Even at present, in the case of a verb like to smell, the relation between the subject and predicate differs essentially when we say, I smell the flower; or, The flower smells. An effort on the part of our linguistic imagination is again needed, but the effort need not be very difficult, in order to enable us to realise that in a sentence like John smell flower, or John strike Alfred, BOTH nouns may once have been felt as standing in the subject relation to the predicate; so that, again, in the latter sentence, gestures or circumstances were needed in order to make it clear who was the acting subject and who the suffering subject, whereas, in the former sentence, no such confusion could arise.
If we take a sentence like ‘Give him a book,’ we feel both the person and the thing as objects of the action; and observation of this fact will enable us further to understand still more clearly that, at an older period of language, two subjects may have stood in the same sentence with the same predicate, though the relation between them and that predicate was not the same. It may further aid us to understand how, when once one of these subjects had developed into the grammatical category of OBJECT, the possible relations of such objects were so varied that the differentiation into various grammatical categories of accusative, dative, etc., becomes intelligible and natural.
The object, when once developed, may and often does become, by the nature of its relation to the predicate, a mere limitation or definition of such predicate, instead of remaining a member of the sentence equivalent in importance and weight with the subject, as it is, e.g., in such sentence as John strikes Alfred: whilst in a sentence like John runs a mile, the object is a mere attribute to the predicate, and the sentence can no longer be looked upon as tripartite, but must be regarded as consisting of two parts, i.e. (1) the subject, and (2) the predicate with its extension. These two cases, however, are not separated by any clear line of demarcation.
And just as the predicate may receive such a defining word, so may the subject and the object developed from it. These now commonly occur in the shape of attributes, whether substantival or adjectival, and genitives of substantives; as, The cattle are the farmer’s best; The cattle are beautifully fat. This could not be expressed at all in languages which have as yet developed no inflections: these could merely employ the defining word in juxtaposition to the word defined; as, in Chinese, T’su sin heu sin t’u ye, literally meaning ‘Origin Sin prince Sin spring final part,’ i.e. ‘Originally the prince of Sin sprang from Sin,’ i.e. ‘was born of a woman of the Kingdom of Sin.’ The fact that the determinant attached to the subject is not a predicate can then only be discovered by the presence of a third word which is detached from the two words that together make up the subject by a greater stress or, it may be, by a slight pause. Thus, if we say, liber pulcher, it is impossible to say whether pulcher is a predicate or merely the attribute to liber, unless we add some verb like est or habetur, or unless the custom of the language leads us to apprehend pulcher, from its position, as a predicate.
In truth the determinant, in this case ‘pulcher,’ is nothing but a degraded predicate, uttered not so much for its own sake, i.e. for the information it conveys, as in order to assign to this group of subject and determinant a further predicate, which predicate then conveys the real information; as, Liber pulcher nobis gaudio est: Hæc res agetur nobis, vobis fabula (Plautus, Captivi, Prologue.)
We have stated that the determinant is merely a degenerate or degraded predicate. The meaning of this statement may be most easily apprehended from cases in which the finite verb is affected by this degeneration, so that of the two predicates one might be logically replaced by a relative sentence; as, There is a devil haunts thee (Henry IV., Pt. I., Act II., iv.); I have a mind presages me (Merchant of Venice, I. i.); He groneth as our bore lith in our stie (Chaucer, Canterbury Tales, 7411); And was war of a pistel stood under a wal (Tale of Gamelyn); I’ll have none shall touch what I shall eat (Massinger, City Madam, I. i.); I can tell you news will comfort you (ibid., III. i.); The price is high shall buy thy vengeance (Middleton, Spanish Gipsy, V. i. 443).
A similar construction was found in the older stages of the Romance languages; cf. O.Ital. Non vi rimasse un sol non lacrimassi (‘There remained none did not cry’); O.Fr. Or n’a baron ne li envoit son fil (‘There is no baron does not send him his son’). Nor must we suppose that this construction is one peculiar to the Indo-European languages, and entirely inherited from an early stage in their development. Its use in Teutonic languages becomes more general towards the end of the Middle Ages than before that time. But even in Semitic languages like Arabic, we meet with expressions such as ‘I passed by a man slept.’
In the above instances, we have seen that the finite verb could sink into the position of a mere attributival determinant. In other words, in such a sentence as ‘There is a devil haunts thee,’ the very words show that the important word, in which the chief information lies, is devil, while the verb haunts might almost as well be expressed by an adjectival attributive, as ‘haunting.’ It is plain that if a verb could thus easily lose its predicatival character, a predicate bearing no distinguishing marks of its verbal character could, with even more facility, be similarly degraded. The border-land between meus in ‘liber meus’ (= the book is mine) and liber meus amittitur is a very narrow one.
It is very necessary to distinguish between the various functions of the determinant—the differences in which, however, commonly remain undenoted by us by any corresponding verbal difference, though they are, logically speaking, of the greatest importance. The determinant may leave the extent of the subject untouched; in other words, the epithet may apply to all the objects or ideas which the substantive by itself, or limited as it is by other circumstances, denotes: this is the case in mortal man; the almighty God. On the other hand, it may serve to restrict the meaning of the substantive; as when we say, old houses, an old house, a (or the) son of the king, the journey to Paris, Charles the Great. Similarly, if we say, the old house, meaning to contrast it with the new one, it is obvious that we individualise the meaning of house: while the expression would come under the first head in a sentence like Lo, the place where I was born! Humble as it is, I love the old house. In the latter class of instances, the determinant must be expressed, because without it the predicate is meaningless or untrue. If we say, A journey obliges us to cross the channel, we ascribe by these words to all journeys what is true of some only, e.g., of a journey to Paris. In the first category, in considering the epithet, we may notice that it may already be known as commonly attached to the word to which it is appended, as in This red wine (the speaker holding it up) I prefer to many more expensive ones; or it may tell us something new, as in the case of That poor man has no children, where the sentence without poor would state the same fact, the word poor conveying additional information. In this case it approaches the nature of a true predicate, and we often employ a relative sentence to express it: thus, instead of saying, Poor Charles has had to emigrate; if we wished to emphasise the adjective, we should say, Charles, who was poor, etc. Again, the determinant need stand in no direct relation to the predicate, as in our above example, where the fact that the man has no children is independent of his being poor; but it may also stand to the predicate in the relation of cause and effect, as in The cruel man would not listen to his victim’s prayers, where the determinant ‘cruel’ is applied owing to the fact mentioned in the predicate.
We have now seen that attributes are degenerated predicates. There are sentences in which the determinant has, as yet, a somewhat greater independence than is the case with the ordinary attributes, and which, therefore, may be said to represent a transition stage. In a sentence like He arrived safe and sound, the determinant safe and sound is still predicate, in the wider sense of the term, to he, but subordinate to the other predicate arrived, which alone in present grammar would bear this name. Safe and sound are, IN COMPARISON WITH arrived, a mere attribute to he, and nowadays such determinants are, for the linguistic consciousness, what has been very correctly termed PREDICATIVE ATTRIBUTES. These are distinguished from ordinary attributes by a greater freedom in the place they may occupy in the sentence, and thereby manifest their greater independence.