that Weber is right in referring the romance of Sir Amadas to this district. The mixture of the Anglian forms gwo, gwon, bwons, boyd-word (in pure Northumbrian), gae, gane, banes, bod-worde, with the northern terms, tynt, kent, bathe, mare, and many others of the same class, could hardly have occurred in any other part of England."^[81]

Yorkshire, North and part of West Riding.—The Anglo-Saxon specimens of this area have been noticed in § 692.

1. The Craven.—This, in northern localities, "becomes slightly tinctured with Northumbrian."—Quart. Rev. ut supra.

2. The Cleveland.—With not only Northumbrian, but even Scotch characters. Quart. Rev. ut supra.

All these dialects, if rightly classified, belong to the Northumbrian division of the Angle branch of the Anglo-Saxon language; whilst, if the primâ facie view of their affiliation or descent, be the true one, they are the dialects of § 692, in their modern forms.

§ 696. The classification which gives this arrangement now draws a line of distinction at the river Ribble, in Lancashire, which separates South from North Lancashire; whilst in Yorkshire, the East Riding, and that part of the West which does not belong to the Wapentake of Claro, belong to the class which is supposed to exclude the previous and contain the following dialects:—

§ 697. South Lancashire and Cheshire.—Sub-varieties of {553}the same dialects, but not sub-varieties of the previous ones.

The plural form in -en is a marked character of this dialect—at least of the Lancashire portion.

These last two statements apply to all the forthcoming areas north of Essex. The latter is a simple historical fact; the former supposes an amount of difference between the Angle and the Saxon which has been assumed rather than proved; or, at any rate, which has never been defined accurately.

The elements of uncertainty thus developed, will be noticed in §§ 704-708. At present it is sufficient to say, that if the South Lancashire dialect has been separated from the north, on the score of its having been Mercian rather than Northumbrian, the principle of classification has been based upon political rather than philological grounds; and as such is exceptionable.

§ 698. Shropshire, Staffordshire, and West Derbyshire.—Supposing the South Lancashire and Cheshire to be the Mercian (which we must remember is a political term), the Shropshire, Staffordshire, and West Derbyshire are Mercian also; transitional, however, in character.

Here, also, both the a priori probabilities and the known facts make the Danish intermixture at its minimum.

§ 699. East Derbyshire and Nottinghamshire.—Here the language is considered to change from the mode of speech of which the South Lancashire is the type, to the mode of speech of which the Norfolk and Suffolk dialect is the type.

Danish elements may now be expected, Derbyshire being the most inland Danish area.

Hallamshire.—This means the parts about Sheffield {554}extended so as to include that portion of the West Riding of Yorkshire which stands over from § 696. Probably belonging to the same group with the South Lancashire.

East Riding of Yorkshire.—It is not safe to say more of this dialect than that its affinities are with the dialects spoken to the north rather than with those spoken to the south of it, i.e., that of—

Lincolnshire.—Frontier—On the Nottinghamshire and Leicestershire frontier, passing into the form of speech of those counties. Pretty definitely separated from that of Norfolk. Less so from that of North Cambridgeshire. Scarcely at all from that of Huntingdonshire, and North Northamptonshire.

Danish admixture.—The number of towns and villages ending in the characteristic Danish termination -by, at its maximum; particularly in the neighbourhood of Spilsby.

Language not Danish in proportion to the other signs of Scandinavian intermixture.

Specimens of the dialects in its older form—Havelok^[82] the Dane (?), Manning's Chronicle (supposing the MS. to have been transcribed in the county where the author was born).

Provincial peculiarities (i.e., deviations from the written language) nearly at the minimum.

Huntingdonshire, North Northamptonshire, and Rutland.—Anglo-Saxon period.—The latter part of the Saxon Chronicle was written at Peterboro. Probably, also, the poems of Helena and Andreas. Hence, this area is that of the old Mercian in its most typical form; whilst South Lancashire is that of the new—a practical instance of the inconvenience of applying political terms to philological subjects.

§ 700. Norfolk, Suffolk, and the fen part of Cambridgeshire.—Here the population is pre-eminently Angle. The political character East-Anglian rather than Mercian.

Specimens of the dialects in the Anglo-Saxon stage.—The Natale St. Edmundi, in Thorpe's Analecta Anglo-Saxonica.

§ 701. Leicestershire, Warwickshire, and South Northamptonshire.—Mercian (so-called) rather than West-Saxon (so-called).

Probably, approaching the written language of England more closely than is the case with the dialects spoken to the south of them.

Certainly, approaching the written language of England less closely than is the case with the dialect of Huntingdonshire, North Northamptonshire, and South Lincolnshire.

§ 702. These remarks have the following import. They bear upon the question of the origin of the written language of England.

Mr. Guest first diverted the attention of scholars from the consideration of the West Saxon of the chief Anglo-Saxon writers as the mother-dialect of the present English, to the Mercian; so turning their attention from the south to the centre of England.

The general principle that a central locality has the a priori likelihood in its favour, subtracts nothing from the value of his suggestion.

Neither does the fact of the nearest approach to the written language being found about the parts in question; since the doctrine to which the present writer commits himself, viz., that in the parts between Huntingdon and Stamford, the purest English is most generally spoken, is, neither universally recognised, nor yet part of Mr. Guest's argument.

Mr. Guest's arguments arose out of the evidence of the MSS. of the parts in question.

That the dialect most closely allied to the dialect (or dialects) out of which the present literary language of England is developed, is to be found either in Northamptonshire or the neighbouring counties is nearly certain. Mr. Guest looks for it on the western side of that county (Leicestershire); the present writer on the eastern (Huntingdonshire).

§ 703. It is now convenient to pass from the dialects of {556}the water-system of the Ouse, Nene, and Welland to those spoken along the lower course of the Thames.

These, to a certain extent, may be dealt with like those to the north of the Humber. Just as the latter were, in the first instance, and in the more general way, thrown into a single class (the Northumbrian), so may the dialects in question form the provisional centre of another separate class. For this we have no very convenient name. The dialects, however, which it contains agree in the following points.

1. These are considered to be derived from that variety of the Anglo-Saxon which is represented by the chief remains of the Anglo-Saxon literature, i.e., the so-called standard or classical language of Alfred, Ælfric, the present text of Cædmon, &c.

2. About half their present eastern area consists of the counties ending in -sex; viz., Sussex, Essex, and Middlesex.

3. Nearly the whole of their original area consisted in kingdoms (or sub-kingdoms) ending in -sex; viz., the districts just enumerated, and the kingdom of Wessex.

a.—Considered with reference to their literary history.—They are dialects whereof the literary development began early, but ceased at the time of the Norman Conquest, being superseded by that of the central dialects (Mercian so-called) of the island. The truth of this view depends on the truth of Mr. Guest's doctrine noticed in page 555. If true, it is by no means an isolated phænomenon. In Holland the present Dutch is the descendant of some dialect (or dialects) which was uncultivated in the earlier periods of the language; whereas the Old Frisian, which was then the written language, is now represented by a provincial dialect only.

"In speaking of the Anglo-Saxon language, scholars universally intend that particular form of speech in which all the principal monuments of our most ancient literature are composed, and which, with very slight variations, is found in Beowulf and Cædmon, in the Exeter and Vercelli Codices, in the translation of the Gospels and Homilies, and in the works {557}of Ælfred the Great. For all general purposes this nomenclature is sufficiently exact; and in this point of view, the prevalent dialect, which contains the greatest number of literary remains, may be fairly called the Anglo-Saxon language, of which all varying forms were dialects. It is, however, obvious that this is in fact an erroneous way of considering the subject; the utmost that can be asserted is, that Ælfred wrote his own language, viz., that which was current in Wessex; and that this, having partly through the devastations of heathen enemies in other parts of the island, partly through the preponderance of the West-Saxon power and extinction of the other royal families, become the language of the one supreme court, soon became that of literature and the pulpit also."—Kemble. Phil. Trans. No. 35.

b.—Considered in respect to their political relations.—Subject to the influence of the Wessex portion of the so-called Heptarchy, rather than to the Mercian,

c.—Considered ethnologically—Saxon rather than Angle. The exceptions that lie against this class will be noticed hereafter.

§ 704. Kent—Theoretically, Kent, is Jute rather than Saxon, and Saxon rather than Angle.

Predominance of local terms compounded of the word -hurst; as, Penshurst, Staplehurst, &c.

Frisian hypothesis.—The following facts and statements (taken along with those of §§ 15-20, and §§ 129-131), pre-eminently require criticism.

1. Hengest the supposed father of the Kentish kingdom is a Frisian hero—Kemble's Sächsische Stamtaffel.

2. The dialect of the Durham Gospels and Ritual contain a probably Frisian form.

3. "The country called by the Anglo-Saxons Northumberland, and which may loosely be said to have extended from the Humber to Edinburgh, and from the North Sea to the hills of Cumberland, was peopled by tribes of Angles. Such, at least, is the tradition reported by Beda, who adds that Kent was first settled by Jutes. Who these Jutes were is {558}not clearly ascertained, but from various circumstances it may be inferred that there was at least a considerable admixture of Frisians amongst them. Hengest, the supposed founder of the Kentish kingdom, is a Frisian hero, and Jutes, 'ëotenas,' is a usual name for the Frisians in Bëówulf. Beda, it is true, does not enumerate Frisians among the Teutonic races by which England was colonized, but this omission is repaired by the far more valuable evidence of Procopius, who, living at the time of some great invasion of Britain by the Germans, expressly numbers Frisians among the invaders. Now the Anglo-Saxon traditions themselves, however obscurely they may express it, point to a close connection between Kent and Northumberland: the latter country, according to these traditions, was colonized from Kent, and for a long time received its rulers or dukes from that kingdom. Without attaching to this legend more importance than it deserves, we may conclude that it asserts an original communion between the tribes that settled in the two countries; and consequently, if any Frisic influence is found to operate in the one, it will be necessary to inquire whether a similar action can be detected in the other. This will be of some moment hereafter, when we enter upon a more detailed examination of the dialect. The most important peculiarity in which the Durham Evangeles and Ritual differ from the Psalter is the form of the infinitive mood in verbs. This in the Durham books is, with exception of one verb, beán, esse, invariably formed in -a, not in -an, the usual form in all the other Anglo-Saxon dialects. Now this is also a peculiarity of the Frisic, and of the Old Norse, and is found in no other Germanic tongue; it is then an interesting inquiry whether the one or the other of these tongues is the origin of this peculiarity; whether, in short, it belongs to the old, the original Frisic form which prevailed in the fifth, sixth and seventh centuries, or whether it is owing to Norse influence, acting in the ninth and tenth, through the establishment of Danish invaders and a Danish dynasty in the countries north of the Humber."—Kemble. Phil. Trans. No. 35.

The details necessary for either the verification or the overthrow of the doctrine of a similarity of origin between {559}portions of the Northumbrian^[83] and portions of the Kentish population have yet to be worked out.

So have the differentiæ between the dialects of Kent, and the dialects of Sussex, Essex, Middlesex, and Wessex.

§ 705. Sussex.—The characteristics are involved in those of Kent—thus, if Kent be simply Saxon the two counties have the same ethnological relation; whilst if Kent be Frisian or Jute(?) Sussex may be either like or unlike.

Hampshire.—Theoretically, Saxon rather than Angle, and West Saxon (Wessex) rather than south, east, or Middle-Saxon.

Jute elements in either the Hants or Isle of Wight dialects, hitherto undiscovered. Probably, non-existent.

Present dialect certainly not the closest representative of the classical Anglo-Saxon, i. e., the so-called West Saxon.

Berkshire.—Present dialect, probably, the closest representative of the classical Anglo-Saxon.

Devonshire and West Somerset.—Present dialect strongly marked by the use of z for s (Zomerzet=Somerset).

Worcestershire.—The language of the Anglo-Saxon period is characterized by the exclusive, or nearly exclusive, use of s in the forms usse and usses for ure and ures. See Codex Diplomaticus, Nos. 95 and 97.

North Glostershire.—Politically, both North Gloster and Worcestershire are Mercian rather than West-Saxon.

What does this prove? Not that the West Saxon dialect {560}extended into Mercia, but that a political nomenclature is out of place in philology.

The Welsh frontier.—Herefordshire, &c.—Celtic elements. General character of the dialects, probably, that of the counties immediately to the east of them.

Essex.—Theoretically, Saxon rather than Angle. No such distinction, however, is indicated by the ascertained characteristic of the Essex dialects as opposed to the East Anglian, Suffolk, and the Mercian.

Hertfordshire.—I am not aware of any thing that distinguishes the South Hertfordshire form of speech from those of—

Middlesex.—Here, as far as there are any characteristics at all, they are those of Essex. The use of v for w, attributed (and partially due) to Londoners, occurs—not because there is any such thing as a London dialect, but because London is a town on the Essex side of Middlesex.

Surrey.—The name (Suð rige=southern kingdom) indicates an original political relation with the parts north rather than south of the Thames.

§ 706. Supposed East-Anglian and Saxon frontier.—For the area just noticed there are two lines of demarcation—one geographical, and one ethnological.

b. Ethnological.—The line which separates Middlesex and Essex (so-called Saxon localities) from Herts and Suffolk (so-called Angle localities).

Of these the first line involves an undeniable fact; the second a very doubtful one. No evidence has been adduced in favour of disconnecting Saxon Essex from Anglian Suffolk, nor yet for connecting it with Sussex and Wessex. The termination -sex is an undoubted fact; the difference between the Saxons and Angles which it is supposed to indicate is an assumption.

