THE CASES.

It is commonly said that the cases are of local origin. Denoting in the first instance relations of place, they became by degrees transferred to relations of a more abstract character. This theory has gained favour because it is in harmony with that general law of progress by which the human mind mounts from what is tangible and concrete to what is abstract and intellectual. It is intelligible, and therefore possibly true. It seems to explain the striking contrast between the fewness of the forms of the cases and the multitude of their uses. It is also borne out to some extent by an examination of the forms of particular cases, especially the genitive and locative. On the other hand, it has met with opposition, as incompatible both with the philosophy and the forms of the cases. Professor Curtius regards it as untenable. Were it true, he says, that the nominative denoted the starting-point of the action of the verb, and the accusative denoted the goal, then, of the three categories of space, unde, ubi, quo, two are absorbed by the nominative and accusative, and but one, ubi is left for the remainder, five at least in number. Moreover if the nominative originally signified motion from a place, it would be included in the same category with the ablative and genitive.

These objections are alternatives, and cannot be urged simultaneously. If the local theory is unsatisfactory because it leaves but one category for all the cases except the nominative and accusative, it cannot also labour under the objection that it includes nominative and ablative in one and the same category. They assume, moreover, (1) That the nominative is a case in the same sense in which the accusative, genitive, and

ablative are cases; (2) That all the cases came into being at one and the same time; (3) That in the period of the formation of language the human mind was capable of forming such general notions as categories of space; and (4) That these categories are indivisible.

Let us examine these assumptions briefly. (1) The ancient grammarians distinguished between the casus rectus and the casus obliqui. In other words, they regarded the nominative as the standard form of a word from which the other cases were declined. In this they were most undoubtedly mistaken. The nominative is a separate formation from the stem, and requires a termination no less than the accusative; λóyo-s no less than Ayo-v. On the other hand, modern grammarians take great credit for putting the nominative on the same level with the other cases, and consider that in doing so they have removed an anomaly from the theory of declension. Yet it is a remarkable fact that the article (so called) ó, n, Tó has one form for the masculine and feminine nominative, and another form for the neuter nominative and oblique cases; and that it is this very article or pronoun which is applied in the formation of the nominative case, for the s of λóyo-s, xápi-s is in Bopp's opinion a remnant of the demonstrative sa, which in Greek appears as ò. In the personal pronouns also we find the nominative formed from one stem, and the other cases from another stem. The inference to be drawn from these facts seems to be that the old grammarians were correct in separating the nominative from the other cases, and regarding it as standing on a different level.

(2) The objection that the local theory leaves but one category, ubi, for five cases out of the seven, has no point unless it is assumed that all the cases came into existence at one and the same time. For as one case denoting this local relation became applied to other uses, its place would be supplied by a new formation of strict and obvious local meaning. In this manner there might without difficulty arise as many as five different forms, all in the first instance signifying a similar local relation. But such an assumption will scarcely be maintained in the face of what we know about the nature

of language. Forms are not created in groups but arise gradually as the need of them is felt. Thus prepositions have supplanted the cases; and auxiliary verbs have taken the place of the moods, but in neither instance has the change been sudden or the new uses introduced otherwise than singly. In language again, as spoken at the present day-and it is only from language actually living that we can gain any true notion of the life of language—we see the same gradual and isolated change. Unless therefore we suppose the processes which governed language in the earliest eras to have been quite different from those in force now, this assumption is quite untenable.

(3) and (4). Still less can we admit that the formation of language was carried on in subservience to such general and abstract notions as the categories. The categories are not necessary forms of thought, but merely summaries of relations under which we are accustomed to look at things. Such summaries cannot be made until the mind has become acquainted with abstractions, and language has become fixed: we require the aid of language in forming them. But language as we know from the study of barbarous dialects—begins with the individual impression, and proceeds by slow degrees to what is general. Thus the notions of creeping,' 'running,' 'walking,' 'riding,' are prior to the more general notion of 'going.' And similarly from beneath,' 'from above,' 'from the side of,' 'from out of,' are prior to the more general conception unde. What wonder then if more than one case can be subsumed under the category unde! Who would attempt to prove the existence of è as a preposition governing the genitive impossible because we have already πapà with the same case denoting motion from? 'From' = 'out of' is not by any means identical with 'from' in the sense of removal from.' The notions require two different prepositions; we may therefore suppose it possible that they were felt to require two different cases. Even therefore if it were admissible to make use of arguments based on the categories in investigating the origin of the cases, the mode in which they are here applied is more than doubtful

