Chapter XII · ON THE INFLUENCE OF CHANGE IN FUNCTION ON ANALOGICAL FORMATION.

The careful consideration of such a form as I breakfasted will lead us to understand another phase in the life history of our words, and in the development of their syntactical combinations. It is well known that the word (to) breakfast is really a compound of the verb to break and the noun fast (ieiunium). Accordingly, we find, about the year 1400 A.D., ‘Ete and be merry, why breke yee nowt your fast;’ in 1653, Izaak Walton wrote, ‘My purpose is to be at Hodsden before I break my fast;’ and as late as 1808, Scott writes in his Marmion, ‘and knight and squire had broke their fast.’110 In these and similar cases, the words have retained their full and original meaning of ‘to put an end to fasting by eating;’ and the natural apprehension of this compound when employed as a noun was in the sense of the meal whereby this process is effected after the night’s fasting, i.e. the first meal taken in the day. When once the verb had thus acquired the meaning of ‘to take the first meal in the day,’ and was next applied even in cases where so little food had been taken before that meal as to be hardly worth considering a ‘meal,’ the meaning of ‘breaking the fast’ had been effaced by the new sense of eating the first IMPORTANT meal of the day. The change of meaning, coupled with the change in function, disconnected the compound from the linguistic groups to which it had hitherto belonged, and so it came about that, after the analogy of other verbs formed from nouns, to breakfast was conjugated as a weak verb. Thus, in 1679, Everard writes, After breakfasting peaceably; and about a century later, the word is used transitively in the sense of ‘to entertain at breakfast,’ e.g., They will breakfast you, or I was breakfasted.111

This and all the following examples to be discussed in this chapter illustrate the point that, in the unconscious grouping of our words into material and modal groups, it is mainly the function of the word which causes such grouping; and that a change of function, entailing, as it does, a change in the grouping, will often expose the word which has thus altered its meaning to the influence of analogy with other groups, though as long as it preserved its original meaning it stood quite apart from them. No doubt, however, similarity of form conduces also sometimes to this end. The group to which the word once belonged will then follow its own path of development, while the detached member will go on its new way.

We have a similar instance in vouchsafe: The king vouches it SAUE (Robert of Brunne, early in fourteenth century), where we should now say: The king vouchsafes. The verb to backbite is most probably a derivative from the compound nouns back-biting (of which the earliest instance dates from 1175) and backbitter (which is found as early as 1230); while in the Early English Psalter (A.D. 1300) the past tense is still formed bac-bate. Gower (1393) already formed the past participle back bited.112 Again, the noun browbeating (from ‘to beat one’s brows,’ i.e. ‘to lower the brows,’ ‘to frown’), found as early as 1581,113 became, from a compound noun, a simple one with the meaning of scolding or teasing; and gave rise to a verb to browbeat, of which the earliest known instance dates from 1603. It is, however, doubtful whether this verb has hitherto been definitely separated from the group to which etymologically speaking it belongs. The past participle brow-beat (1803; Jane Porter, Thaddeus) occurs, it is true, but the more usual form is as yet browbeaten.

The most ordinary results of this process are, of course, all the numerous formations from nouns that have been pressed into the service of verbs; as, I box, He boxed; (to) dust, (to) soap, (to) dog, etc., etc.: in the case of all which, the change of function must have preceded all forms due to analogy with the groups into which the word entered solely in consequence of that change. So, again, as long as a word has an adjectival function, and even when it is used substantively, but retains its original attributive meaning, it is, in English, not declined: as the POOR men; the POOR ye have with you always; the BLUE hats. When, then, only certain individuals belonging to the class designated by the adjective have to be indicated—and not, as in the case of the poor,—all the individuals possessing the quality of poverty,—we resort to the addition of the word ones: as, I do not like those green hats; I prefer the blue ONES. As soon, however, as the word loses its real signification, and passes into a proper noun, it is at once declined: as, the Grays, the Pettys, the Quicklys; the Blues, the Liberals, the Conservatives, etc.114

It may happen that the position of the accent aids to produce change of function, as in the case of prófecto (pró facto), and in the very interesting case of igitur, which has been shown to be the enclitic form of agitur, originating in the common Plautine phrase (Quḯd agitur) Quíd igitur.115

The case is similar with the adverbial termination -ment in French and -mente in Italian, from the Latin mente. Cruellement (crudeli mente) and fièrement are intelligible formations; but solidement, lourdement, etc., are formed upon their analogy. At first applied only to adverbs of manner, the termination was transferred to adverbs of time and space; as, anciennement, largement. Our English termination -ly (from like) is a familiar instance of the same degradation of the final syllable: cf. godlike, by the side of godly.