§ 707. The dialects of the remaining counties have, probably, the transitional characters, indicated by their geographical position.

Buckingham, Beds, Northampton.—These connect the two most convenient provisional centres of the so-called West-Saxon of Alfred, &c., and mother-dialect of the present written English, viz.: Wantage and Stamford (or Huntingdon); and in doing this they connect dialects which, although placed in separate classes (West-Saxon and Mercian), were, probably, more alike than many subdivisions of the same group.

To investigate the question as to the Mercian or West-Saxon origin of the present written English without previously stating whether the comparison be made between such extreme dialects as those of the New Forest, and the neighbourhood of Manchester, or such transitional ones as those of Windsor and Northampton is to reduce a real to a mere verbal discussion.

Warwickshire, Staffordshire.—From their central position, probably transitional to both the north and south, and the east and west groups.

§ 708. The exceptions suggested in §§ 703, 704, lie not only against the particular group called West-Saxon, but (as may have been anticipated) against all classifications which assume either—

1. A coincidence between the philological divisions of the Anglo-Saxon language, and the political division of the Anglo-Saxon territory.

§ 709. English dialects not in continuity with the mother-tongue.—Of these the most remarkable are those of—

1. Little England beyond Wales.—In Pembrokeshire, and a part of Glamorganshire, the language is English rather than Welsh. The following extracts from Higden have effected the belief that this is the result of a Flemish colony. "Sed {562}et Flandrenses, tempore Regis Henrici Primi in magna copia juxta Mailros ad orientalem Angliæ plagam habitationem pro tempore accipientes, septimam in insula gentem fecerunt: jubente tamen eodem rege, ad occidentalem Walliæ partem, apud Haverford, sunt translati. Sicque Britannia ... his ... nationibus habitatur in præsenti ... Flandrensibus in West Wallia."

A little below, however, we learn that these Flemings are distinguished by their origin only, and not by their language:—"Flandrenses vero qui in Occidua Walliæ incolunt, dimissa jam barbarie, Saxonice satis loquuntur."—Higden, edit. Gale, p. 210.

On the other hand, Mr. Guest has thrown a reasonable doubt upon this inference; suggesting the probability of its having been simply English. The following vocabulary collected by the Rev. J. Collins,^[84] in the little peninsula of Gower, confirms this view. It contains no exclusively Flemish elements.

§ 710. The baronies of Forth and Bargie in the County Wexford.—The barony of Forth "lies south of the city of Wexford, and is bounded by the sea to the south and east, and by the barony of Bargie to the west. It is said to have been colonized by the Welshmen who accompanied Strongbow in his invasion of Ireland; but by the term Welshmen, as here used, we must no doubt understand the English settlers of Gower and Pembroke. Vallancey published a specimen of their language. Some of the grammatical forms can hardly {564}fail to interest the English scholar, and we may venture more particularly to call his attention to the verbal ending th. In no other of our spoken dialects do we find the th still lingering as an inflection of the plural verb."

§ 711. Americanisms.—These, which may be studied in the excellent dictionary of J. R. Bartlett, are chiefly referable to five causes—

2. Influence of the languages introduced from Europe anterior to the predominance of English; viz.: French in Louisiana, Spanish in Florida, Swedish in Pennsylvania and Delaware, and Dutch in New York.

3. Influence, &c., subsequent to the predominance of the English; viz.: German in Pennsylvania, and Gaelic and Welsh generally.

4. Influence of the original difference of dialect between the different portions of the English population.

5. Influence of the preponderance of the Anglo-Saxon over the Anglo-Norman element in the American population in general.

§ 712. Extract.—In a sound and sagacious paper upon the Probable Future Position of the English Language,^[86] Mr. Watts, after comparing the previous predominance of the French language beyond the pale of France, with the present spread of the German beyond Germany, and after deciding in favour of the latter tongue, remarks that there is "The existence of another language whose claims are still more commanding. That language is our own. Two centuries ago the proud position that it now occupies was beyond the reach of anticipation. We all smile at the well-known boast of Waller in his lines on the death of Cromwell, but it was the loftiest that at the time the poet found it in his power to make:—

"'I care not,' said Milton, 'to be once named abroad, though perhaps I could attain to that, being content with these islands as my world.' A French Jesuit, Garnier, in 1678, laying down rules for the arrangement of a library, thought it superfluous to say anything of English books, because, as he observed, 'libri Anglicâ scripti linguâ vix mare transmittunt.' Swift, in the earlier part of the eighteenth century, in his 'Proposal for correcting, improving, and {567}ascertaining the English Tongue,' observed, 'the fame of our writers is usually confined to these two islands." Not quite a hundred years ago Dr. Johnson seems to have entertained far from a lofty idea of the legitimate aspirations of an English author. He quotes in a number of the 'Rambler' (No. 118, May 4th, 1751), from the address of Africanus as given by Cicero, in his Dream of Scipio:—'The territory which you inhabit is no more than a scanty island inclosed by a small body of water, to which you give the name of the great sea and the Atlantic Ocean. And even in this known and frequented continent what hope can you entertain that your renown will pass the stream of Ganges or the cliffs of Caucasus, or by whom will your name be uttered in the extremities of the north or south towards the rising or the setting sun? So narrow is the space to which your fame can be propagated, and even there how long will it remain?' 'I am not inclined,' remarks Johnson, 'to believe that they who among us pass their lives in the cultivation of knowledge or acquisition of power, have very anxiously inquired what opinions prevail on the further banks of the Ganges.... The hopes and fears of modern minds are content to range in a narrower compass; a single nation, and a few years have generally sufficient amplitude to fill our imagination.' What a singular comment on this passage is supplied by the fact that the dominions of England now stretch from the Ganges to the Indus, that the whole space of India is dotted with the regimental libraries of its European conquerors, and that Rasselas has been translated into Bengalee! A few years later the great historian of England had a much clearer perception of what was then in the womb of Fate. When Gibbon, as has been already mentioned, submitted to Hume, a specimen of his intended History of Switzerland, composed in French, he received a remarkable letter in reply: 'Why,' said Hume, 'do you compose in French and carry faggots into the wood, as Horace says with regard to Romans who wrote in Greek? I grant that you have a like motive to those Romans, and adopt a language much more generally diffused than your native tongue, but have you not remarked the fate {568}of those two ancient languages in following ages? The Latin, though then less celebrated and confined to more narrow limits, has in some measure outlived the Greek, and is now more generally understood by men of letters. Let the French therefore triumph in the present diffusion of their tongue. Our solid and increasing establishments in America, where we need less dread the inundation of barbarians, promise a superior stability and duration to the English language.'

"Every year that has since elapsed has added a superior degree of probability to the anticipations of Hume. At present the prospects of the English language are the most splendid that the world has ever seen. It is spreading in each of the quarters of the globe by fashion, by emigration, and by conquest. The increase of population alone in the two great states of Europe and America in which it is spoken, adds to the number of its speakers in every year that passes, a greater amount than the whole number of those who speak some of the literary languages of Europe, either Swedish, or Danish, or Dutch. It is calculated that, before the lapse of the present century, a time that so many now alive will live to witness, it will be the native and vernacular language of about one hundred and fifty millions of human beings.

"What will be the state of Christendom at the time that this vast preponderance of one language will be brought to bear on all its relations,—at the time when a leading nation in Europe and a gigantic nation in America make use of the same idiom,—when in Africa and Australasia the same language is in use by rising and influential communities, and the world is circled by the accents of Shakspeare and Milton? At that time such of the other languages of Europe as do not extend their empire beyond this quarter of the globe will be reduced to the same degree of insignificance in comparison with English, as the subordinate languages of modern Europe to those of the state they belong to,—the Welsh to the English, the Basque to the Spanish, the Finnish to the Russian. This predominance, we may flatter ourselves, will be a more signal blessing to literature than that of any other language could possibly be. The English is essentially a {569}medium language;—in the Teutonic family it stands midway between the Germanic and Scandinavian branches—it unites as no other language unites, the Romanic and the Teutonic stocks. This fits it admirably in many cases for translation. A German writer, Prince Pückler Muskau, has given it as his opinion that English is even better adapted than German to be the general interpreter of the literature of Europe. Another German writer, Jenisch, in his elaborate 'Comparison of Fourteen Ancient and Modern Languages of Europe,' which obtained a prize from the Berlin Academy in 1796, assigns the general palm of excellence to the English. In literary treasures what other language can claim the superiority? If Rivarol more than sixty years back thought the collective wealth of its literature able to dispute the pre-eminence with the French, the victory has certainly not departed from us in the time that has since elapsed,—the time of Wordsworth and Southey, of Rogers and Campbell, of Scott, of Moore, and of Byron.

"The prospect is so glorious that it seems an ungrateful task to interrupt its enjoyment by a shade of doubt: but as the English language has attained to this eminent station from small beginnings, may it not be advisable to consider whether obstacles are not in existence, which, equally small in their beginnings, have a probability of growing larger? The first consideration that presents itself is that English is not the only language firmly planted on the soil of America, the only one to which a glorious future is, in the probable course of things, assured.

"A sufficient importance has not always been attached to the fact, that in South America, and in a portion of the northern continent, the languages of the Peninsula are spoken by large and increasing populations. The Spanish language is undoubtedly of easier acquisition for the purposes of conversation than our own, from the harmony and clearness of its pronunciation; and it has the recommendation to the inhabitants of Southern Europe of greater affinity to their own languages and the Latin. Perhaps the extraordinary neglect which has been the portion of this language for the last {570}century and a half may soon give place to a juster measure of cultivation, and indeed the recent labours of Prescott and Ticknor seem to show that the dawn of that period has already broken. That the men of the North should acquire an easy and harmonious southern language seems in itself much more probable than that the men of the south should study a northern language, not only rugged in its pronunciation, but capricious in its orthography. The dominion of Spanish in America is, however, interrupted and narrowed by that of Portuguese, and to a singular degree by that of the native languages, some of which are possibly destined to be used for literary purposes in ages to come.

"At the time when Hume wrote his letter to Gibbon, the conquest of Canada had very recently been effected. The rivalry of the French and English in North America had been terminated by the most signal triumph of the English arms. Had measures been taken at that time to discourage the use of French and to introduce that of English, there can be little doubt that English would now be as much the language of Quebec and Montreal as it is of New York and the Delaware. Those measures were not taken. At this moment, when we are approaching a century from the battle of the Heights of Abraham, there is still a distinction of races in Canada, nourished by a distinction of language, and both appear likely to continue.

"Within the United States themselves, a very large body of the inhabitants have remained for generation after generation ignorant of the English language. The number is uncertain. According to Stricker, in his dissertation 'Die Verbreitung des deutschen Volkes über die Erde,' published in 1845, the population of German origin in the United States in 1844 was 4,886,632, out of a total of 18,980,650. This statement, though made in the most positive terms, is founded on an estimate only, and has been shown to be much exaggerated. Wappaus (in his 'Deutsche Auswanderung und Colonisation'), after a careful examination, arrives at the conclusion that the total cannot amount to a million and a half. Many of these are of course acquainted with both {571}languages—in several cases where amalgamation has taken place, the German language has died out and been replaced by the English,—but the number of communities where it is still prevalent is much larger than is generally supposed. In Pennsylvania, Ohio, and Missouri, to say nothing of other states, there are masses of population of German origin or descent, who are only acquainted with German. This tendency has of late years increased instead of declining. It has been a favourite project with recent German emigrants to form in America a state, in which the language should be German, and from the vast numbers in which they have crossed the Atlantic, there is nothing improbable in the supposition, that, by obtaining a majority in some one state, this object will be attained. In 1835 the legislature of Pennsylvania placed the German language in its legal rights on the same footing with the English.

"It may be asked if any damage will be done by this? The damage, it may be answered, will be twofold. The parties who are thus formed into an isolated community, with a language distinct from that of those around them, will be placed under the same disadvantages as the Welsh of our own day, who find themselves always as it were some inches shorter than their neighbours, and have to make an exertion to be on their level. Those of them who are only masters of one language are in a sort of prison; those who are masters of two, might, if English had been their original speech, have had their choice of the remaining languages of the world to exert the same degree of labour on, with a better prospect of advantage. In the case of Welsh, the language has many ties: even those who see most clearly the necessity of forsaking it, must lament the harsh necessity of abandoning to oblivion the ancient tongue of an ancient nation. But these associations and feelings could not be pleaded in favour of transferring the Welsh to Otaheite; and when these feelings are withdrawn, what valid reason will remain for the perpetuation of Welsh, or even, it may be said, of German?

"The injury done to the community itself is perhaps the greatest; but there is a damage done to the world in general. It will be a splendid and a novel experiment in modern society, if a single language becomes so predominant over all others as {572}to reduce them in comparison to the proportion of provincial dialects. To have this experiment fairly tried, is a great object. Every atom that is subtracted from the amount of the majority has its influence—it goes into the opposite scale. If the Germans succeed in establishing their language in the United States, other nations may follow. The Hungarian emigrants, who are now removing thither from the vengeance of Austria, may perpetuate their native Magyar, and America may in time present a surface as checkered as Europe, or in some parts, as Hungary itself, where the traveller often in passing from one village to another, finds himself in the domain of a different language. That this consummation may be averted must be the wish not only of every Englishman and of every Anglo-American, but of every sincere friend of the advancement of literature and civilization. Perhaps a few more years of inattention to the subject will allow the evil to make such progress that exertion to oppose it may come too late."

