The predicate or attribute, instead of agreeing with the subject, or with the word which it serves to define, may agree with a genitive dependent on that subject; as, Ἦλθε δ’ ἐπί ψυχή Θηβαίου Τειρεσίαο χρύσεον σκῆπτρον ἔχων (Homer, Od., xi. 90), ‘The soul (fem.) of the Theban Teresias (masc.) came having (masc.) a golden sceptre.’ In English we find ‘There are eleven days’ journey from Horeb unto Kadesh-barnea’ (Deut. i. 2).
In French it is customary to say La plupart de ses amis l’abandonnèrent, ‘The most part of his friends abandoned (plur.) him;’ but La plupart du peuple voulait, ‘The most part of the people wished (sing.):’ in the former case the quantity of individuals is regarded; in the latter the people are looked upon as a totality divided.
The attribute sometimes in Latin and Greek, referring to the person addressed, appears in the vocative: as, Quibus Hector ab oris Expectate venis? (Vergil, Æn., ii. 282), ‘From what shores, Hector, O long expected, dost come?’ Stemmate quod Tusco ramum millesime ducis (Persius, iii. 28), ‘Because thou, O thousandth, dost draw thy lineage from an Etruscan tree.’ Thus, in Greek, Ὄλβιε, κῶρε, γένοιο (Theocr., Id., xvii. 66), ‘Mayst thou be happy, O boy,’ lit. ‘O happy, O boy, mayst thou be!’
Such examples as these may aid us to understand the way in which concord has spread beyond the area to which it strictly belonged. And we may gather from these some idea of the way in which this process grew up in prehistorical times. We must remember, however, that concord was not felt so indispensable in the earliest stages of language, because absolute forms without inflectional suffixes were then the rule.
The question now comes, What were the rudiments from which concord proceeded? We must suppose that a period once existed in which substantives coalesced with the stem of the verb, and in which pronouns could precede the stem, just as our actual verbal inflections seem to owe their origin in many cases to the coalition of pronouns with the stem. We must therefore suppose that, just as it was possible to say Διδω-μι (‘Give I’), so it was possible to say ‘Go father,’ ‘Father go’ (for ‘Father goes’); and ‘I go,’ just as it was possible to say ‘Go I,’ ‘Go thou,’ ‘Go he’ (instead of ‘I go,’ etc.). There are actually some non-Indo-European languages in which the third person singular differs from the other persons by dispensing with any suffix. Such is Hungarian,[176] in which the root ‘fog,’ ‘seize,’ is thus declined—fog-ok, fogo-s, fog. Here, then, the original plan maintains itself, of coalition according to the formula ‘Go-father,’ or ‘Father-go.’ In the next stage, the subject is repeated, as, when we say Ἔγω δίδωμι, we are really saying ‘I give I.’ This process is very common in some modern languages, especially in poetry, when emphasis is to be given to the subject: as, The night it was still, and the moon it shone (Kirke White, Gondoline);[177] The skipper he stood beside the helm (Longfellow): Je le sais, moi; Il ne voulut pas, lui; Toi, tu vivras vil et malheureux,—‘I know it, I;’ ‘He would not, he;’ ‘Thou, thou shalt live vile and wretched.’ Similar is the anticipation of the subject by an indefinite il; as, Il suffisait un mot, ‘There sufficed a word.’ The pronoun was originally doubled only where it was specially emphasised, just as in uneducated conversation at the present day we hear such forms as I says, says I. But such pronominal reduplication must have spread, and have affected the verbal forms when they were completely formed, just as it, at an earlier period, affected the tense-stems. It is, however, by this time so far forgotten that the termination of such a word as legit represents a personal pronoun, that its most common use is to indicate its relationship with the subject by mere concord; as Pater legit, lit. ‘Father read—he,’ i.e. ‘father reads.’ In fact, the personal endings at the present day merely serve to mark the verb as such, and sometimes to express the difference between different moods.
In the case of nouns, the concord of gender and number, at any rate, is first formed in the pronoun to which reference is made, to which gender, too, owes its origin, as in such cases as illæ mulieres, ‘those women (nom.);’ illas mulieres (acc.).
Concord in case appears first in apposition; as, Imperatoris Cæsaris exercitus, ‘The army of Cæsar (gen.) the commander (gen.),’ where it serves to show that both nouns have the same relation to exercitus. But here there is no more actual necessity for employing the case-ending twice, than there is for repeating the pronominal suffix in the case of the verb. This we may see in such cases as King Arthur’s seat; La gloire de la nation française, ‘The glory of the French nation.’ A concord in gender and number occurs, even at the present day, only where it is demanded by the nature of the case; as, La dame sur le visage de laquelle les grâces étaient peintes (Fénelon), ‘The lady on the face of whom the graces were painted.’
The concord of substantives in apposition having been the first to form itself—as in Cæsaris imperatoris Romani, ‘Of Cæsar (gen.) the Roman-commander (gen.)’—we must suppose the concord of the attributival and predicatival adjective to have been modelled upon that use; as, Cæsaris domini potentis, ‘Of Cæsar (gen.) the powerful master (gen.),’ or Cæsaris invicti, ‘Of Cæsar (gen.) unconquered (gen.).’ In other words, their origin reaches back to a time when the adjective still occupied the same category as the substantive, and was not yet thought of as occupying a category of its own. The transition is marked by such substantives as are called, in Latin grammars, Mobilia, which in the forms of their genders resemble adjectives. Such as coquus, ‘cook’ (masc.); coqua, ‘cook’ (fem.): dominus, ‘lord;’ domina, ‘lady:’ rex, ‘king;’ regina, ‘queen.’ As these substantives passed into adjectives, they maintained the concord, and it then came to be regarded as of the essence of the adjective.