Chapter VII · CHANGE OF MEANING IN SYNTAX.

We have considered, in Chapter IV., the different ways in which words change their meanings: and have remarked that change of meaning consists in the widening or narrowing of the scope or application of each word. We wish, in this chapter, to point out that these processes are not confined to words, but that whole syntactical combinations are constantly undergoing changes of meaning of a similar nature. It may be well to give at the outset an instance illustrative of such difference. Let us take the sentence, ‘The book reads like a translation.’ In this sentence the meaning which we attach to the word book has developed from that attached to A.S. bóc, a beech tree.42 The word read has been specialised in meaning from the more primitive signification ‘to interpret.’ In the same way, translation meant originally nothing more than a transference of any kind, but has been specially applied to a transference of the ideas expressed by one language into those of another. Such, then, are examples of changes of meaning which have occurred in words.

But besides these changes, it is obvious that we have here a sentence in which the relation between the subject and predicate differs considerably from that which is the usual one. We do not in the aforesaid sentence mean to say that the subject book performs the action reads, but we wish to assert that the subject is of such a nature as to admit of some person performing the action in question. This usage of the subject and predicate, though, when employed circumspectly, it need cause no obscurity, yet is an exceptional usage, or, as we have elsewhere called it, an occasional one. Such a construction might, however, easily spread, and become habitual or usual. In that case we should have to admit that the meaning of the general syntactical relation between subject and predicate connected by a verb in the active voice had widened in extent, and contracted in content. Instead of stating that the subject does the action, we should now have to adapt the statement to the wider but more indefinite relation—the subject either does or admits of the action. We shall have occasion to return to these and similar phrases later on.

Now let us take the phrase ‘He reads himself into the mind of his author.’ In this case we shall find that the meaning of reads is the same as that which we usually attach to it; the peculiar meaning lies not in the separate words, but in the phrase taken as a whole. The particular, occasional use of the accusative himself, together with the combination of the words, is what expresses the whole thought implied; and thus we have here an instance of a specific construction in which the force of the accusative connected with the word is different from the force of the case in more common usage. Though the application of the accusative in the way we have just mentioned must originally have been an occasional one, yet the phrase, though it has indeed become specific, has become so common, that we may in this combination call its meaning usual. We have, then, in studying change of meaning in syntactical relations, besides the classification of occasional and usual, another distinction to draw; that between (a) a change of meaning in a general relation, without reference to the individual terms which happen to stand in that relation (such as subject and predicate, verb and object, noun with accompanying genitive, preposition and its régime), and (b) a change in meaning of a case, or other syntactical relation, with regard to a specific word or expression, in connection with which it has come to express a new shade of thought. These two classifications are independent of each other, and cross one another. It is further to be noticed that, just as it is impossible to draw a hard and fast line of distinction between the occasional and usual in the meaning of a word, so it is impossible to always clearly formulate when the change in meaning of a syntactical relation is general or special; nay, it would in many cases be difficult to decide whether a change of meaning in a group of words is owing to a change of meaning in the words, or in their syntactical relations. Yet it is necessary to keep the distinction in view.

Instances of these syntactical changes are common in all languages. We might take, as a simple instance, from the Latin, the syntactical change which is brought about in the relationship of the transitive verb and its accusative. Transitive verbs commonly take the accusative of the direct object; as, Grecia capta ferum victorem cepit. But many words not originally transitive become so when composed with a preposition; as, accedere, præcellere, transgredi, just as to forego in English is transitive, while to go is intransitive. This construction was then felt as usual. But besides these we find a quantity of verbs strictly intransitive employed with the accusative; as, ambulare maria, (to walk the seas: Cicero, de Finibus, ii. 34); ludere Appium (just as we say, to play the fool: Cicero, ad Quint. Fratr., ii. 15); saltare Cyclopa (to dance the Cyclops dance: Horace, Sat. I. v.); stupere donum, (Vergil); etc. It was felt that the relationship between ambulare and maria, e.g., was closely enough related to that of regere currum on the one hand, and to that of ambulare super maria on the other, to enable analogy to become widely operative in extending this use. The result was that some of the constructions passed into regular usage; some stood out longer, and must always have appeared as exceptional or occasional; as, sudare mella (Vergil, Eclogue iv. 30).