But these assumptions and the objections based upon them

are of a general character. It is to the forms of the cases that we must look for reliable evidence of their original meaning. Professor Curtius divides the cases into two groups, the first comprising the nominative, accusative and vocative; the second comprising the remaining cases. Without acquiescing in the division or in any arguments based upon it we may adopt. it for convenience' sake in going through the cases. In regard to the first group Professor Curtius entirely rejects the local theory, his chief objection being that were the nominative originally a case signifying unde and the accusative originally a case signifying quo, there would be a confusion of diametrically. opposite notions in the use of the accusative for the nominative in the neuter nouns, which use is very ancient.

We have seen reason to dissent from those modern grammarians (of whom Curtius is one) who place the nominative on the same level as the oblique cases. The sa, which appears as the s of the nominative, was found to be confined to the masculine and feminine gender and nominative case, the remaining cases of masculine and feminine and the neuter gender in all cases being formed from a different stem ta. Now this connection between the animate genders and the nominative case on the one hand and between the oblique cases and the neuter gender on the other hand is sufficiently striking. It suggests the inference that the s of the nominative is a suffix denoting animate gender. And what more natural or simple means could be adopted to denote the subject of an action in an operative sense than a gender-suffix, significative of life and action? If this be true the nominative may be at once removed from the number of cases of originally local origin. For the distinction of animate and inanimate gender is quite as primary as distinctions of a local nature.

The vocative is not a case at all. It does not bring a word into relation with the other words of the sentence. The termination, when it has one, is identical with that of the nominative, and the addition is due to false analogy.

The termination of the accusative is in the animate genders m, Greek v. The neuter has as a rule no termination—in the second declension only do we find m (Greek v); the pronouns Journal of Philology. VOL. II.

14

have d, e.g. quod, quid. This last-mentioned suffix is apparently identical with the pronoun ta which we have already found to be used only in the neuter gender throughout and the oblique cases of the animate genders. And as we found a connection between the nominative and animate gender so we may also infer some connection between the oblique cases and inanimate gender, for the oblique cases are all to a certain extent cases of the object and what is inanimate may be regarded as objective. The pronoun root ta then denotes inanimateness, or objectivity in general, in contrast to sa (ô, ǹ), which denotes life and animateness. Hence it is peculiarly applicable for the signification of the neuter in stems of indeterminate gender such as the pronouns (except the personal) seem to be.

When on the other hand the stem is animate or regarded as animate, another suffix is required for the case of the object. Thus m (Greek v) is added to all masculine and feminine nouns and to neuter nouns of the 2nd declension. Of the origin of this suffix we can say nothing certain. It has been suggested that it is identical with ma, which we find in the oblique cases of the first personal pronoun. However this may be, the real force of the suffix m would seem to be that it transfers what is animate into a case in which it may be regarded as inanimate. If this be true the stems of neuter nouns of the second declension must all or at least a portion of them which gave the type to the rest have been at one time of animate gender. In support of this hypothesis it may be said (1) that there are no neuters of the first declension (which is ultimately identical with the 2nd); that the neuters of the second declension are anomalous; that only in the 1st declension can the s of the nominative be dropped, which seems to imply some notion of agency in the stem; and lastly, that Professor Schleicher considers the a or o in such stems as bhara-, pepo- to be a 'determinative pronoun, signifying the doer.' In form, then, eóv and TÉKVOV are identical; both are to be regarded as inanimate, both are accusatives, both are neuters.

With regard to these three cases then, the theory of local origin appears untenable, and if what has been said is true we hardly require the additional proof which Professor Curtius