§ 713. Of the Gypsy language I need only say, that it is not only Indo-Germanic, but that it is Hindoo. Few words from it have mixed themselves with our standard (or even our provincial) dialects.

Thieves' language, or that dialect for which there is no name, but one from its own vocabulary, viz. Slang, is of greater value in philology than in commerce. It serves to show that in speech nothing is arbitrary. Its compound phrases are either periphrastic or metaphorical; its simple monosyllables are generally those of the current language in an older form. The thieves of London are conservators of Anglo-Saxonisms. In this dialect I know of no specimens earlier than the reign of Queen Elizabeth. In the dramatic literature of that age they are rife and common. The Roaring Girl, the Jolly Beggars, amongst the plays, and Deckar's Bellman amongst the tracts, preserve us a copious vocabulary, similar to what we have now, and similar to what it was in Gay's time. Of this the greater part is Saxon. Here and there appears a word of Latin origin, e.g., pannum, bread; cassons, cheese. Of the Gypsy language I have discovered no trace. {573}

§ 714. The Talkee-Talkee is a Lingua Franca based on the English, and spoken by the Negroes of Surinam.

It is Dutch rather than English; it shows, however, the latter language as an element of admixture.

§ 715. That the Anglo-Norman of England was, in the reign of Edward III., not the French of Paris (and most probably not the Franco-Norman of Normandy), we learn from the well-known quotation from Chaucer:—

§ 716. The concluding extract from the Testamenta Eboracensia, published by the Surtees' Society, is from the will of a gentleman in Yorkshire. To me it seems to impugn the assertion of Higden, that the Norman was spoken throughout England without a variety of pronunciation: "Mirandum videtur quomodo nativa propria Anglorum lingua, in unica insula coartata, pronunciatione ipsa fit tam diversa, cum tamen Normannica lingua, quæ adventicia est, univoca maneat penes cunctos."—Ed. Gale, p. 210.

§ 717. Relations of dialects (so-called) to languages (so-called).—"It is necessary clearly to conceive the nature and character of what we call dialects. The Doric, Æolic, and Ionic for example, in the language of grammarians, are dialects of the Greek: to what does this assertion amount? To this only, that among a people called the Greeks, some being Dorians spoke a language called Doric, some being Æolians spoke another language called Æolic, while a third class, Ionians, spoke a third language called, from them, Ionic. But though all these are termed dialects of the Greek, it does not follow that there was ever a Greek language of which these were variations, and which had any being apart from these. Dialects then are essentially languages: and the name dialect itself is but a convenient grammarian's phrase, invented as part of the machinery by which to carry on reasonings respecting languages. We learn the language which has the best and largest literature extant; and having done so, we treat all very nearly resembling languages as variations from what we have learnt. And that dialects are in truth several languages, will readily appear to any one who perceives the progressive development of the principle of separation in cognate tongues. The language of the Bavarian highlander or High Dutch, the language of the Hanoverian lowlander or Low Dutch, are German dialects: elevate, as it is called, regulate, and purify the one, and it assumes the {576}name and character of a language—it is German. Transplant the other to England, let nine centuries pass over it, and it becomes a language too, and a language of more importance than any which was ever yet spoken in the world, it has become English. Yet none but practised philologists can acknowledge the fact that the German and English languages are dialects of one Teutonic tongue."

§ 718. Relation of dialects to the older stages of the mother-tongue.—This has been noticed in § 691. The following extract from Mr. Kemble's paper just quoted, illustrates what he calls the spontaneity of dialects:—

The doctrine that languages become dead when they lose a certain power of evolving new forms out of previously existing ones, is incompatible with views to which the present writer has committed himself in the preface. If the views there exhibited be true the test of the vitality of a language, if such metaphors must be used, is the same as the test of vitality in material organisms, i.e., the power of fulfilling certain functions. Whether this is done by the evolution of new forms out of existing materials, or by the amalgamation (the particular power of the English language) of foreign terms is a mere difference of process.

§ 719. Effect of common physical conditions.—I again quote the same paper of Mr. Kemble's:—

If true, the bearings of this is important; since, if common physical conditions effect a common physiognomy of language, we may have a certain amount of resemblance without a corresponding amount of ethnological affinity.

PRAXIS.

The following extracts are given in the form of simple texts. They are meant, more especially, to be explained by masters to their classes; and as such were used by myself during the time that I was Professor of the English language and literature at University College. They are almost all taken from editions wherein either a translation or a full commentary can be found by reference. To have enlarged the present Appendix into a full Praxis, would have been to overstep the prescribed limits of the present work.

MŒSO-GOTHIC.

1. 2. Anastodeins aivaggeljons ïesuis xristaus sunaus guþs. sve

gameliþ ïst ïn esaï in praufetau. sai. ïk ïnsandja aggilu meinana

faura þus. saei gamanveiþ vig þeinana faura þus. stibna vopjandins

3. ïn auþidai. manveiþ vig fraujins. raihtos vaurkeiþ

4. staigos guþs unsaris. vas ïohannes daupjands ïn auþidai jah

5. merjands daupein ïdreigos du aflageinai fravaurhte. jah usïddjedun

du ïmma all ïudaialand jah ïairusaulymeis jah daupidai

vesun allai ïn ïaurdane awai fram ïmma andhaitandans fravaurhtim

6. seinaim. vasuþ-þan ïohannes gavasiþs taglam ulbandaus

jah gairda filleina bi hup seinana jah matida þramsteins

7. jah miliþ haiþivisk jah merida qiþands. qimiþ svinþoza mis sa

afar mis. þizei ïk ni ïm vairþs anahneivands andbindan skaudaraip

8. skohe is. aþþan ïk daupja ïzvis ïn vatin. ïþ ïs daupeiþ ïzvis

{582}

9. ïn ahmin veihamma. jah varþ ïn jainaim dagam. qam

ïesus fram nazaraiþ galeilaias jah daupiþs vas fram ïohanne ïn

10. ïaurdane. jah suns usgaggands us þamma vatin gasaw usluknans

11. himinans jah ahman sve ahak atgaggandan ana ïna. jah

stibna qam us himinam. þu ïs sunus meins sa liuba. ïn þuzei

12. vaila galeikaida. jah suns sai. ahma ïna ustauh ïn auþida.

13. jah vas in þizai auþidai dage fidvortiguns fraisans fram satanin

14. jah vas miþ diuzam jah aggileis andbahtidedun ïmma. ïp afar

þatei atgibans varþ ïohannes. qam ïesus ïn galeilaia merjands

15. aivaggeljon þiudangardjos guþs qiþands þatei usfullnoda þata

mel jah atnewida sik þiudangardi guþs. ïdreigoþ jah galaubeiþ

16. ïn aivaggeljon. jah warbonds faur marein galeilaias gasaw

seimonu jah andraian broþar ïs. þis seimonis. vairpandans

17. nati ïn marein. vesun auk fiskjans. jah qaþ ïm ïesus. hirjats

18. afar mis jah gatauja ïgqis vairþan nutans manne. jah suns

19. affetandans þo natja seina laistidedun afar ïmma. jah jainþro

ïnngaggands framis leitil gasaw ïakobu þana zaibaidaiaus jah

20. ïohanne broþar ïs jah þans ïn skipa manvjandans natja. jah

suns haihait ïns jah affetandans attan seinana zaibaidaiu ïn þamma

skipa miþ asnjam galiþun afar ïmma jah galiþun ïn kafarnaum.

21. jah suns sabbato daga galeiþands ïn synagogen laisida

22. ïns jah usfilmans vaurþun ana þizai laiseinai ïs. unte vas laisjands

23. ïns sve valdufni habands jah ni svasve þai bokarjos. jah

vas ïn þizai synagogen ïze manna ïn unhrainjamma ahmin jah

24. ufhropida qiþands. fralet. wa uns jah þus ïesu nazorenai.

qamt fraqistjan uns. kann þuk was þu ïs. sa veiha guþs.

25. jah andbait ïna ïesus qiþands. þahai jah usgagg ut us þamma.

26. ahma unhrainja. jah tahida ïna ahma sa unhrainja jah hropjands

27. stibnai mikilai usïddja us ïmma. jah afslauþnodedun

allai sildaleikjandans. svaei sokidedun miþ sis misso qiþandans.

wa sijai þata. wo so laiseino so niujo. ei miþ valdufnja jah

ahmam þaim unhrainjam anabiudiþ jah ufhausjand ïmma.

28. usïddja þan meriþa ïs suns and allans bisitands galeilaias.

29. jah suns us þizai synagogen usgaggandans qemun ïn garda seimonis

30. jah andraiïns miþ ïokobau jah ïohannem. ïþ svaihro

31. seimonis log ïn brinnon. jah suns qeþun ïmma bi ïja. jah

duatgaggands urraisida þo undgreipands handu ïzos. jah affailot

32. þo so brinno suns jah andbahtida ïm. andanahtja þan vaurþanamma.

þan gasaggq sauïl. berun du ïmma allans þans ubil

{583}

33. habandans jah unhulþons habandans. jah so baurgs alla garunnana

34. vas at daura. jah gahailida managans ubil habandans

missaleikaim sauhtim jah unhulþons managos usvarp jah ni

35. fralailot rodjan þos unhulþons. unte kunþedun ïna. jah air

uhtvon usstandans usïddja jah galaiþ ana auþjana staþ jah jainar

36. baþ. jah galaistans vaurþun ïmma seimon jah þai miþ

37. ïmma. jah bigitandans ïna qeþun du ïmma þatei allai þuk

38. sokjand. jah qaþ du ïm. gaggam du þaim bisunjane haimom

39. jah baurgim. ei jah jainar merjau. unte duþe qam. jah

vas merjands ïn synagogim ïze and alla galeilaian jah unholþons

40. usvairpands. jah qam at ïmma þrutsfill habands bidjands

ïna jah knivam knussjands jah qiþands du ïmma þatei. jabai

41. vileis. magt mik gahrainjan. ïþ ïesus ïnfeinands ufrakjands

handu seina attaitok ïmma jah qaþ ïmma. viljau. vairþ hrains.

42. jah biþe qaþ þata ïesus. suns þata þrutsfill affaiþ af ïmma jah

43. hrains varþ. jah gawotjands ïmma suns ussandida ïna jah qaþ

44. du ïmma. saiw ei mannhun ni qiþais vaiht ak gagg þuk silban

ataugjan gudjin jah atbair fram gahraineinai peinai. þatei

45. anabauþ moses du veitvodiþai ïm. ïþ ïs usgaggands dugann

merjan filu jah usqiþan þata vaurd. svasve ïs juþan ni mahta

andaugjo ïn baurg galeiþan ak uta ana auþjaim stadim vas.

jah ïddjedun du ïmma allaþro.

II.

OLD HIGH-GERMAN.

... sîn ta piqueme,

Das er towian scal,

Wanta sâr so sih dui sêla

In dem sind arhevit,

Ente si den lîhhamun

Likkan lâzzit;

So quimith ein heri

Fona himilzungalon;

Daz andar fona pehhe:

{584}

Dar pâgant siu umpi.

Sorgên mac diu sêla,

Unzi diu suona argêt,

Za wideremo herie,

Si gihalot werde.

Wanta ipu sia daz Satanazsses

Kisindi giwinnit,

Das leitet sia sâr

Dar iru leid wirdit,

In fiur enti in finstri,

Dazu ist reht virinlih ding.

Upi sia avar kihalont die,

Die dar fona himile quemant,

Enti si dero engilo eigan wirdit,

Die pringant sia sâr ûf in himilo rîhhi,

Darî est lîp âno tôd, lioht âno finstri,

Selida âno sorgun; dar nist neoman suih.

Denne der mar in pardîsu

Pû kiwinnit,

Hûs in himile,

Dar quimit imu hilfa kinuok

Pidiu ist durft mihhil allero manno welilihemo

Daz in es sîn muot kispane,

Daz er kotes willun

Kerno tuo,

Ente hella fuir

Harto wîsê,

Pehhes pina,

Dar piutit den Satanaz altist

Heizzan lauc. So mac huckan za diu,

Sorgên drâto

Der sih suntigen weiz.

Wê demo in vinstrî scal

Sîno virina stuen,

Prinnan in pehhe;

Daz ist rehto palwig ding—

Daz man den harêt ze gote,

Ente imo helfa ni quimit;

Wânit sih kinâda

{585}

Diu wênaga sêla

Ni ist in kihuctin

Himiliskin gote,

Wanta hiar in werolti

After ni werkôta.

So denne der mahtigo khuninc

Daz mahal kipannit

Dara scal queman

Chunno kilîhhaz

Denne ni kitar parno nohhein

Den pan furisizzan,

Dî allero manno welîh

Ze demo mahale sculi,

Der scal er, vora demo ricche,

Az rahhu stantan,

Pî daz er, in werolti,

Kiwerkota hapêta.

Daz hôrt ih rahhon

Dia werolt-rehtwîson,

Daz sculi der Antichristo

Mit Eliase pâgan.

Der warch ist kiwâfanit;

Denne wirdit untar in wîk arhapan;

Khensun sind so kreftic,

Diri kosa ist so mihhil.