One of the most ordinary changes brought about by relations in syntax is that due to the relationship of what is commonly called the governing word and its case. The signification, for example, borne by an accusative standing in the relation of object to a verb may cause the verb to bear a meaning more special than its ordinary meaning. Thus, in the case of such a phrase as I beat, it is clear that in to beat a dog, to beat the enemy, to beat the air, different values are attached to the meaning of the word ‘(to) beat,’ and the word thereby is narrowed in its definition and correspondingly enriched in its contents. It seems natural to examine a little more in detail the relationship borne by the cases to the word which governs them: there seems no objection to the use of the word governs, provided only that it be understood with due limitations; that certain particular forms are commonly devoted to the expression of certain ideas or relationships, and that the idea be not entertained that there is anything in the nature of the meanings of the words indissolubly connected with a particular form.

To deal with the Cases first. It is impossible to set together the different uses of the genitive, and to draw from these by induction any certain proof of the functions which this case fulfilled in the primitive Indo-European languages. For instance, the use of the genitive when it depends on verbs seems to have nothing in common with that of the same case when connected with substantives. In the former case, for instance, in the Classical languages, we find merely a few isolated instances of the genitive regularly governed by verbs, especially those verbs which signify ruling over, remembering, lacking, etc. The genitive with nouns, on the other hand, seems most probably to have been used in Indo-European for the expression of any relation between two substantives, as indeed it was in classical Greek, and, to a less extent, in Latin; cf. such different usages as Cæsaris horti; docendi gratia; reus Milonis; urbis instar; me Pompeii esse scio (Cicero, Fam., ii. 13); Germanicus Ægyptum proficiscitur cognoscendæ antiquitatis (Tacitus, Annals, xi. 59); hoc præmii; ut adhuc locorum (Plautus, Captivi, 382). In modern English, on the contrary, the function of the genitive in connection with substantives is greatly restricted. Many usages possible in Anglo-Saxon are at the present day obsolete; for instance, Criste is ALLRE kinge king (Orm., 3588), MÁDMA mænigo (Beowulf, 41), ðaer wæs MÁDMA fela (ibid., 36), RINCA manige (ibid., 729), he ÐAES WÆPNES onláh sélran sweord-frecan = he lent the weapon to the brave hero (ibid., 1468-69), tó gebídanne ÓÐRES YRFEWEARDES = to expect another heir (ibid., 2453,) he ʒef Horse MÁDMES inoʒe (L.I. 163, Fiedler and Sachs, ii. p. 277).43 The genitive at the present day is confined to certain characteristically special usages, and possesses several apparently independent significations. It must, however, be noticed that the true inflectional genitive in English is that which characterises the possessive case; as, John’s hat. In other cases in Modern English, we have commonly dropped the inflection, and are accustomed to render the genitival relation by a periphrasis with the preposition of. Using the word genitive in this sense, we may say that the typical usages of the genitive in modern English are the possessive genitive (the man’s brother), the partitive genitive (a cup of wine), and the genitive denoting that the governing substantive is what it is in virtue of what depends upon it (the writer of the work). This last division falls naturally into two sub-divisions in the case of nouns of action: the subjective genitive (surly Gloster’s governance—Shakespeare, 2 Henry VI., I. iii.) and the objective genitive (the government of the country). These usages have survived the various original methods of the application of the genitive, and they must thus be counted amongst genuine grammatical categories.

The relation of the accusative to its governing verb resembles the relation of the genitive to its governing substantive. The most general definition of the meaning of the accusative might be that it denotes any and every kind of relation that a substantive can bear to a verb, except that of a subject to its predicate. It is, however, true that, in English, we are unable to employ it in every case to denote such relation: nor, indeed, does this use seem to have been permissible in the original Indo-European languages; though it is true that the accusative was used more freely and commonly in old Greek and Latin, for instance, than in later times: cf. such constructions as ἄπορα πόριμος (Æsch., Prom. Vinctus); Quid tibi hanc rem tactio est? (Plautus, Pœnulus, V. v. 29), humeros exsertus uterque (Statius, Thebais, v. 439). Hence, in considering the different uses of the accusative, we must at the very outset place those meanings side by side which have gradually become independent.