The word self was originally an adjective meaning in Anglo-Saxon and Middle-English ‘the same,’ and declined in apposition with the noun or personal pronoun to which it was attached to mark emphasis. It then stood in the same case, number, and gender, he selfe, his selfes, him selfum, hine selfne, etc., gen. and dat. sing. fem. hyre selfre, etc. The history of the development from this usage to our present one is not quite clear; but we should remember that the terminations of the adjective were among the first to wear off completely, or at least to become confused and indistinct; and, further, that the accusative of the personal pronouns, was at an early date merged into the dative. We thus obtain the following schematic declension:—

Singular.

Nom. I self thou self he, she, it self Gen. my self thy self his, her, his self Acc. } Dat. } me self thee self him, her, him self

Plural.

Nom. they selve Gen. their selve Acc. } Dat. } them selve

Now, if we bear in mind that in these combinations the accent fell upon the word self (or selve), and that consequently the proclitic forms my, me, and thy, thee, in the genitive and dative had the same sound respectively,—and, further, that in the feminine of the third person singular (herself), these two cases were also alike,—it does not seem strange (1) that these two cases (genitive and dative) became confused, and (2) that the word self became a noun, as exemplified in such phrases as I said it to herself. Once having changed its function, the word assumed the flection of the new group to which its new function had attached it, and a plural form, as of a noun, arose—themselves, ourselves, theirselves.116 When once a single form served in three (genitive, dative, accusative) of the four cases, it not unnaturally succeeded in ousting the last, and succeeded all the more easily as I self was, of course, wrong, if self was a noun.

It is not, however, an invariable rule that the new associations into which a word enters in consequence of its change of function entail a change of form in the word. In Latin the word frugi was originally the dative case of a word frux, gen. frugis, meaning fruit, profit, advantage; and is actually employed by Plautus, with the full consciousness of its origin, in the phrase bonæ frugi esse (Asin., III. iii. 12). In fact, this use is exactly parallel to the use of usui in bono usui estis nulli, in Plautus, Curculio, l. 499; but in this case, usui, owing to its frequent occurrence, preserved the memory of its origin fresh. Cicero, however, treats frugi simply as an indeclinable adjective: Homines et satis fortes et satis plane frugi et sobrii (In Verrem, v. 27). Instances are also frequent where a change in meaning brings about a change in syntactical construction. Thus, for instance, in Latin we find that the nominative quisque is coupled with the reflective pronoun in the plural almost in the signification of singulatim.117 In Plautus we find præsente testibus (Amphitruo, II. ii. 203), and, in Afranius and Terence, absente nobis (Eunuchus, IV. iii. 7); in these cases the participles approach the characteristics of prepositions. A similar development gave to the present participle considering its present prepositional force. Macte is used similarly. Age! in Latin is used as generally as Come! in English, irrespective of the number of persons addressed; cave is used in the same way. Paucis is used for ‘a little’ in ausculta paucis (Terence, Andria, 536). Hélas is used in French by women equally as by men; φέρε, ἰδού, in Old Greek, are addressed to either one or many persons indifferently. In the same way, in late Greek, ὤφελον and ὤφελε were employed simply as conjunctions, without any consideration of number or person, the original construction having been Ὀλέσθαι ὤφελον τῇδ’ ἡμέρᾳ = ‘Would that I had perished on that day!’ In English albeit is used simply as a conjunction, and may be, in the sense of perhaps, is showing a tendency to fuse into one word, as it is actually written in American conversational language mebbe.