Elias strîtît

Pî den ewigon lîp,

Wili den rehtkernon

Daz rîhhi kistarkan;

Pidiu scal imo halfan

Der himiles kiwaltit.

Der Anticristo stêt

Pî dem Altfiante

Stêt pî demo Satanase,

Der inan farsenkan scal;

Pidiu scal er in der wîcsteti

Wunt pivallan,

Enti in demo sinde

Sigalos werdan.

{586}

Doh wânit des vila gotmanno,

Daz Elias in demo wîge arwartit (werdit).

Sâr so daz Eliases pluot

In erda kitruifit,

So inprinnant die perga,

Poum ni kistentit

Einic in erdu,

Aha artruknênt,

Muor varsuilhet sih,

Suilizot lougui der himil

Mâno vallit,

Prinnit mittilagart,

Stein ni kistentit einik in erdu.

Verit denne stuatago in lant,

Verit mit diu viuriu

Viriho wîsôn,

Dar ni mai denne mâk andremo

Helfan vora dema Muspille.

Denne daz preita wasal

Allaz varprinnit,

Enti viur enti luft

Iz allaz arfurpit,

War ist denne diu marha,

Dar man dar eo mit sînem magon

(Diu marha ist farprunnan

Diu sêla stêt pidungan),

Ni weiz mit win puoze;

Sâr verit si za wîze.

Pidui ist dem manne so guot,

Denne er ze demo mahale quimit,

Daz er rahhono welihha

Rehto arteile;

Denne ni darf er sorgên,

Denne er ze deru suonu quimit.

Denne varant engila;

Uper dio marho,

Wecchant diota,

Wîssant ze dinge;

Denne scal manno gelîh

{587}

Fona deru moltu arsten;

Lôssan sih ar dero lêuuo vazzon

Scal imo avar sîn lîp piqueman,

Daz er sîn reht allaz

Kirahhon muozzi,

Enti imo after sînen tâtin

Arteilet werde.

Denne der gisizzit,

Der dar suonnan scal,

Enti arteillan scal,

Tôten enti quekken,

Denne stêt darumpi

Engilo menigi,

Quotero gomono girust so mihhil.

Dara quimit ze deru rightungu so vilo dia dar arstent,

So dar manno nohhein

Wiht pimîdan ni mak;

Dar scal denne hant sprehhan,

Houpit sagên,

Allero lido wehh

Unsi id den luzigun vinger.

Ni weiz der wênago man

Wielihhan urteil er habêt;

Denne er mit den miaton

Marrit daz rehta,

Daz der tiuval darpî

Kitarnit stentit;

Der habêt in ruovu

Rahhono welihha,

Daz der man er enti sîd

Upiles kifrumita,

Daz er iz allaz kisagêt,

Denne or ze deru suonu quimit.

* * * * * *

III.

ANGLO-SAXON.

IV.

VI.

From Warton's History of English Poetry, Ed. 1840. Vol. I. p. lxvii. Translated by R. Taylor.

VII.

VIII.

OLD SAXON.

3. Gidruouit bin an tilogon minro, in mistrot bin fan stimmon fiundes, in fan arbeide sundiges.

7. In ic quad "uuie sal geuan mi fetheron also duuon, in ic fliugon sal, in raston sal."

9. Ic sal beidan sin, thie behaldon mi deda fan luzzilheide geistis in fan geuuidere. {599}

10. Bescurgi, herro, te deile tunga iro, uuanda ic gesag unriht in fluoc an burgi.

11. An dag in naht umbefangan sal sia ouir mura ira, unreht in arbeit an mitdon iro in unreht.

13. Uuanda of fiunt flukit mi, is tholodit geuuisso; in of thie thie hatoda mi, ouir mi mikila thing spreke, ic burge mi so mohti geburran, fan imo.

16. Cum dot ouir sia, in nithir stigin an hellon libbinda. Uuanda arheide an selethe iro, an mitdon ini.

18. An auont in an morgan in an mitdondage tellon sal ic, in kundon; in he gehoron sal.

19. Irlosin sal an frithe sela mina fan then, thia ginacont mi, uuanda under managon he uuas mit mi.

21. Ne geuuisso ist ini uuihsil; in ne forchtedon got. Theneda hant sina an uuitherloni.

IX.

MODERN DUTCH OF HOLLAND.

2. Gelijk geschreven is in de Profeten: ziet, Ik zend mijnen Engel voor uw aangezigt, die uwen weg voor u heen bereiden zal.

3. De stem des roependen in de woestijn: bereidt den weg des Heeren, maakt zijne paden regt!

4. Johannes was doopende in de woestijn, en predikende den doop der bekeering tot vergeving der zonden.

5. En al het Joodsche land ging tot hem uit, en die van Jerûzalem; en werden allen van hem gedoopt in the rivier de Jordaan, belijdende hunne zonden.

6. En Johannes was gekleed met kemelshaar, en met eenen {600}lederen gordel om zijne lendenen, en at sprinkhannen en wilden honig.

7. En hij predikte, zeggende: na mij komt, die sterker is dan ik, wien ik niet waardig ben, nederbukkende, den riem zijner schoenen te ontbinden.

8. Ik heb ulieden wel gedoopt met water, maar hij zal u doopen met den Heiligen Geest.

9. En het geschiedde in diezelve dagen, dat Jezus kwam van Názareth, gelegen in Galiléa, en werd van Johannes gedoopt in de Jordaan.

10. En terstond, als hij uit het water opklom, zag bij de hemelen opengaan, en den Geest, gelijk eene duive, op hem nederdalen.

11. En er geschiedde eene stem nit de hemelen: gij zijt mijn geliefde Zoon, in denwelken Ik mijn welbehagen heb!

13. En hij was aldaar in de woestijn vertig dagen, verzocht van den Satan; en was bij de wilde gedierten; en de Engelen dienden hem.

14. En nadat Johannes overgeleverd was, kwam Jezus in Galiléa, predikende het Evangelie van het Koningrijk Gods,

15. En zeggende: de tijd is vervuld, en het Koningrijk Gods nabij gekomen; bekeert u, en gelooft het Evangelie.

16. En wandelende bij de Galilésche zee, zag hij Simon en Andréas, zijnen broeder, werpende het net in de zee (want zij waren visschers);

17 En Jezus zeide tot hen: volgt mij na, en ik zal maken, dat gij visschers der menschen zult worden.

19. En van daar een weinig voortgegaan zijnde, zag hij Jacobus, den zoon van Zebedéüs, en Johannes, zijnen broeder, en dezelve in het schip hunne netten vermakende.

20. En terstond riep hij hen; en zij, latende hunnen vader Zebedéüs in het schip, met de huurlingen, zijn hem nagevolgd.

21. En zij kwamen binnen Kapernaüm; en terstond op den Sabbatdag in de Synagoge gegaan zijnde, leerde hij.

22. En zij versloegen zich over zijne leer: want hij leerde hen, als magt hebbende, en niet als de Schriftgeleerden. {601}

23. En er was in hunne Synagoge een mensch, met eenen onreinen geest, en hij riep uit,

24. Zeggende: laat af, wat hebben wij met u te doen, gij Jezus Nazaréner! zijt gij gekomen, om ons to verderven? Ik ken u, wie gij zijt, namelijk de Heilige Gods.

26. En de onreine geest, hem scheurende, en roepende met eene groote stem, ging uit van hem.

27. En zij werden allen verbaasd, zoodat zij onder elkander vraagden, zeggende: wat is dit? wat nieuwe leer is deze, dat hij met magt ook den onreineen geesten gebiedt, en zig hem gehoorzaam zijn!

28. En zijn gerucht ging terstond uit, in het geheel omliggen land van Galiléa.

29. En van stonde aan uit de Synagoge gegaan zijnde, kwamen zij in het huis van Simon en Andréas, met Jacobus en Johannes.

30. En Simons vrouws moeder lag met de koorts; en terstond zeiden zij hem van haar.

31. En hij, tot haar gaande, vattede hare hand, en rigtte ze op; en terstond verliet haar de koorts, en zij diende henlieden.

32. Als het nu avond geworden was, toen de zon onderging, bragten zij tot hem allen, die kwalijk gesteld, en van den duivel bezeten waren.

34. En hij genas er velen, die door verscheidene ziekten kwalijk gesteld waren; en wierpe vele duivelen uit, en liet de duivelen niet toe te spreken, omdat zij hem kenden.

35. En des morgens vroeg, als het nog diep in den nacht was, opgestaan zijnde, ging hij uit, en ging henen in eene woeste plaats, en bad aldaar.

38. En hij zeide tot hen: laat ons in de bijliggende vlekken gaan, opdat ik ook daar predike: want daartoe ben ik uitgegaan.

39. En hij predikte in hunne Synagogen, door geheel Galiléa, en wierp de duivelen uit.

40. En tot hem kwam een melaatsche, biddende hem, en vallende {602}voor hem op de knieën, en tothem zeggende: indien gij wilt, gij kunt mij reinigen.

41. En Jezus, met barmhartigheid innerlijk bewogen zijnde, strekte de hand uit, en raakte hem aan, en zeide tot hem: ik wil, word gereinigd.

42. En als hij dit gezegd had, ging de melaatschheid terstond van hem, en hy werd gereinigd.

43. En als hij hem strengelijk verboden had, deed hij hem terstond van zich gaan;

44. En zeide tot hem: zie, dat gij niemand iets zegt; maar ga heen en vertoon u zelven den Priester, en offer voor uwe reiniging, hetgeen Mozes geboden heeft, hun tot eene getuigenis.

45. Maar hij vitgegaan zijnde, begon vele dingen te verkondigen, en dat woord te verbreiden, alzoo dat hij niet meer openbaar in de stad kon komen, maar was buiten in de woeste plaatsen; en zij kwamen tot hem van alle kanten.

OLD NORSE.

XI.

ICELANDIC.

Sva er sagt, at kringla heimsins, sú er mannfólkit byggir, er mjök vag-skorin: gánga höf stór úr útsjánum inn í jordina. Er þat kunnigt, at haf gengr af Njorvasundum, ok allt út til Jórsala-lands. Af hafinu gengr lángr hafsbotn til landnordrs, er heitir Svartahaf: sa skilr heims þridjúngana: heitir fyrin austan Asia, en fyrir vestan kalla sumir Evrópa, en sumir Enea. En nordan at Svartahafi gengr Sviþjod in mikla eda in kalda. Svíþjód ena miklu kalla sumir menn ecki minni enn Serkland hít mikla; sumir jafna henni vid Bláland hit mikla. Hinn neyrdri lutr Svíþjódar liggr óbygdr af frosti ok kulda, swa sem hinn sydri lutr Blálands er audr af sólarbruna. I Svíþjód eru stór hérut mörg: þar eru ok margskonar þjodir undarligar, ok margar túngur: þar eru risar, ok þar eru dvergar: þar eru ok blámenn; þar eru dýr ok drekar furdulega stórin. Ur Nordri frá fjöllum þeim, er fyrir utan eru bygd alla, fellr á um Svíþjód, sú er at rettu heitir Tanais; hún var fordum köllut Tanaqvísl edr Vanaquísl; hún kémur til sjávar inu i Svarta-haf. I Vanaqlvíslum var þa kallat Vanaland, edr Vanheimr; sú á skiir heimsþridjúngana; heitir fyrir austan Asia, en fyrir vestan Evrópa. {605}

Fyrir austan Tanaqvísl í Asía, var kallat Asa-land edr Asaheimr; en höfutborgina, er í var landinu, kölludu þeir Asgard. En í borginni var höfdíngi sá er Odinn var kalladr, þar var blótstadr mikill. Þar var þar sidr at 12 hofgodar vóru æztir; skyldu þeir ráda fyrir blótum ok dómum manna í milli; þat eru Diar kalladir edr drottnar: þeim skyldi þjónustu veita allr folk ok lotníng. Odinn var hermadr mikill ok mjök vidförull, ok eignadiz mörg riki: han var sva Sigrfæll, at í hvörri orustu feck hann gagn. Ok sva kom at hans menn trúdu því, at hann ætti heimilann sigr í hverri orustu. Þat var háttr hans ef ann sendi menn sína til orustu, edr adrar sendifarar, at hann lagdi adr hendur í höfut þeim, ok gaf þeim bjanak; trúdu þeir at þá mundi vel faraz. Sva var ok um hans menn, hvar sem þeir urdu í naudum staddir á sjá edr á landi, þá kölludu þeir á nafn hans, ok þóttuz jafnan fá af þvi fro; þar þottuz þeir ega allt traust er hann var. Hann fór opt sva lángt í brot, at hann dvaldiz í ferdinni mörg misseri.

XII.