The first distinction which we must remark in the use of the accusative is that between the free accusative, or accusative which is independent of the nature of the verb which it follows,—as, to buy a hat,—and the attached accusative, which is connected with a few verbs only by a close tie, and in each case with a restricted signification,—as, to blow a gale, to row a race. The free accusative is more freely used in English than in French or German; many of the relations which in those languages are expressed by the genitive and dative are in English expressed by the case under consideration.

One of the original usages of the free accusative was the expression of an extension over space and time; and in this case, it is not always found with verbs. We have in Latin, Cæsar tridui iter processit (Cæsar, Bell. Gallic., i. 38); Unguem non oportet discedere (Cicero, ad Att., xiii. 20): and, in English, such uses as To write of victories next year (Butler, Hudibras, II., III., 173); My troublous dream this night (Henry VI., Part II., Act. II., ii.); where the dative was usual in Anglo-Saxon (see Koch, ii., p. 94; Mason, p. 147). As instances of the attached accusative, we must especially consider the accusative of such substantives as are ETYMOLOGICALLY CONNECTED with the verb; as, to fight a hard fight; to see a strange sight; sangas ic singe (Ps. xxvi. 7).44 This ‘cognate accusative’ most probably furnishes the cue to such constructions as Come and trip it as you go, where it seems to replace some noun, as, e.g., tripping. Once established, this use of it instead of a cognate noun in the accusative, would easily be extended to cases like to foot it for to dance a dance, where the use of the verb to foot is but an ‘occasional’ one, and apparently too unusual to admit of the formation of the noun footing in the sense of dance. We must, then, suppose that the word it stands for a dance, i.e. for an accusative not cognate with the verb actually used, but with another and synonymous verb. The use of the accusative of towns in Latin, in answer to the question Whither?—as, Ire Romam, Tarentum, etc., further illustrates the attached accusative with which we may compare expressions in English, as to go west; flying south, etc.

The usage, now common in English, whereby a predicative adjective is connected with an intransitive verb seems to be of later origin. Cf. to cry one’s eyes red; to wash one’s forehead cool; to eat one’s-self full; to dance one’s-self tired; to shout one’s-self hoarse. In these cases the predicatival force of the accusative must be regarded as a widening of the signification. No doubt, however, special factors must have aided to bring this construction into use: such as the survival of the memory of the general signification of the accusative, as representing the goal of the verbal action; and, again, the analogy of such cases as to shoot a man dead; to buy a man free; to strike a man dumb; to beat black and blue;—where the accusative serves to define the verb, and indeed, almost enters into composition with it, as it in fact actually does in many cases in German, like tot schlagen; cf. the English dumb-foundered. There are a large number of colloquial phrases which are similar,45 such as to talk a person’s head off; to worm one’s-self into another’s confidence; to read one’s-self into an author; to laugh a man down, etc.

There is, next, the case of the accusative after compound verbs, where the simple verbs are intransitive or govern a different kind of accusative from that taken by the verb when compounded. Such are circumdare and præcellere in Latin, and, in English, to forego, to underrate, to withstand, to outlast; or, A.S. ofer-swimman, forestandan, etc.; e.g., (hé) oferswam sioleða bigong—He swam across the sea (Beowulf, 2368): Wið ord and wið ecge ingang forstód—He withstood entrance against sword and spear (ibid., 1550).46 These are on the border line of ‘free’ and ‘attached’ accusatives.

There are certain verbs composed with certain prefixes which, in virtue of their composition, receive a transitive force; as, belabour, begrudge, bewitch, belie, befleck, etc., and which, in some cases, receive in addition the power of adopting a different kind of object, generally calling in the aid of metaphor to extend their meaning; as, embody, encompass, enthral, overrule.