In German we find expressions like Heb hinten über sich das glas, ‘Raise your glass high’ (Uhland, Volkslieder) instead of über dich. In the same way we find in Latin suo loco, etc.; and in Latin law formulæ, Si sui juris sumus, where we should expect Si nostri juris sumus (i.e. ‘If we stand in our own rights’). In Old Norse a middle and passive is formed by the aid of a reflective -sik (sese), which is, of course, properly applicable to the third person only: it appears later as -st; thus, at kalla, ‘to call;’ at kallast, ‘to be called.’ In the same way, we have in English the words (to) bask and (to) busk,118 where the proper meaning of the termination has so completely died out that it is possible to write busk ye. The passive is similarly formed in the Slavonic languages.

Again, change of meaning influences the construction in the case of numerous verbs in Latin, which are properly intransitive, but are used as transitives. Such are perire,119 deperire; demori, used in the sense of ‘to be mortally enamoured of;’ stupere, ‘to marvel at;’ ardere, ‘to love with fire:’ the last-mentioned two words approximate in sense to mirari and amare respectively, and hence the instinct of language employs them in the same government.

The verb to doubt, in the etymological signification of hesitating between two beliefs, was, and is still constructed with whether. If, however, Spenser (Faëry Queene) says— ‘That makes them doubt their wits be not their aine,’ it is because the word is employed in this case, as indeed it frequently is in Shakespeare, in the sense of ‘to fear.’

The verb to babble, originally used intransitively, means to prattle or to chatter. When, however, it is employed in the sense of ‘to speak foolish words.’ or of ‘to reveal by talking,’ it is used with an object in the accusative case, and a passive is formed of it; e.g., Griefs too sacred to be babbled to the world. Again, compound words, as long as they are felt as such by the speakers, are naturally treated as such; cf. the Latin word respublica, which, though we write it as a single word, was declined in both its parts, respublica, reipublicæ, etc. But, when it had once become an indivisible unit—when the form république in French, or the English word republic, was formed with its various meanings, all closely resembling, but not identical with, that of the original compound, the word came to be treated after the analogy of other nouns, and the same derivatives are formed from it as from a simple form; cf. republican, etc. This fact is, again, instanced by such forms as high-spirited (high-spirit + ed, and not high + spirited), gentleman-like (gentleman + like, not gentle + manlike), good-natured (goodnature + ed, not good + natured).

Similarly, the Latin compound i (a demonstrative pronoun) + pse was at first declined as eumpse (e.g., Plautus, Truc., I. ii. 64), eampse, eopse, eapse, etc., all which forms are found in Plautus.120 When, however, the word came to be looked on as a simple word, it was declined as such: ipsum, ipsam, ipso, ipsa, etc.

In German there are many instances of words compounded with adverbs of place which are specially instructive as to the way in which a word may become detached from its previous use by a change of meaning. For instance, in modern German the usage is to say wirken AUF etwas, and not IN etwas, which was the usage even in the last century. In the same way, we speak of influence over as much as of influence on, showing that we have forgotten the significance of in.121

The word welcome in such phrases as I made them welcome is employed as an adjective, as, indeed, it is commonly apprehended to be. It was originally a substantive, and was derived from the infinitive mood of the verb, its meaning being pleasure-comer. The word is popularly supposed to derive from well and come; but the first element in the compound is really related to will—the true sense being the will-comer, i.e. he who comes to please another’s will. (Cf. Ger. willkommen.) The change in meaning seems due to Scandinavian influence, for in the Scandinavian languages the word is really composed of the adjective well and the past participle come; cf. Danish velkommen (welcome).122

The expression Quin conscendimus equos (Livy, i. 57) is properly Why do we not mount our horses? but is understood as Let us mount our horses; and in accordance with such usage quin may take after it an imperative, as quin age; or a hortative subjunctive, as quin experiamur? The sense of cur in some cases approximates to that of quod; and hence we find the word followed by a similar construction, in Horace, Ep. I. 8. 9;—irascar amicis, Cur me funesto properent arcere veterno. The O.Fr. car underwent a similar change. Derived from quare it meant, in the first instance, then; as, Cumpainz Rolond, l’oliphant KAR sunez (Chanson de Roland), i.e. Compagnon Roland sonnez DONC l’oliphant;123 it next came to be used like que or parceque after phrases like la raison est; and it then comes to be used with the conditional and imperative in the sense of utinam (cf. Diez, iii. 214).