Hákon konúngr hafði þá fylkt liði síno, ok segja menn at hann steypti af sèr brynjunni áðr orrostan tækist; Hákon konúngr valdi mjök menn með sèr í hirð at afli ok hreysti, svâ sem gert hafði Haraldr konúngr faðir hans; þar var þá með konúngi Þorálfr hinn sterki Skólmsson, ok gekk á aðra hlið konúngi; hann hafði hjálm ok skjöld, kesju ok sverð þat er kallat var Fetbreiðr; þat var mælt at þeir Hákon konúngr væri jafnsterkir; þessa getr Þórðr Sjáreksson í drápu þeirri er hann orti um Þórálf:

En er fylkíngar gengu saman, var fyrst skotit spjótum, þvínæst brugðu menn sverðum; Gerðist þá orostan óð ok mannskjæd; Hákon konúngr ok Þórálfr gengu þá fram um merkin ok hjöggu til beggja handa; Hákon konúngr var auðkendr, meiri enn aðrir menn, lýsti ok mjök af hjálmi hans er sólin shein á; þá varð vopnaburðr mikill at konúngi; tók þá Eyvindr Finnsson hatt einn, ok setti yfir hjálm konúngsins; þá kallaði hátt Eyvindr Skreyja: leynist hann nú Norðmanna konúngr, eðr hefir hann flýit, þvíat horfinn er nú gullhjálmrinn? Eyvindr ok Álfr bróðir hans gengu þá hart fram svâ sem óðir ok galnir væri, hjöggu til beggja handa; þa mælti Hákon konúngr hátt til Eyvindar: haltu svâ fram stefnunni ef þú vill finna hann Norðmanna konúng, Var þá skampt at bíða at Eyvindr kom þar, reiddi upp sverþit ok hjó til konúngs; Þórálfr skaut við honum Eyvindi skildinum, svâ at hann stakaði við; konúngr tók þá tveim höndum sverþit Kvernbít, ok hjó til Eyvindar, klauf hjálminn ok höfuðit alt í herþar niðr; í því bili drap Þórálfr Álf Askmann. Svâ segir Eyvindr Skáldaspillir:

Eptir fall þeirra bræðra gekk Hákon konúngr svâ hart fram at alt hravkk fur honum; sló þá felmt ok flótta á lið Eiríks sona, en Hákon konúngr var í öndverðri sinni fylkíng, ok fylgði fast flóttamönnum, ok hjó tídt ok hart; þá fló ör ein, er Fleinn er kallaðr, ok kom í hönd Hákoni konúngi uppi í músina firir neþan öxl, ok er þat margra manna sögn at skósveinn Gunnhildar, sá er Kispíngr er nefndr, ljóp fram í þysinn ok kallaði: gefi rúm konúngs bananum, ok skaut þá fleinnum til konúngs; en sumir segja at engi vissi hverr skaut; má þat ok vel vera, firir því at örvar ok spjót ok önnur skotvâpn flugu svâ þykkt sem drífa; fjöldi manns fèll þar af Eiríks sonum, en honúngarnir allir komust á skipin, ok rèro þegar undan, en Hákonar menn eptir þeim; svâ segir Þórðr Sjáreksson: {607}

XIII.

MODERN SWEDISH.

THE END.

NOTES

[1] Qu. the people of Euten, in Holstein.

[2] Zeus, p. 591.

[3] From Zeuss, v. v. Frisii, Chauci.

[4] The chief works in the two dialects or languages.

[5] Probably, for reasons, too long to enter upon, those of Grutungs and Tervings; this latter pointing to Thuringia, the present provincial dialect of which tract was stated, even by Michaelis, to be more like the Mœso-Gothic than any other dialect of Germany.

[6] Nearly analogous to Ostro-goth, and Visi-goth.

[7] The meaning of these terms is explained in § 90-92. The order of the cases and genders is from Rask. It is certainly more natural than the usual one.

[8] Compare with the Anglo-Saxon adjectives in § 85.

[9] Compare with the Anglo-Saxon adjectives in § 85.

[10] The syllables vulg-, and Belg-, are quite as much alike as Teuton-, and Deut-sch; yet how unreasonable it would be for an Englishman to argue that he was a descendant of the Belgæ because he spoke the Vulgar Tongue. Mutatis mutandis, however, this is the exact argument of nine out of ten of the German writers.

[11] Tacitus, De Mor. Germ. 40.

[12] And on the west of the Old Saxons is the mouth of the river Elbe and Friesland; and then north-west is the land which is called Angle and Sealand, and some part of the Danes.

[13] He sailed to the harbour which is called Hæðum, which stands betwixt the Wends (i.e. the Wagrian Slaves, for which see § 42) and Saxons, and Angle, and belongs to Denmark ... and two days before he came to Hæðum, there was on his starboard Gothland, and Sealand, and many islands. On that land lived Angles, before they hither to the land came.

[14] Zeus, in voc.

[15] Zeus, in voc.

[16] Zeus, in voc.

[17] See G. D. S. Vol. ii. II.

[18] Zeus, p. 492.

[19] As in Amherst and inherent.

[20] The meaning of the note of interrogation is explained in § 148.

[21] Edinburgh Philosophical Magazine.

[22] Natural History of Man.

[23] This list is taken from Smart's valuable and logical English Grammar.

[24] As in Shotover Hill, near Oxford.

[25] As in Jerusalem artichoke.

[26] A sort of silk.

[27] Ancient Cassio—"Othello."

[28] This class of words was pointed out to me by the very intelligent Reader of my first edition.

[29] V. Beknopte Historie van't Vaderland, i. 3, 4.

[30] Hist. Manch. b. i. c. 12.

[31] Dissertation of the Origin of the Scottish Language.—Jamieson's Etymological Dictionary, vol. i. p. 45, 46.

[32] Sir W. Betham's Gael and Cymry, c. iii.

[33] Scripturæ Linguæque Phœniciæ Monumenta, iv. 3.

[34] To say, for instance, Chemist for Chymist, or vice versâ; for I give no opinion as to the proper mode of spelling.

[35] Mr. Pitman, of Bath, is likely to add to his claims as an orthographist by being engaged in the attempt to determine, inductively, the orthoepy of a certain number of doubtful words. He collects the pronunciations of a large number of educated men, and takes that of the majority as the true one.

[36] Gesenius, p. 73.

[37] Write one letter twice.

[38] Rev. W. Harvey, author of Ecclesiæ Anglicanæ Vindex Catholicus.

[39] Murray's Grammar, vol. i. p. 79.

[40] Used as adverbs.

[41] Used as the plurals of he, she, and it.

[42] Different from ilk.

[43] Guest, ii. 192.

[44] Or call-s.

[45] Thou sangest, thou drankest, &c.—For a reason given in the sequel, these forms are less unexceptionable than sungest, drunkest, &c.

[46] Antiquated.

[47] As the present section is written with the single view of illustrating the subject, no mention has been made of the forms τυπῶ (typô), and ἔτυπον (etypon).

[48] Obsolete.

[49] Obsolete.

[50] Obsolete.

[51] The forms marked thus^[51] are either obsolete or provincial.

[52] Obsolete.

[53] Sounded wun.

[54] Obsolete.

[55] Præterite, or Perfect.

[56] Philological Museum, ii. p. 387.

[57] Vol. ii. p. 203.

[58] Found rarely; bist being the current form.—Deutsche Grammatik, i. 894.

[59] Over, under, after.—These, although derived forms, are not prepositions of derivation; since it is not by the affix -er that they are made prepositions. He went over, he went under, he went after—these sentences prove the forms to be as much adverbial as prepositional.

[60] In the first edition of this work I wrote, "Verbs substantive govern the nominative case." Upon this Mr. Connon, in his "System of English Grammar," remarks, "The idea of the nominative being governed is contrary to all received notions of grammar. I consider that the verb to be, in all its parts, acts merely as a connective, and can have no effect in governing anything." Of Mr. Connon's two reasons, the second is so sufficient that it ought to have stood alone. The true view of the so-called verb substantive is that it is no verb at all, but only the fraction of one. Hence, what I wrote was inaccurate. As to the question of the impropriety of considering nominative cases fit subjects for government it is a matter of definition.

[61] The paper On certain tenses attributed to the Greek verb has already been quoted. The author, however, of the doctrine on the use of shall and will, is not the author of the doctrine alluded to in the Chapter on the Tenses. There are, in the same number of the Philological Museum, two papers under one title: first, the text by a writer who signs himself T. F. B.; and, next, a comment, by the editor, signed J. C. H. (Julius Charles Hare). The usus ethicus of the future is due to Archdeacon Hare; the question being brought in incidentally and by way of illustration.

The subject of the original paper was the nature of the so-called second aorists, second futures, and preterite middles. These were held to be no separate tenses, but irregular forms of the same tense. Undoubtedly this has long been an opinion amongst scholars; and the writer of the comments is quite right in stating that it is no novelty to the learned world. I think, however, that in putting this forward as the chief point in the original paper, he does the author somewhat less than justice. His merit, in my eyes, seems to consist, not in showing that real forms of the aoristus secundus, futurum secundum, and præteritum medium were either rare or equivocal (this having been done before), but in illustrating his point from the English language; in showing that between double forms like συνελέχθην and συνελέγην, and double forms like hang and hanged, there was only a difference in degree (if there was that), not of kind; and, finally, in enouncing the very legitimate inference, that either we had two preterites, or that the Greeks had only one. "Now, if the circumstances of the Greek and English, in regard to these two tenses, are so precisely parallel, a simple and obvious inquiry arises, Which are in the right, the Greek grammarians or our own? For either ours must be wrong in not having fitted up for our verb the framework of a first and second preterite, teaching the pupil to say, 1st pret. I finded, 2d pret. I found; 1st pret. I glided, 2d pret. I glode: or the others must be so in teaching the learner to imagine two aorists for εὑρίσκω, as, aor. 1, εὕρησα, aor. 2, ἑῦρον; or for ἀκούω, aor. 1, ἤκουσα, aor. 2, ἤκοον."—p. 198.

The inference is, that of the two languages it is the English that is in the right. Now the following remarks, in the comment, upon this inference are a step in the wrong direction:—"The comparison, I grant, is perfectly just; but is it a just inference from that comparison, that we ought to alter the system of our Greek grammars, which has been drawn up at the cost of so much learning and thought, for the sake of adapting it to the system, if system it can be called, of our own grammars, which are seldom remarkable for anything else than their slovenliness, their ignorance, and their presumption? Is the higher to be brought down to the level of the baser? is Apollo to be drest out in a coat and waistcoat? Rather might it be deemed advisable to remodel the system of our own grammars."

This, whether right or wrong as a broad assertion, was, in the case in hand, irrelevant. No general superiority had been claimed for the English grammars. For all that had been stated in the original paper they might, as compared with the Greek and Latin, be wrong in ninety-nine cases out of a hundred. All that was claimed for them was that they were right in the present instance; just as for a clock that stands may be claimed the credit of being right once in every twelve hours. That the inference in favour of altering the system of the Greek grammars is illegitimate is most undeniably true; but then it is an inference of the critic's not of the author's. As the illustration in question has always seemed to me of great value,—although it may easily be less original than I imagine,—I have gone thus far towards putting it in a proper light.

Taking up the question where it is left by the two writers in question, we find that the difficulties of the so-called second tenses in Greek are met by reducing them to the same tense in different conjugations; and, according to the current views of grammarians, this is a point gained. Is it so really? Is it not rather the substitution of one difficulty for another? A second conjugation is a second mode of expressing the same idea, and a second tense is no more. Real criticism is as unwilling to multiply the one as the other. Furthermore, the tendency of English criticism is towards the very doctrines which the Greek grammarian wishes to get rid of. We have the difficulty of a second conjugation: but, on the other hand, instead of four past tenses (an imperfect, perfect, pluperfect, and aorist), we have only one (the aorist). Now, when we find that good reasons can be given for supposing that the strong preterite in the Gothic languages was once a reduplicate perfect, we are at liberty to suppose that what is now the same tense under two forms, was, originally, different tenses. Hence, in English, we avoid the difficulty of a second conjugation by the very same process which we eschew in Greek; viz., the assumption of a second tense. But this we can do, as we have a tense to spare.

Will any process reconcile this conflict of difficulties? I submit to scholars the following hypotheses:—

1. That the true second future in Greek (i.e., the future of verbs with a liquid as a characteristic) is a variety of the present, formed by accentuating the last syllable; just as I beát you=I will beat you.

2. That this accent effects a change on the quantity and nature of the vowel of the penultimate.

3. That the second aorist is an imperfect formed from this secondary present.

4. That the so-called perfect middle is a similar perfect active.

[62] Transactions of Philological Society. No. 90, Jan. 25, 1850.

[63] Notwithstanding the extent to which a relative may take the appearance of conjunction, there is always one unequivocal method of deciding its true nature. The relative is always a part of the second proposition. A conjunction is no part of either.

[64] Unless another view be taken of the construction, and it be argued that ἔδωκε is, etymologically speaking, no aorist but a perfect. In form, it is almost as much one tense as another. If it wants the reduplication of the perfect, it has the perfect characteristic κ, to the exclusion of the aorist σ; and thus far the evidence is equal. The persons, however, are more aorist than perfect. For one of Mathiæ's aorists (μεθῆκε) a still better case might be made, showing it to be, even in etymology, more perfect than aorist.

Κτείνει με χρυσοῦ, τὸν ταλαίπωρον, χάριν

Ξένος πατρῷος, καὶ κτανὼν ἐς οἶδμ' ἁλὸς

Μεθῆχ', ἵν' αὐτὸς χρυσὸν ἐν δόμοις ἔχῃ.

Κεῖμαι δ' ἐπ' ἀκταῖς.

Eur. Hec.

[65] It is almost unnecessary to state that the sentence quoted in the text is really a beautiful couplet of Withers's poetry transposed. It was advisable to do this, for the sake of guarding against the effect of the rhyme. To have written,

What care I how fair she is

If she be not fair to me?

would have made the grammar seem worse than it really was, by disappointing the reader of a rhyme. On the other hand, to have written,

What care I how fair she were,

If she were not kind as fair?

would have made the grammar seem better than it really was, by supplying one.