An ‘attached’ accusative, or one properly attached adverbially, in a defining and qualifying sense,47 to one definite individual verb, has, as a rule, only one single meaning, limited by use. But sometimes we find that in this case, too, several applications have set in; such may have been in some cases original, and in others due to the fact that the one ‘usual’ signification has extended by ‘occasional’ transgression. Take such cases as to blow a gale, to blow a sail, to strike a blow; to strike a man, to strike terror; to run a race, to run a man down; to stone a man, to stone cherries; pacing the ground, the morrice pacings; to keep a man from harm, to keep harm from a man; to stick a man with a knife, to stick a stamp; and in Latin, defendere aliquem ab ardore solis, defendere ardorem solis ab aliquo; prohibere calamitatem a provincia, prohibere provinciam calamitate; mutare equum mercede, mutare mercedem equo. So, too, in Greek: ἀρκεῖν τινα ἀπὸ κινδύνου; ἀρκεῖν κίνδυνον ἀπό τινος.

Poetry has a strong tendency to aid such ‘occasional’ constructions to become ‘usual:’ for it is a part of the technique of poetry to produce strong impressions by using its material in a fresh and striking way: thus we find in Latin, vina cadis onerare (Vergil, Æneid, i. 199: a variation for cados vinis); liberare obsidionem (Livy, xxvi. 8), instead of liberare urbem obsidione; vina coronant (Vergil, Æneid, iii. 526) instead of pocula vinis coronant: δάκρυα τέγγειν = ‘to stain tears,’ instead of ‘to stain with tears’ (Pindar): αἷμα δεύειν = ‘to stain blood,’ instead of ‘to stain with blood’ (Sophocles). Thus, in English, we have The Attic warbler pours her throat (Gray); to languish a drop of blood a day (Shakespeare, Cymbeline, I. ii.) The relation expressed by the accusative may in itself be more than a single one; and thus the connection of a single verb with several accusatives to express different ideas is quite natural.

It seems hardly true to state that the Indo-European prepositions governed any particular case. The case which followed the preposition was actually referred to the verb; the general meaning of the verb was still felt and was merely specialised by the preposition; whence it comes that the same preposition is followed by different cases, each bearing its own special meaning. The Greek language offers good examples of this, and seems to stand nearer the original state, as far as usage goes. Take, for instance, a preposition like πάρα. Its general meaning may be defined as ‘from:’ when followed by the genitive it signifies ‘proceeding from;’ when followed by the accusative, ‘to,’ reference to the source not being overlooked: similarly with κατά, μετά, etc. In English, more than in most European languages, the tendency has been to multiply the use of prepositions, and to employ them independently of any feeling for the case. The case has thus become more and more independent of the preposition: the connection of the latter with the case has become merely matter of custom; and the consciousness of the original signification of the case has become fainter. With regard to the Latin prepositions which govern one case only (like ex, ab), or which govern more than one without affecting the sense (like tenus), the employment of the case is merely traditional, and no value can be attached to it. Between the absolute fixity of the one use and the original freedom of the other use stands the employment of in, sub, and super, sometimes with the ablative, sometimes with the accusative, but with different meanings for the respective cases.

The changes that have appeared in Syntax in the case of prepositions are very well exemplified in English, in which language their use has so greatly spread, and plays such an important part. They were, in the first place, prefixed to the verb, which they qualified adverbially,48 forming, in fact, a compound with it; as, ‘to overtake,’ ‘overreach,’ ‘overlook.’ They were next detached from the verb, but not prefixed to the noun; as, ‘to take over,’ ‘to reach over,’ ‘to look over;’ and the difference in meaning between these three pairs of phrases will show us how the preposition came to lose memory of the proper signification of the case. In a later stage still, they appear prefixed to nouns, and serve to particularise the relations of actions to things—relations which, in the inflected state of language, were expressed by the case endings of nouns; cf. Bigstandað me strange genéatas (Cædmon) = ‘Stout vassals bystand me;’ He heom stód wið (Layamon) = ‘He them stood against;’ or Again the false paiens the Christens stode he by (P. Langtoft) = ‘Against the false pagans the Christians he stood by;’ i.e. ‘He stood by the Christians.’