In O.Fr. the word par (Latin per) was used for much. It took this sense from its use in combinations like perficere, perraro, etc., but it was detached from the verb, and was habitually used in O.Fr. in such combinations as par fut proz = il fut très preux; and in some cases coupled with other adverbs, like moult and tant; as, tant par fut bels = il était si beau, literally tant beaucoup (Chanson de Roland). The phrase survives in par trop.124

The Greek οὐκ οῦν, originally not therefore, like the Latin nonne, serves to introduce a question expecting an affirmative answer. It then comes to be used to introduce direct positive assertions; thus, οὐκοῦν ἐλευθερία ἡμᾶς μένει; from meaning ‘Does not, then, freedom await us?’ comes to mean simply ‘Therefore freedom awaits us.’ The word nanu in Sanskrit has gone through a similar development. Ne in Latin, properly the interrogative particle, comes to be used as the correlative of an:—faciatne an non faciat; or even faciat, necne. Similarly, in Russian, the interrogative particle li comes to be used as the correlative of ili (or); as ugodno-li vam eto? (‘Is this agreeable to you?’); but we then get combinations like dyélaet-li, ili ne dyélaet (‘whether he does it or no’).

The accusative with an infinitive could originally only stand in connection with a transitive verb as long as the accusative of the subject was regarded as the object of the finite verb, as audio te venire; but the accusative and infinitive came to be regarded as a dependent sentence with the accusative as its subject, and then we find the construction after words like gaudeo, horreo (Livy, xxxiv. 4. 3), doleo (Horace, Odes, iv. 4. 62), etc., which can properly speaking take no accusative of the object connected with them; as gaudere, dolere, infitias ire; nay, we find it after combinations such as spem habeo, etc. The accusative and infinitive construction then passes into sentences which depend on another accusative and infinitive, as (1) into relative sentences loosely connected; e.g. mundum censent regi numine Deorum—ex quo illud natura consequi (Cic. de Fin., iii. 19, § 64): (2) into sentences of comparison; e.g. ut feras quasdam nulla mitescere arte sic immitem ejus viri animum esse (Livy, xxxiii. 45): (3) into indirect questions; e.g. quid sese inter pacatos facere, cur in Italiam non revehi (Livy, xxviii. 24);125 (4) into temporal and causal sentences; e.g. crimina vitanda esse, quia vitari metus non posse (Seneca, Epist., 97. 13). A similar extension of the use is found in Greek.

The possessive cases mine, thine, his, her, its, our, your, their have passed into the category of adjectives, as in the case of Shall I not take mine ease in mine inn? (1 Henry IV., III. iii. 93). The instinct of language regarded mine, thine, etc., as the equivalents of of me, of thee, etc.; and marked the function by the addition of the possessive preposition of, as in this inn of mine. Thus, again, a gerund like killing,126 from having the same form as the participle, can be used in expressions like the killing a man, instead of the killing of a man.

We not only find that the word which changes its function undergoes the consequent changes in form or in syntax, but it also often happens that, owing to functional changes participated in by a certain group of words, such a group becomes detached, and thereby gains independence enough to influence other words that have cognate meanings. There are in Old English, as in German, many adverbs which are in their origin the genitives singular of strong masculine and neuter substantives, such as dæges (by day); but the origin of the termination has been forgotten, and the s has come to be looked upon as a merely adverbial termination. Consequently we find the adverb nihtes (by night), though niht is really feminine, and its genitive case is properly nihte. Similar formations are hereabouts, inwards, othergates (Shakespeare, Twelfth Night, V. i. 198), towards, whereabouts, etc. In the same way, the genitive plural of Anglo-Saxon substantives in -ung (later -ing) could be used adverbially; as,—án-ung-a, án-ing-a, (altogether), genitive plural of ân-ung, a substantive formed from án (one): after this analogy others were formed: as, hedling, afterwards altered to headlong; darkling, etc.