[66] In the first edition of the present work I inaccurately stated that neither should take a plural and either a singular verb; adding that "in predicating something concerning neither you nor I, a negative assertion is made concerning both. In predicating something concerning either you or I, a positive assertion is made concerning one of two." This Mr. Connon (p. 129) has truly stated to be at variance with the principles laid down by me elsewhere.

[67] Latin Prose Composition, p. 123.

[68] Quoted from Guest's English Rhythms.

[69] To the definition in the text, words like old and bold form no exception. At the first view it may be objected that in words like old there is no part preceding the vowel. Compared, however, with bold, the negation of that part constitutes a difference. The same applies to words like go and lo, where the negation of a part following the vowel is a point of identity. Furthermore, I may observe, that the word part is used in the singular number. The assertion is not that every individual sound preceding the vowel must be different, but that the aggregate of them must be so. Hence, pray and bray (where the r is common to both forms) form as true a rhyme as bray and play, where all the sounds preceding a, differ.

[70] For prosópa. The Greek has been transliterated into English for the sake of showing the effect of the accents more conveniently.

[71] For the sake of showing the extent to which the accentual element must be recognised in the classical metres, I reprint the following paper On the Doctrine of the Cæsura in the Greek senarius, from the Transactions of the Philological Society, June 23, 1843:—

"In respect to the cæsura of the Greek tragic senarius, the rules, as laid down by Porson in the Supplement to his Preface to the Hecuba, and as recognized, more or less, by the English school of critics, seem capable of a more general expression, and, at the same time, liable to certain limitations in regard to fact. This becomes apparent when we investigate the principle that serves as the foundation to these rules; in other words, when we exhibit the rationale, or doctrine, of the cæsura in question. At this we can arrive by taking cognizance of a second element of metre beyond that of quantity.

"It is assumed that the element in metre which goes, in works of different writers, under the name of ictus metricus, or of arsis, is the same as accent, in the sense of that word in English. It is this that constitutes the difference between words like týrant and resúme, or súrvey and survéy; or (to take more convenient examples) between the word Aúgust, used as the name of a month, and augúst, used as an adjective. Without inquiring how far this coincides with the accent and accentuation of the classical grammarians, it may be stated that, in the forthcoming pages, arsis, ictus metricus, and accent (in the English sense of the word), mean one and the same thing. With this view of the arsis, or ictus, we may ask how far, in each particular foot of the senarius, it coincides with the quantity.

First Foot.—In the first place of a tragic senarius it is a matter of indifference whether the arsis fall on the first or second syllable; that is, it is a matter of indifference whether the foot be sounded as týrant or as resúme, as Aúgust or as augúst. In the following lines the words ἡκω, παλαι, εἰπερ, τινας, may be pronounced either as ἡ´κω, πα´λαι, ει´περ, τι´νας, or as ἡκω´, παλαι´, ειπερ´, τινα´ς, without any detriment to the character of the line wherein they occur.

Ἡ´κω νεκρων κευθμωνα και σκοτου πυλας.

Πα´λαι κυνηγετουντα και μετρουμενον.

Ει´περ δίκαιος εσθ' εμος τα πατροθεν.

Τι´νας ποθ' ἑδρας τασδε μοι θοαζετε.

or,

Ἡκω´ νεκρων κευθμωνα και σκοτου πυλας.

Παλαι´ κυνηγετουντα και μετρουμενον.

Ειπερ´ δικαιος εσθ' εμος τα πατροθεν.

Τινα´ς ποθ' ἑδρας τασδε μοι θοαζετε.

Second Foot.—In the second place, it is also a matter of indifference whether the foot be sounded as Aúgust or as augúst. In the first of the four lines quoted above we may say either νε´κρων or νεκρω´ν, without violating the rhythm of the verse.

Third Foot.—In this part of the senarius it is no longer a matter of indifference whether the foot be sounded as Aúgust or as augúst; that is, it is no longer a matter of indifference whether the arsis and the quantity coincide. In the circumstance that the last syllable of the third foot must be accented (in the English sense of the word), taken along with a second fact, soon about to be exhibited, lies the doctrine of the penthimimer and hepthimimer cæsuras.

The proof of the coincidence between the arsis and the quantity in the third foot is derived partly from a posteriori, partly from a priori evidence.

1. In the Supplices of Æschylus, the Persæ, and the Bacchæ, three dramas where licences in regard to metre are pre-eminently common, the number of lines wherein the sixth syllable (i. e., the last half of the third foot) is without an arsis, is at the highest sixteen, at the lowest five; whilst in the remainder of the extant dramas the proportion is undoubtedly smaller.

2. In all lines where the sixth syllable is destitute of ictus, the iambic character is violated: as

Θρηκην περασα´ντες μογις πολλῳ πονῳ.

Δυοιν γεροντοι´ν δε στρατηγειται φυγη.

These are facts which may be verified either by referring to the tragedians, or by constructing senarii like the lines last quoted. The only difficulty that occurs arises in determining, in a dead language like the Greek, the absence or presence of the arsis. In this matter the writer had satisfied himself of the truth of the two following propositions:—1. That the accentuation of the grammarians denotes some modification of pronunciation other than that which constitutes the difference between Aúgust and augúst; since, if it were not so, the word ἄγγελον would be sounded like mérrily, and the word ἀγγέλων like disáble; which is improbable, 2. That the arsis lies upon radical rather than inflectional syllables, and out of two inflectional syllables upon the first rather than the second; as βλε´π-ω, βλεψ-α´σ-α, not βλεπ-ω´, βλεψ-ασ-α´. The evidence upon these points is derived from the structure of language in general. The onus probandi lies with the author who presumes an arsis (accent in the English sense) on a non-radical syllable. Doubts, however, as to the pronunciation of certain words, leave the precise number of lines violating the rule given above undetermined. It is considered sufficient to show that wherever they occur the iambic character is violated.

The circumstance, however, of the last half of the third foot requiring an arsis, brings us only half way towards the doctrine of the cæsura. With this must be combined a second fact, arising out of the constitution of the Greek language in respect to its accent. In accordance with the views just exhibited, the author conceives that no Greek word has an arsis upon the last syllable, except in the three following cases:—

1. Monosyllables, not enclitic; as σφω´ν, πα´ς, χθω´ν, δμω´ς, νω´ν, νυ´ν, &c.

2. Circumflex futures; as νεμω´, τεμω´, &c.

3. Words abbreviated by apocope; in which case the penultimate is converted into a final syllable; δω´μ', φειδεσ´θ', κεντει´τ', εγω´γ', &c.

Now the fact of a syllable with an arsis being, in Greek, rarely final, taken along with that of the sixth syllable requiring, in the senarius, an arsis, gives as a matter of necessity, the circumstance that, in the Greek drama, the sixth syllable shall occur anywhere rather than at the end of a word; and this is only another way of saying, that, in a tragic senarius, the syllable in question shall generally be followed by other syllables in the same word. All this the author considers as so truly a matter of necessity, that the objection to his view of the Greek cæsura must lie either against his idea of the nature of the accents, or nowhere; since, that being admitted, the rest follows of course.

As the sixth syllable must not be final, it must be followed in the same word by one syllable, or by more than one.

1. The sixth syllable followed by one syllable in the same word.—This is only another name for the seventh syllable occurring at the end of a word, and it gives at once the hepthimimer cæsura: as

Ἡκω νεκρων κευθμω´να και σκοτου πυλας.

Ἱκτηριοις κλαδοι´σιν εξεστεμμενοι.

Ὁμου τε παιανω´ν τε και στεναγματων.

2. The sixth syllables followed by two (or more) syllables in the same word. This is only another name for the eighth (or some syllable after the eighth) syllable occurring at the end of a word; as

Οδμη βροτειων αἱ´ματων με προσγελα.

Λαμπρους δυναστας εμ´πρεποντας αιθερι.

Now this arrangement of syllables, taken by itself, gives anything rather than a hepthimimer; so that if it was at this point that our investigations terminated, little would be done towards the evolution of the rationale of the cæsura. It will appear, however, that in those cases where the circumstance of the sixth syllable being followed by two others in the same words, causes the eighth (or some syllable after the eighth) to be final, either a penthimimer cæsura, or an equivalent, will, with but few exceptions, be the result. This we may prove by taking the eighth syllable and counting back from it. What follows this syllable is immaterial: it is the number of syllables in the same word that precedes it that demands attention.

1. The eighth syllable preceded in the same word by nothing.—This is equivalent to the seventh syllable at the end of the preceding word: a state of things which, as noticed above, gives the hepthimimer cæsura.

Ανηριθμον γελα´σμα παμ|μητορ δε γη.

2. The eighth syllable preceded in the same word by one syllable.—This is equivalent to the sixth syllable at the end of the word preceding; a state of things which, as noticed above, rarely occurs. When however it does occur, one of the three conditions under which a final syllable can take an arsis must accompany it. Each of these conditions requires notice.

α). With a non-enclitic mono-syllable the result is a penthimimer cæsura; since the syllable preceding a monosyllable is necessarily final.

Ἡκω σεβι´ζων σο´ν Κλυ´ται|μνηστρα κρατος.

No remark has been made by critics upon lines constructed in this manner, since the cæsura is a penthimimer, and consequently their rules are undisturbed.

β). With poly-syllabic circumflex futures constituting the third foot, there would be a violation of the current rules respecting the cæsura. Notwithstanding this, if the views of the present paper be true, there would be no violation of the iambic character of the senarius. Against such a line as

Καγω το σον νεμω´ ποθει|νον αυλιον

there is no argument a priori on the score of the iambic character being violated; whilst in respect to objections derived from evidence a posteriori, there is sufficient reason for such lines being rare.

γ). With poly-syllables abbreviated by apocope, we have the state of things which the metrists have recognised under the name of quasi-cæsura; as

Κεντειτε μη φειδε´σθ' εγω | 'τεκον Παριν.

3. The eighth syllable preceded in the same word by two syllables.—This is equivalent to the fifth syllable occurring at the end of the word preceding: a state of things which gives the penthimimer cæsura; as

Οδμη βροτειων αἱ´ ματῶν | με προσγελα.

Λαμπρους δυναστας εμ´πρεπον τας αιθερι.

Αψυχον εικω προ´σγελῳσα σωματος.

4. The eighth syllable preceded in the same word by three or more than three syllables.—This is equivalent to the fourth (or some syllable preceding the fourth) syllable occurring at the end of the word preceding; a state of things which would include the third and fourth feet in one and the same word. This concurrence is denounced in the Supplement to the Preface to the Hecuba; where, however, the rule, as in the case of the quasi-cæsura, from being based upon merely empirical evidence, requires limitation. In lines like

Και ταλλα πολλ' επει´κασαι | δικαιον ην,

or (an imaginary example),

Τοις σοισιν ασπιδη´στροφοις|ιν ανδρασι,

there is no violation of the iambic character, and consequently no reason against similar lines having been written; although from the average proportion of Greek words like επεικασαι and ασπιδηστροφοισιν, there is every reason for their being rare.

After the details just given, the recapitulation is brief.

1. It was essential to the character of the senarius that the sixth syllable, or latter half of the third foot, should have an arsis, ictus metricus, or accent in the English sense. To this condition of the iambic rhythm the Greek tragedians, either consciously or unconsciously, adhered.

2. It was the character of the Greek language to admit an arsis on the last syllable of a word only under circumstances comparatively rare.

3. These two facts, taken together, caused the sixth syllable of a line to be anywhere rather than at the end of a word.

4. If followed by a single syllable in the same word, the result was a hepthimimer cæsura.

5. If followed by more syllables than one, some syllable in an earlier part of the line ended the word preceding, and so caused either a penthimimer, a quasi-cæsura, or the occurrence of the third and fourth foot in the same word.

6. As these two last-mentioned circumstances were rare, the general phænomenon presented in the Greek senarius was the occurrence of either the penthimimer or hepthimimer.

7. Respecting these two sorts of cæsura, the rules, instead of being exhibited in detail, may be replaced by the simple assertion that there should be an arsis on the sixth syllable. From this the rest follows.

8. Respecting the non-occurrence of the third and fourth feet in the same word, the assertion may be withdrawn entirely.

9. Respecting the quasi-cæsura, the rules, if not altogether withdrawn, may be extended to the admission of the last syllable of circumflex futures (or to any other polysyllables with an equal claim to be considered accented on the last syllable) in the latter half of the third foot.

[72] Sceolon, aron, and a few similar words, are no real exceptions, being in structure not present tenses but preterites.

[73] Quarterly Review, No. clxiv.

[74] Quarterly Review, No. clxiv.

[75] From the Quarterly Review, No. cx.

[76] From the Quarterly Review, No. cx.

[77] Apparently a lapsus calami for spede.

[78] J. M. Kemble, "On Anglo-Saxon Runes," Archæologia, vol. xxviii.

[79] But not of Great Britain. The Lowland Scotch is, probably, more Danish than any South-British dialect.

[80] In opposition to the typical Northumbrian.

[81] Quarterly Review—ut supra.

[82] The subject is a Lincolnshire tradition; the language, also, is pre-eminently Danish. On the other hand, the modern Lincolnshire dialect is by no means evidently descended from it.

[83] For some few details see Phil. Trans., No. 36.

[84] Transactions of the Philological Society. No. 93.

[85] Philological Transactions. No. 84.

[86] Transactions of the Philological Society, No. 92.