We sometimes find the partitive use of the genitive replaced by apposition. The simplest and most natural example of this is where the apposition is made up of several members which are collectively the equivalent of the substantive to which they are appended; for instance, ‘They went, one to the right, the other to the left;’ ‘Postero die terrestrem navalemque exercitum, non instructos modo, sed hos decurrentes, classem in portu, simulacrum et ipsam edentem navalis pugnæ ostendit’ (Livy, xxix. 22). ‘Duæ filiæ harum, altera occisa, altera capta est’ (Cæsar, Bell. Gallic., i. 53); ‘Diversa cornua, dextrum ad castra Sammitium, lævum ad urbem tendit’ (Livy, x. 41); ‘Capti ab Iugurtha, pars in crucem acti, pars bestiis objecti sunt’ (Sall., Iug.). But the same appositional construction appears when the whole apposition represents only a part of the expression or phrase of which it is the expansion; as, ‘Volsci maxima pars cæsi,’ (Livy): ‘Cetera multitudo decimus quisque ad supplicium lecti’ (Livy); ‘Nostri ceciderunt tres’ (Cæsar); ‘My arrival, although an only son, unseen for four years, was unable to discompose, etc.’ (Scott, Rob Roy, i.); ‘Tuum, hominis simplicis, pectus vidimus’ (Cicero, Phil., ii. 43). This is also the case where the subject is expressed only by the personal termination of the verb; as, ‘Plerique meminimus’ (Livy); ‘Simoni adesse me quis nuntiate’ = ‘Tell Simo, one or the other of you!’ (Plautus). Similarly, in the case of the designation of materials, we find an apposition taking the place of the partitive genitive; thus we find, in Latin, ‘aliquid id genus’ for ‘something of that kind;’ ‘Scis me antea orationes aut aliquid id genus solitum scribere’ (Cicero, Att., xiii. 12); ‘Pascuntur omne genus objecto frumento maxime ordeo’ (Varro, de Re Rustica, iii. 6);49 ‘arma magnus numerus’ (Livy). Thus, ‘He gained the sur-addition Leonatus’ (Shakespeare, Cymbeline, I. i.).

This more simple and primitive appositional construction is very common in modern xml:lang; as, ein stück brot, ein glas wasser: in Middle High German it was rarer; in modern Scotch it is common in such instances as a wee bit body, a curran days (a number of days): it was common in Anglo-Saxon; as, ‘scóp him Heort naman’ (Beowulf, 78); Emme broðer ðe queene (Robert of Gloucester); The Duke of Burgoys, Edmonde sonne (Wa., i. 87); David Kingdom (R. of G., i. 7.):50 and is found in Chaucer,—Gif us a busshel whet or malt or reye (Canterbury Tales, 7328); half a quarter otes (ibid., 7545): and has survived even in modern English, in such cases as The Tyrol passes (Coleridge, Picc., i. 10); Through Solway sands, through Tarras moss (Scott, Lay of Last Minstrel, i. 21). We must regard this method of apposition as the most primitive in language; the two words in apposition are simply placed side by side like two Chinese roots, and must be looked upon as the simple stems without any inflection.

Even the subject of a verb may deviate from previous usage in the way whereby it denotes a relation: cf. such phrases as The cistern is running dry; The roof drips with water; The trees drop honey. Thus we can say, The river is running over; The wood is resonant with song; The window will not shut; The fire will not draw; The kettle boils; This sample tastes bad; The hall thick swarming now with complicated monsters (Milton): in Italian, Le vie correvano sangue (Malespini): in Spanish, Corrieron sangue los rios: Sudare mella (Vergil, Ecl. iv., 30); cf. also, the use of sapere, in Latin, in such cases as cum sapimus patruos (Persius, Sat. i., 11); sentir, in French, as Cela sent la guerre. In these cases we should expect the subject and object to be inverted.

