tree," or to determine the order in which they separated from each other (see Schmidt's Die Verwandtschaftsverhältnisse der Indogermanischen Sprachen, Weimar, 1872). But the prevailing view is still that of Lottner, Curtius, Jolly, Fick, and Scherer, that we may with confidence assume the first division to have been that between the Aryan or Asiatic (Indo-Persian) and the European groups, and that there are sufficient points of agreement between all the European languages to warrant us in assuming that there was a period of some duration during which the European peoples remained united.
Of these points of agreement the most important are the following:—
1. The vowel a is found to have "split" on European soil into the three vowels e, a, o,—that is to say, there are numerous instances in which the European languages agree in degrading a primitive a into e or o when the Asiatic tongues either retain the a or weaken it quite independently into i.
2. The Europeans agree in softening a primitive r into l, where the Asiatics have retained r.
3. There are a large number of new words, and apparently even some new roots, common to most, if not to all, the European languages, of which no trace is to be found among the Indo-Persians.[1]
These facts cannot be set aside by instances of agreement in inflexion or syntax between Greek and Sanskrit, for example, for it is much easier to believe that at the comparatively late date at which any Teutonic language is known to us,—and much more so, at the far later date of the earliest Celtic records,—the inflexions which they presumably once had in common with Greek had become to a large extent worn away and unrecognizable and the syntactic constructions modified, than to suppose that such numerous instances of agreement were wholly fortuitous.
A similar course of argument fairly leads to the presumption of a common Græco-Italic nationality. The agreement in vocabulary is still closer than that between the various members of the united European group: for instance, while the general terms for agriculture are shared not only by Greeks and Italians, but also by Teutons, Celts, Slavs (though not by Indo-Persians), there are many special terms which are only found on Greek and Italian soil, the most interesting among them being perhaps the words for wine and oil. Other words, again, which are used with a more indefinite meaning by the Europeans generally are specialized and differentiated in Græco-Italic (Fick, Vergl. Wörterb., ii. pp. 1–288; Curtius's Principles, 230b, 234, 597, &c.). Whether we may also assume (as is done, e.g., by Professor E. Curtius, History of Greece, vol. i. p. 19) that there was a common Græco-Italic law of accentuation is very doubtful, in face of the arguments advanced by Corssen in favour of a freer law of accents in the earlier times, both in Greek and in Latin. It is much more probable that the rigidity of the Latin system, and the exquisite flexibility and harmony of the Greek, were developed quite separately from a more fluid state. But undoubtedly there is a far greater similarity in the inflexional system of Greek and Latin than can be established between either of these and any other member of the group.[2]
At the time when the common Indo-European unity was first broken up, the language had reached a stage of development which may be given with some confidence as follows. The steps assumed are those which have been established by Professor Curtius in his monograph, Zur Chronologie der Indogermanischen Sprachforschung (2d edition, Leipsic, 1873). In spite of the criticisms to which this scheme has been subjected, by far the most important of which are those by Max Müller and Ascoli, it may fairly be said to maintain its ground, and it is reasserted with full confidence in Curtius's admirable work on the Greek verb.
We start with the period in which roots alone were employed as words. As to the origin of these roots, philology is as yet quite unable to speak with any positiveness: all that can be said is that the imitative or onomatopoetic theory has not been proved to be capable of producing all the roots which we are compelled to postulate, while, on the other hand, no theory has been generally recognized as fit to be regarded as a serious rival. It is clear, however, that we must admit an extremely early, if not an absolutely primitive, distinction of roots into verbal and pronominal roots, i.e., (1) such combinations of sound as were significant, and carried with them a notion which was vague and general, if not philosophically abstract, and (2) such as had no meaning in themselves, but only served to denote relations.
The second stage is that of the "determination" of roots, wherein, by the addition of different phonetic elements, they acquired a differentiated meaning—e.g., when the very vague ju, "join," became ju-g, "join together," ju-dh, "join in battle." (It may here, however, be open to question whether the fuller forms were developed from the shorter by additions, or the shorter abstracted from the group of similar fuller forms, as Max Müller is rather inclined to hold.)
The third stage is that of the formation of verbs, by the close combination of a verbal root with one or more pronominal roots, to denote the character of the subject of the verb. It is in the nature of this combination that we find the distinguishing feature of the Indo-Germanic stock of languages. At the same time we find (1) the "strengthening" of the vowel of the root, by the addition of the simplest vowel-sound a, to denote repeated or continuous as distinct from momentary action; (2) reduplication, originally producing the same effect, but afterwards, in a specialized form, denoting the continued result in the present of an act done in the past; (3) the augment, a particle, originally demonstrative in its nature, prefixed to a verb to denote that the action expressed by the verb took place at a time removed from the present, i.e., in the past. To the same stage (though possibly to a later part of it) belongs the further development of terminations, so as to mark an action as having a special reference to the subject; this produces what is in Greek conventionally called the middle voice, but what is really a reflexive formation. We may take as types of the words created during this stage such forms as da-ta, "give there," i.e., he gives; da-da-fa, "he is giving;" a-da-m, "I gave;" da-da-mai, "I give with a view to myself."
In a fourth stage we get the expansion of the root into a stem, occasioned apparently, in the first instance, by the increasing need of distinguishing the noun from the verb. The earliest method of forming a stein was by the addition of a "thematic vowel" a to the root, to convey the notion of a continuous action; thus from bhar, "carry," came bhar-a, "carrying." Sometimes the vowel of the root was "strengthened" along with the addition of the thematic vowel; thus rik, "leave;" raika, "leaving." Afterwards other similar formative elements (or pronominal roots) such as ta, na, ma, tra, &c., were added to produce nominal stems of many various kinds. There is no reason to suppose that these were at first strictly differentiated in meaning; thus par-nu- is "filled" not "filling," but su-nu- may be taken either actively or passively, "the begetter" or "the begotten," and tap-nu-is "the burning" fever. Subsequently the instinct of language availed itself of variations in form to distinguish various relations, especially of gender. Again, when noun-stems came to be used, as the roots had previously been used, to form verbs by the addi-
- ↑ See Curtius, Ueber die Spaltung des A-Lautes, Leipsic, 1864; Fick, Die ehemalige Spracheinheit der Indogermanen Europas, Göttingen, 1873; Vergleichendes Wörterbuch der Indogermanischen Sprachen, 3d edition, Göttingen, 1876, where the list of common European words fills vol. i. pp. 471–843. Compare Grundzüge, p. 93.
- ↑ Schleicher's doctrine as to the close relation of the Latin and the Celtic inflexions is not now generally accepted.