[87] Quarterly Review, vol. xliii.

FOR THE SAXON OF THE CONTINENT.	FOR THE SAXON OF ENGLAND.
1. Continental Saxon.	Insular Saxon.
2. German Saxon.	English Saxon.
3. Westphalian Saxon.	Hanoverian Saxon.
4. South-Saxon.	North Saxon.
5. Cheruscan Saxon.	Angle Saxon.
6. Saxon of the Heliand.^[4]	Saxon of Beowulf.^[4]

		Saxon.	Icelandic.
		Neuter.	Neuter.
Sing.	Nom.	Eáge (an eye).	Auga (an eye).
	Acc.	Eáge	Auga.
	Dat.	Eágan	Auga.
	Gen.	Eágan	Auga.
{39} Plur.	Nom.	Eágan	Augu.
	Acc.	Eágan	Augu.
	Dat.	Eágan	Augum.
	Gen.	Eágan	Augna.
		Masculine.	Masculine.
Sing.	Nom.	Nama (a name).	Bogi (a bow).
	Acc.	Naman	Boga.
	Dat.	Naman	Boga.
	Gen.	Naman	Boga.
Plur.	Nom.	Naman	Bogar.
	Acc.	Naman	Boga.
	Dat.	Namum	Bogum.
	Gen.	Namena	Boga.
		Feminine.	Feminine.
Sing.	Nom.	Tunge (a tongue).	Túnga (a tongue).
	Acc.	Tungan	Túngu.
	Dat.	Tungan	Túngu.
	Gen.	Tungan	Túngu.
Plur.	Nom.	Tungan	Túngur.
	Acc.	Tungan	Túngur.
	Dat.	Tungum	Túngum.
	Gen.	Tungena	Túngna.

		Saxon.	Icelandic.
		Neuter.	Neuter.
Sing.	Nom.	Leáf (a leaf).	Skip (a ship).
	Acc.	Leáf	Skip.
	Dat.	Leáfe	Skipi.
	Gen.	Leáfes	Skips.
Plur.	Nom.	Leáf	Skip.
	Acc.	Leáf	Skip.
	Dat.	Leáfum	Skipum.
	Gen.	Leáfa	Skipa.
		Masculine.	Masculine.
Sing.	Nom.	Smið (a smith).	Konungr (a king).
	Acc.	Smið	Konung.
	Dat.	Smiðe	Konungi.
	Gen.	Smiðes	Konungs.
{40} Plur.	Nom.	Smiðas	Konungar.
	Acc.	Smiðas	Konunga.
	Dat.	Smiðum	Konungum.
	Gen.	Smiða	Konunga.
		Feminine.	Feminine.
Sing.	Nom.	Sprǽc (a speech).	Brúðr (a bride).
	Acc.	Sprǽce	Brúi.
	Dat.	Sprǽce	Brúði.
	Gen.	Sprǽce	Brúðar.
Plur.	Nom.	Sprǽca	Brúðir.
	Acc.	Sprǽca	Brúðir.
	Dat.	Sprǽcum	Brúðum.
	Gen.	Sprǽca	Brúða.

			a.
		Neut.	Masc.	Fem.
Sing.	Nom.	Hitt	Hinn	Hin.
	Acc.	Hitt	Hinn	Hina.
	Dat.	Hinu	Hinum	Hinni.
	Gen.	Hins	Hins	Hinnar.
Plur.	Nom.	Hin	Hinir	Hinar.
	Acc.	Hin	Hina	Hinar.
	Dat.	Hinum	Hinum	Hinum.
	Gen.	Hinna	Hinna	Hinna.
			b.
Sing.	Nom.	—it	—inn	—in.
	Acc.	—it	—inn	—ina (-na).
{41}	Dat.	—nu	—num	—inni (-nni).
	Gen.	—ins	—ins	—innar (-nnar).
Plur.	Nom.	—in	—nir	—nar.
	Acc.	—in	—na	—nar.
	Dat.	—num	—num	—num.
	Gen.	—nna	—nna	—nna.

		Neut.	Masc.	Fem.
Sing.	Nom.	Augat	Boginn	Túngan.
	Acc.	Augat	Boginn	Túnguna.
	Dat.	Auganu	Boganum	Túngunni.
	Gen.	Augans	Bogans	Túngunnar.
Plur.	Nom.	Augun	Bogarnir	Túngurnar.
	Acc.	Augun	Bogana	Túngurnar.
	Dat.	Augunum	Bogunum	Túngunum.
	Gen.	Augnanna	Boganna	Túngnanna.

Saxon.				Icelandic.
Definite.^[7]				Definite.^[7]
Singular.				Singular.
	Neut.	Masc.	Fem.		Neut.	Masc.	Fem.
Nom.	Góde	Góda	Góde.	Nom.	Haga	Hagi	Haga.
Acc.	Góde	Gódan	Gódan.	Acc.	Haga	Haga	Högu.
Abl.	Gódan	Gódan	Gódan.	Abl.	Haga	Haga	Högu.
Dat.	Gódan	Gódan	Gódan.	Dat.	Haga	Haga	Högu.
Gen.	Gódan	Gódan	Gódan.	Gen.	Haga	Haga	Högu.

Plural.				Högu is the Plural form for all the Cases and all the Genders.
Nom.	Gódan	Gódan	Gódan.
Acc.	Gódan	Gódan	Gódan.
Abl.	Gódum	Gódum	Gódum.
Dat.	Gódum	Gódum	Gódum.
Gen.	Gódena	Gódena	Gódena.

Indefinite.				Indefinite.
Singular.				Singular.
	Neut.	Masc.	Fem.		Neut.	Masc.	Fem.
Nom.	Gód	Gód	Gód.	Nom.	Hagt	Hagr	Hög.
Acc.	Gód	Gódne	Góde.	Acc.	Hagt	Hagan	Hög.
Abl.	Góde	Góde	Gódre.	Abl.	Högu	Högum	Hagri.
Dat.	Gódum	Gódum	Gódre.	Dat.	Högu	Högum	Hagri.
Gen.	Gódes	Gódes	Gódre.	Gen.	Hags	Hags	Hagrar.

Plural.				Plural.
Nom.	Góde	Góde	Góde.	Nom.	Hög	Hagir	Hagar.
Acc.	Góde	Góde	Góde.	Acc.	Hög	Haga	Hagar.
Abl.	Gódum	Gódum	Gódum.	Abl.	Högum	Högum	Högum.
Dat.	Gódum	Gódum	Gódum.	Dat.	Högum	Högum	Högum.
Gen.	Gódra	Gódra	Gódra.	Gen.	Hagra	Hagra	Hagra.

Sing.	Nom.	Sitt	Sinn	Sín.
	Acc.	Sitt	Sinn	Sína.
	Dat.	Sínu	Sínum	Sinni.
	Gen.	Sins	Sins	Sinnar.
Plur.	Nom.	Sín	Sínir	Sínar.
	Acc.	Sín	Sína	Sínar.
	Dat.	Sínum	Sínum	Sínum.
	Gen.	Sinna	Sinna	Sinna.

	Saxon.	Icelandic.
1.	Án	Eitt, einn, ein.
2.	Twá	Tvö, tveir.
3.	Þreó	Þrju, þrir.
4.	Feower	Fjögur, fjórir.
5.	Fíf	Fimm.
6.	Six	Sex.
7.	Seofon	Sjö.
8.	Eahta	Átta.
9.	Nigon	Niu.
10.	Tyn	Tiu.

		Saxon.	Icelandic.
Pres.	Sing.	1. Bærne	Brenni.
		2. Bærnst	Brennir.
		3. Bærnð	Brennir.
	Plur.	1. Bærnað	Brennum.
		2. Bærnað	Brennið.
		3. Bærnað	Brenna.

The Auxiliary Verb.
	Saxon.	Icelandic.
Indicative. Present.
Sing.	1. Eom (I am)	Em.
	2. Eart.	Ert.
	3. Is.	Er.
{45} Plur.	1. Synd (Syndon)	Erum.
	2. Synd (Syndon)	Eruð.
	3. Synd (Syndon)	Eru.
Indicative. Past.
Sing.	1. Wǽs	Var.
	2. Wǽre	Vart.
	3. Wǽs	Var.
Plur.	1. Wǽron	Vorum.
	2. Wǽron	Voru.
	3. Wǽron	Voru.
Subjunctive. Present.
Sing.	1. Sý	Sé.
	2. Sý	Sér.
	3. Sý	Sé.
Plur.	1. Sýn	Séum.
	2. Sýn	Seuð.
	3. Sýn	Séu.
Subjunctive. Past.
Sing.	1. Wǽre	Væri.
	2. Wǽre	Værir.
	3. Wǽre	Væri.
Plur.	1. Wǽron	Værum.
	2. Wǽron	Væru.
	3. Wǽron	Væruð.
Infinitive.
	Wesan	Vera.
Participle.
	Wesende	Verandi.

Neuter.
	Singular.	Plural.
Nom.	Áugô (an eye)	Áugôna.
Acc.	Áugô	Áugôna.
Dat.	Áugin	Áugam.
Gen.	Áugins	Áugônê.
Masculine.
Nom.	Manna (a man)	Mannans.
Acc.	Mannan	Mannans.
Dat.	Mannin	Mannam.
Gen.	Mannins	Mannanê.
Feminine.
Nom.	Tuggô (a tongue)	Tuggôns.
Acc.	Tuggôn	Tuggôns.
Dat.	Tuggôn	Tuggôm.
Gen.	Tuggôns	Tuggônô.

Singular.
Nom.	Blindata	Blinds	Blinda.
Acc.	Blindata	Blindana	Blinda.
Dat.	Blindamma	Blindamma	Blindái.
Gen.	Blindis	Blindis	Blindáizôs.
Plural.
Nom.	Blinda	Blindái	Blindôs.
Acc.	Blinda	Blindans	Blindôs.
Dat.	Blindáim	Blindáim	Blindáim.
Gen.	Blindáizê	Blindáizê	Blindáizô.

	Indicative.			Subjunctive.
	M.G.	A.S.		M.G.	A.S.
	Present.			Present.
Sing.	1. Sôk-ja	Lufie.	Sing.	1. Sôkjáu	Lufige.
	2. Sôk-eis	Lufast.		2. Sôkjáis
	3. Sôk-eiþ	Lufað.		3. Sôkjái
Plur.	1. Sôk-jam	Lufiað.	Plur.	1. Sôkjáima	Lufion.
	2. Sôk-eiþ	Lufiað.		2. Sôkjáiþ
	3. Sôk-jand	Lufiað.		3. Sôkjáina
	Præt.		Præt.
Sing.	1. Sôkida	Lufode.	Sing.	1. Sôkidêdjáu	Lufode.
	2. Sôkides	Lufodest.		2. Sôkidêdeis
	3. Sôkida	Lufode.		3. Sôkidêdi
Plur.	1. Sôkidêdum	Lufodon.	Plur.	1. Sôkidêdeima	Lufodon.
	2. Sôkidêduþ	Lufodon.		2. Sôkidêdeiþ
	3. Sôkidêdun	Lufodon.		3. Sôkidêdeina

Indicative. Pres.		Subjunctive. Pres.
Sing.	Plur.	Sing.	Plur.
1. Im (I am)	Sijum.	1. Sijáu	Sijáima.
2. Is	Sijuþ.	2. Sijáis	Sijáiþ.
3. Ist	Sind.	3. Sijái	Sijáina.
Præt.		Præt.
1. Vas	Vêsum.	1. Vêsjáu	Vêseima.
2. Vast	Vêsuþ.	2. Vêseis	Vêseiþ.
3. Vas	Vêsun.	3. Vêsei	Vêseina.
Inf. Visan and Sijan—(to be).
Part. Visands—(being).

(a).
Substantives ending in a vowel.
		Neuter.	Masculine.	Feminine.
Sing.	Nom.	Áre (an ear)	Campa (a champion)	Tunge (a tongue).
	Acc.	Áre	Campa	Tunga.
	Dat.	Ára	Campa	Tunga.
	Gen.	Ára	Campa	Tunga.
Plur.	Nom.	Ára	Campa	Tunga.
	Acc.	Ára	Campa	Tunga.
	Dat.	Áron	Campon	Tungon.
	Gen.	Árona	Campona	Tungona.

Neuter.
	Singular.	Plural.
Nom.	Vaúrd (a word)	Vaúrda.
Acc.	Vaúrd	Vaúrda.
Dat.	Vaúrda	Vaúrdam.
Gen.	Vaúrdis	Vaúrdê.
Masculine.
Nom.	Fisks (a fish)	Fiskôs.
Acc.	Fisk	Fiskans.
Dat.	Fiska	Fiskam.
Gen.	Fiskis	Fiskê.
Feminine.
Nom.	Brûþs (a bride)	Brûþeis.
Acc.	Brûþ	Brûþins.
Dat.	Brûþai	Brûþim.
Gen.	Brûþais	Brûþê.

Singular.
	Neuter.	Masculine.	Feminine.
Nom.	Blindô	Blinda	Blindô.
Acc.	Blindô	Blindan	Blindôn.
Dat.	Blindin	Blindin	Blindôn.
Gen.	Blindins	Blindins	Blindôns.
Plural.
Nom.	Blindôna	Blindans	Blindôns.
Acc.	Blindôna	Blindans	Blindôns.
Dat.	Blindam	Blindam	Blindôm.
Gen.	Blindônê	Blindanê	Blindônô.

