The English Language

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.

Thirteenth Class.

§ 376. Contains the single word choose, in the præterite chose; in Anglo-Saxon, ceóse, ceás.

CHAPTER XXIV.

THE WEAK TENSES.

§ 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 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.

§ 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 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. 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 addition of the vowel takes place only in verbs of the first class.

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 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, 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.

The -s in durst has still to be satisfactorily accounted for.

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 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 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.

In respect to the præterite of do=facio, difficulties present themselves.

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, 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 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.

ON CONJUGATION.

§ 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, 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 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 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.

I. Many strong verbs become weak; whilst no weak verb ever becomes strong.

II. All the strong verbs are of Saxon origin. None are classical.

III. The greater number of them are strong throughout the Gothic tongues.

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.

DEFECTIVENESS AND IRREGULARITY.

§ 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.

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 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 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 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.