A similar departure from ordinary usage occurs in the case of what we commonly speak of as ‘transferred’ epithets; i.e. adjectives referring to merely indirect relations with the substantive to which they are attached. Such are expressions like wicked ways; quiet hours; in ambitious Latin (Carlyle, Past and Present, ii. 2); the blest abodes (Pope, Essay on Man, iii. 259). Many of these linguistic licences have become quite usual, and it is forgotten that the epithet attached to the word does not strictly fit it: thus we speak quite commonly of the happy event, a joyful surprise, happy hours, a learned treatise, an intoxicated condition, in a foolish manner, a gay supper, a bright prospect, etc.; and we can even say, He gives us an unhealthy impression, a stingy gift, etc. The word secure in English, like sûr in French, refers in the first instance to a person who need not be anxious; in the second place, to a thing or person about whom no one need be anxious. Thus we can say, I am safe in saying that he is safe. As soon as these freer combinations are apprehended as an ordinary epithet applied to its substantive, we may state that a change in word-meaning has occurred.

Such licence occurs in the case of the participles and nouns in -ing even more than in that of adjectives; thus we can say, in a dismantled state (Dickens, Pickwick, 2); a smiling answer; this consummation of drunken folly (Scott, Rob Roy, 12); a dazzling prospect; the selling price; the dying day; a parting glass; writing materials; sleeping compartment; dining room; singing lesson; falling sickness; waking moments; the ravished hours (Parnell, Hesiod, 225). So, too, we speak of a talented man; cf. also the common French expressions, thé dansant, café chantant. Tacitus has such uses as Muciano volentia rescripsere (Hist., iii. 52) for volenti, etc.

We may probably compare with this use that of the so-called ‘misrelated participle,’ a freely attached predicatival attribute, which is indeed condemned as ungrammatical and careless, but which still occurs very commonly in even the best authors. Cf. ‘When gone we all regarded each other for some minutes with confusion’ (Goldsmith, Vicar of Wakefield, 13); ‘Thus repulsed, our final hope Is flat despair’ (Milton, Paradise Lost, ii. 142); ‘Amazed at the alteration in his manner, every sentence that he uttered increased her embarrassment’ (Miss Austin, Pride and Prejudice, ch. xliii.).51 We are, indeed, accustomed to say that in this case we must supply a subject, and that the full expression would be ‘Amazed as she was,’ in the last instance cited. But if we use such an expression as ‘a pitying tear,’ we might maintain as well that it is necessary to explain this as, ‘with a tear, shed in sign of his pity.’ The fact is, that these loosely appended predicatival attributes answer to a need felt in language, just as much as such words as regarding, during, vu que, instar, supply a requirement in the prepositional category.

In the case of participial constructions, the participle expresses formally the time-relation in which the condition or action denoted by the participle stands to the finite verb. Thus, ‘Being frightened he runs away’ expresses formally nothing more than the temporal relation between the fright and what follows it. It is, however, possible to understand different relations as implied by this participle; thus there would, in this instance, be a connection of cause and effect. There are many cases in which, were we to extend the participial construction into a separate sentence, we should have to employ different conjunctions; sometimes those denoting the reason,—as, ‘Since he was frightened he ran away;’ sometimes we should have to employ such conjunctions as denote an opposition,—as, ‘Notwithstanding that;’ thus, supposing that the sentence in question ran, ‘Being frightened he did not run away,’ this would naturally be broken up into ‘Notwithstanding that he was frightened, he did not run away.’ Sometimes, again, the participle expresses a condition, as in such common cases as ‘Failing an heir, the property passes to the crown.’

Still it is unnecessary to assert that the participle, as such, denotes these different meanings—such as cause, condition, opposition, etc. These relations are only accidental and occasional. When, however, we have dependent sentences introduced by a temporal conjunction, like quum, since, the accidental relation of this conjunction to the governing sentence may come to attach itself and become permanent; in this case, the conjunction will experience a change of syntactical meaning. Take the case of since, formed by the adverbial genitive suffix es, from sin = sithen (from sið, þ̱am, after that). While, again, from meaning ‘the time that’ (a thing occurred,) has come to denote ‘in spite of the fact that,’ in such phrases as ‘While you pretend that you love me, you act as though you did not.’ In the case of the modern German weil, the temporal signification has completely disappeared; and in the same way prepositions, such as through and by, which possess strictly speaking a local or temporal meaning, pass into a causal meaning.

The instances given above may serve to show the way in which changes are constantly occurring in syntax, and will aid in pointing out how language is constantly aiming at supplying, in an economical fashion, its needs as they successively present themselves.