(b).
Substantives ending in a consonant.
		Neuter.	Feminine.
Sing.	Nom.	Skip (a ship)	Hond (a hand).
	Acc.	Skip	Hond.
{52}	Dat.	Skipe	Hond.
	Gen.	Skipis	Honde.
Plur.	Nom.	Skipu	Honda.
	Acc.	Skipu	Honda.
	Dat.	Skipum	Hondum (-on).
	Gen.	Skipa	Honda.

		Saxon.	Frisian.

Sing.	Nom.	Sunu (a son)	Sunu.
	Acc.	Sunu	Sunu.
	Dat.	Suna	Suna.
	Gen.	Suna	Suna.
Plur.	Nom.	Suna	Suna.
	Acc.	Suna	Suna.
	Dat.	Sunum	Sunum.
	Gen.	Sunena	(Sunena).

(c).
Indefinite Declension of Adjectives.
		Neuter.	Masculine.	Feminine.
Sing.	Nom.	Gód	Gód	Gód.
	Acc.	Gód	Gódene	Góde.
	Dat.	Góda (-um)	Góda (-um).	Gódere.
	Gen.	Gódes	Gódes	Gódere.
Plur.	Nom.	Góde	Góde	Góde.
	Acc.	Góde	Góde	Góde.
	Dat.	Gódum (-a)	Gódum (-a)	Gódum (-a).
	Gen.	Gódera	Gódera	Gódera.
(d).
Definite.
		Neuter.	Masculine.	Feminine.
Sing.	Nom.	Góde	Góda	Góde.
	Acc.	Góde	Góda	Góda.
{53}	Dat.	Góda	Góda	Góda.
	Gen.	Góda	Góda	Góda.
Plur.	Nom.	Góda	Góda	Góda.
	Acc.	Góda	Góda	Góda.
	Dat.	Góda (-on)	Góda (-on)	Góda (-on).
	Gen.	Góda (-ona)	Góda (-ona)	Góda (-ona).

(e).
The Persons of the Present Tense.
Indicative Mood.
Sing.	1. Berne	I burn.
	2. Bernst	Thou burnest.
	3. Bernth	He burns.
Plur.	1. Bernath	We burn.
	2. Bernath	Ye burn.
	3. Bernath	They burn.

Frisian.	Anglo Saxon.	English.
Áge	Eáge	Eye.
Háved	Heáfod	Head.
Kind	Cild	Child.
Erva	Eafora	Heir.
Drochten	Drihten	Lord.
Nacht	Niht	Night.
Réd	Rǽd	Council (Rede).
Déde	Dǽd	Deed.
Nose	Nasu	Nose.
Éin	Ágen	Own.
Kápie	Ceapige	I buy (Chapman).
Dua	Don	To do.
Slá	Sleán	Slay.
Gunga	Gangan	Go (Gang).

Câr, a kinsman.
1. form,	Câr agos, a near kinsman.
2.	Ei gâr, his kinsman.
3.	Ei châr, her kinsman.
4.	Vy nghâr, my kinsman.
Tâd, a father.
1. form,	Tâd y plentyn, the child's father.
2.	Ei dâd, his father.
3.	Ei thâd, her father.
4.	Vy nhâd, my father.
Pen, a head.
1. form,	Pen gwr, the head of a man.
2.	Ei ben, his head.
3.	Ei phen, her head.
4.	Vy mhen, my head.
Gwâs, a servant.
1. form,	Gwâs fydhlon, a faithful servant.
2.	Ei wâs, his servant.
{79} 3.	Vy ngwas, my servant.
Duw, a god.
1. form,	Duw trugarog, a merciful god.
2.	Ei dhuw, his god.
3.	Vy nuw, my god.
Bara, bread.
1. form,	Bara cann, white bread.
2.	Ei vara, his bread.
3.	Vy mara, my bread.
Lhaw, a hand.
1. form,	Lhaw wenn, a white hand.
2.	Ei law, his hand.
Mam, a mother.
1. form,	Mam dirion, a tender mother.
2.	Ei vam, his mother.
Rhwyd, a net.
1. form,	Rhwyd lawn, a full net.
2.	Ei rwyd, his net.
From the Erse.
Súil, an eye.
1. form,	Súil.
2.	A húil, his eye.
Sláinte, health.
2. form,	Do hláinte, your health.

Sharp.	Flat.
p to ^[19]m=h.	b to m.
t to ^[19]n=h.	d to n.
k to ng=h.	g to ng.

Welsh.			Irish.
Singular.		Plural.	Singular.		Plural.
Aber	= a conflux;	ebyr.	Ball	= a spot;	baill.
Barð	= a bard;	beirð.	Cnoc	= a hill;	cnoic.
Bràn	= a crow;	brain.	Poll	= a pit;	poil.
Fon	= a staff;	fyn	Fonn	= a tune;	foinn.
Maen	= a stone;	mein.	Crann	= a tree;	crainn.
Gûr	= a man;	gûyr.	Fear	= a man;	fir.
&c.			&c.

Bull	= tarw, Welsh; tarbh, Gaelic.
Three	= tri, Welsh; tre, Gaelic.
Crane	= garan, Welsh; corr, Gaelic.

Welsh.	English.
Basgawd	Basket.
Berfa	Barrow.
Botwm	Button.
Bràn	Bran.
Clwt	Clout, Rag.
Crochan	Crock, Crockery.
Crog	Crook, Hook.
Cwch	Cock, in Cock-boat.
Cwysed	Gusset.
Cyl, Cyln	Kiln (Kill, provinc.).
Dantaeth	Dainty.
Darn	Darn.
Deentur	Tenter, in Tenterhook.
Fflaim	Fleam, Cattle-lancet.
Fflaw	Flaw.
Ffynnell (air-hole)	Funnel.
Gefyn (fetter)	Gyve.
Greidell	Grid, in Gridiron.
Grual	Gruel.
Gwald (hem, border)	Welt.
Gwiced (little door)	Wicket.
Gwn	Gown.
Gwyfr	Wire.
Masg (stitch in netting)	Mesh.
Mattog	Mattock.
Mop	Mop.
Rhail (fence)	Rail.
Rhasg (slice)	Rasher.
Rhuwch	Rug.
Sawduriaw	Solder.
Syth (glue)	Size.
Tacl	Tackle.

Provincial.	Common Dialect.	Norse.
Braid	Resemble	Bråas, Swed.
Eldin	Firing	Eld, Dan.
Force	Waterfall	Fors, D. Swed.
Gar	Make	Göra, Swed.
Gill	Ravine	Gil, Iceland.
Greet	Weep	Grata, Iceland.
Ket	Carrion	Kiöd=Flesh, Dan.
Lait	Seek	Lede, Dan.
Lathe	Barn	Lade, Dan.
Lile	Little	Lille, Dan.

Sing.			Plur.
Apex	sounded	apec-s	apices
Appendix	—	appendic-s	appendices
Calix	—	calic-s	calices
Cicatrix	—	cicatric-s	cicatrices
Helix	—	helic-s	helices
Index	—	indec-s	indices
Radix	—	radic-s	radices
Vertex	—	vertec-s	vertices
Vortex	—	vortec-s	vortices.

THE

ENGLISH LANGUAGE.

ROBERT GORDON LATHAM, M.D., F.R.S.,

THIRD EDITION,

REVISED AND GREATLY ENLARGED.

LONDON:

TAYLOR, WALTON, AND MABERLY,

1850.

THE REV. WILLIAM BUTCHER, M.A.,

THE AUTHOR.

PREFACE

SECOND EDITION.

CONTENTS.

AN INTRODUCTION

THE ENGLISH LANGUAGE

PART I.

CHAPTER I.

CHAPTER II.

CHAPTER III.

CHAPTER IV.

CHAPTER V.

CHAPTER VI.

CHAPTER VII.

CHAPTER VIII.

PART II.

CHAPTER I.

CHAPTER II.

CHAPTER III.

CHAPTER IV.

PART III.

CHAPTER I.

CHAPTER II.

CHAPTER III.

CHAPTER IV.

CHAPTER V.

CHAPTER VI.

CHAPTER VII.

CHAPTER VIII.

CHAPTER IX.

CHAPTER X.

PART IV.

CHAPTER I.

CHAPTER II.

CHAPTER III.

CHAPTER IV.

CHAPTER V.

CHAPTER VI.

CHAPTER VII.

CHAPTER VIII.

CHAPTER IX.

CHAPTER X.

CHAPTER XI.

CHAPTER XII.

CHAPTER XIII.

CHAPTER XIV.

CHAPTER XV.

CHAPTER XVI.

CHAPTER XVII.

CHAPTER XVIII.

CHAPTER XIX.

CHAPTER XX.

CHAPTER XXI.

CHAPTER XXII.

CHAPTER XXIII.

CHAPTER XXIV.

CHAPTER XXV.

CHAPTER XXVI.

CHAPTER XXVII.

CHAPTER XXVIII.

CHAPTER XXIX.

CHAPTER XXX.

CHAPTER XXXI.

CHAPTER XXXII.

CHAPTER XXXIII.

CHAPTER XXXIV.

CHAPTER XXXV.

CHAPTER XXXVI.

CHAPTER XXXVII.

CHAPTER XXXVIII.

PART V.

Original root.	Plur.	Sing.
Apsid-	apsides	apsis
Cantharid-	cantharides	cantharis
Chrysalid-	chrysalides	chrysalis
Ephemerid-	ephemerides	ephemeris
Tripod-	tripodes	tripos.

	Participle.	Substantive.
Wegferende=Wayfaring.			Wegferend=Wayfarer.
Sing.	Nom.	Wegferende	Wegferend.
	Acc.	Wegferendne	Wegferend.
	Abl.	Wegferende	Wegferende.
	Dat.	Wegferendum	Wegferende.
	Gen.	Wegferendes	Wegferendes.
Plur.	Nom.	Wegferende	Wegferendas.
	Dat.	Wegferendum	Wegferendum.
	Gen.	Wegferendra	Wegferenda.

	1st Person.			2nd Person.
Nom.	Wit	We two.	Nom.	Git	Ye two.
Acc.	Unc	Us two.	Acc.	Inc	You two.
Gen.	Uncer	Of us two.	Gen.	Incer	Of you two.

Indicative Mood.
Pres. Sing. 1.	Lufige.	Plur. 1.	Lufiað.
2.	Lufast.	2.
3.	Lufað.	3.
Subjunctive Mood.
Pres. Sing. 1.	Lufige.	Plur. 1.	Lufion.
2.		2.
3.		3.

Lene.		Aspirate.		Sharp.		Flat.
Sharp.	Flat.	Sharp.	Flat.	Lene.	Aspirate.	Lene.	Aspirate.
p	b	f	v	p	f	b	v
t	d	þ	ð	t	þ	d	ð
k	g	κ	γ	k	κ	g	γ
s	z	σ	ζ	s	σ	z	ζ

Broad.	Slender.
Independent.	Independent.	Dependent.
a, in father	a, in fate	a, in fat.
	e fermé, long	e fermé, short.
e, in meine, Germ.		e, in bed.
	ee, in feet	i, in pit.
	ü, of the German, long	the same, short.
	oo, in book	ou, in could.
	o chiuso	the same, short.
aw, in bawl	o, in note	o, in note.

Liquids.	Mutes.				Semivowels.
	Lene.		Aspirate.
	Sharp.	Flat.	Sharp.	Flat.
m	p	b	f	v	w
n	t	d	þ	ð
l	k	g	κ	γ	y
r	s	z	σ	ζ

n	is doubled in	unnatural, innate, oneness.
l	—	soulless, civil-list, palely.
k	—	book-case.
t	—	seaport-town.

We find	p and h	in the words	haphazard, upholder.
—	b and h	—	abhorrent, cub-hunting.
—	f and h	—	knife-handle, offhand.
—	v and h	—	stave-head.
—	d and h	—	adhesive, childhood.
—	t and h	—	nuthook.
—	th and h	—	withhold.
—	k and h	—	inkhorn, bakehouse.
—	g and h	—	gig-horse.
—	s and h	—	race-horse, falsehood.
—	z and h	—	exhibit, exhort.
—	r and h	—	perhaps.
—	l and h	—	well-head, foolhardy.
—	m and h	—	Amherst.
—	n and h	—	unhinge, inherent, unhappy.

a	to	ĕ,	as	man, men.
a	to	oo,	as	stand, stood.
a	to	u,	as	dare, durst.
a	to	ē,	as	was, were.
ea	to	o,	as	speak, spoken.
ea=ĕ	to	ea=ē,	as	breath, breathe.
ee	to	ĕ,	as	deep, depth.
ea	to	o,	as	bear, bore.
i	to	a,	as	spin, span.
i	to	u,	as	spin, spun.
i=ei	to	o,	as	smite, smote.
i=ei	to	ĭ,	as	smite, smitten.
i	to	a,	as	give, gave.
i=ei	to	a,	as	rise, raise.
ĭ	to	e,	as	sit, set.
ow	to	ew,	as	blow, blew.
o	to	e,	as	strong, strength.
oo	to	ee,	as	tooth, teeth.
o	to	i,	as	top, tip.
o	to	e,	as	old, elder; tell, told.
ŏ	to	e,	as	brother, brethren.
ō=oo	to	i,	as	do, did.
o=oo	to	o=ŭ,	as	do, done.
oo	to	o,	as	choose, chose.