II. MORPHOLOGY. A. ROOTS AND STEMS. THE FORM OF INDO-EUROPEAN WORDS. The Indo-European language is one of the highest morphological regularity, since, besides the adjunction of sounds indicating relativity, it is also capable of flexion, that is, of regular variation of the root itself for the purpose of expressing relativity; this variation of the root consists in the stepraising of its vowel (§ 2). The addition of sounds expressing relativity is found at the end only of a root, never at its beginning (the augment is an originally independent word, which only coalesces with the verb, and which can therefore disappear without detracting from the force of the word). Every Indo-European word actually employed in the language has a sound expressing relativity after the root, which, moreover, can also be reduplicated, e.g. da-dă-mi (I give): naked roots do not appear in Indo-European as words (secondary loss of sounds expressing relativity naturally does not come under consideration here). The unique exception occurs in the vocative of those nouns which possess no stem-formative-element besides their casesuffix, as e.g. stem vāk- (speech, √vak, speak), nom. sg. vāk-s, gen. vāk-as, etc., but voc. vāk. The vocative is, however, no real word, no element of a sentence, but a word which has § 80. § 80. assumed the form of an interjection, a gesture translated into sound. Very rarely (and even then due to relatively secondary processes) are relativity-sounds found in the root itself. This happens in the present form, as e.g. Gk. Xλaμßávw, √λaß (e-λaß-ov), μ is here a relativity-sound of the present; Lat. iungo, viug (cf. iug-um), etc. The earlier forms had here also, probably, the nasal after the root-termination, v. post. Conjugation.' From this present-stem with medial nasal the nasalization has spread further to noun-stems (e.g. Gk. Túμπ-avo-v timbrel, VTUTT strike; Lat. iunc-tu-s, iunc-tura, √iug, etc.). § 81. It is, therefore, a distinctive characteristic of the IndoEuropean language, that all words belonging to it have one and the same morphological construction; a regularly variable root and a regularly variable relativity-sound affixed thereto. The morphological formula for all Indo-European words is therefore Rs (v. Introd. II.). Root-formation. The earliest component parts of the Indo-European words are the roots. By 'root' we generally understand the meaning-sound, the sound that conveys the force of the word in question (as 'to be' is e.g. the root of as-mi I am, as-ti he is, etc.). But the stem- and word-formative suffixes also in Indo-European have arisen from originally independent roots by coalescence with other roots. Thus every Indo-European word may be treated as a whole which has gradually grown out of several roots, at the least out of two; of these roots one (the first) is the root of the word (in the narrower sense of the word 'root' in which it is generally used), and bears the meaning unaided, whilst the others have sunk to the subordinate position of relativitysuffixes to this chief-root, and have become welded on to it; e.g. as-mi (I am), as with meaning 'be'; the ma, here weakened as a suffix to mi, expresses the relation of the 1st pers. (ma as an independent root means 'measure,' think,' · 'man,' 'I'); as-ti (he is) vas+vta 'that' dem., 'he'; bhara-ti § 81. (he bears), √bhar (bear)+√a, here become a suffix, a root of demonstrative function, and ta (as in as-ti); vāk-s (speech n. sg.), vak (speak), +sa demonstrative, here shortened to 8, etc. In more simply constructed languages we can see the early steps still preserved, steps which we must pre-suppose in Indo-European (e.g. in the words adduced forms as ma, as ta, bhar a ta, vak sa). Since the suffixes of the Indo-European language arise thus from roots originally independent, it becomes clear why the suffixes, as regards step-formation, are treated in precisely the same way as the chief-roots (e.g. bhar-a-mi 'I bear,' with a of stem-formative sf. raised to a, beside bhar-a-ti 'he bears,' without step-formation; ta-nau-ti ‘he stretches,' beside ta-nu-masi 'we stretch,' etc.). The exact formula of the Indo-European word is therefore R* s*. We shall now treat of chief-roots, meaning-sounds, alone. We have not hitherto any accurate investigation into the laws of root-formation in Indo-European. What sound-combinations occur in the roots of the Indo-European? Could roots change from one vowel-scale to another at a date as early as that of the Indo-European original language? What extension of meaning is permitted by a root, and in what cases must we assume original roots phonetically identical but nevertheless distinct? Several other similar questions hitherto unanswered might be proposed in this place. Meaning-sounds or roots (chief-roots) are generally separable from the words which they now underlie. The Indo-European roots seem in the first period of the life of the original language to have possessed a sound-form still simpler than that shown by roots existing in the actual language, cf. e.g. √yu-g beside yu (join), ma-t beside ma (measure), etc. In such cases what we see is probably the welding of a second root on to the first. Original reduplication also appears, e.g. ka-k (cook), gi-g (live). The origin of such secondary root-formations is in a great many § 81. cases due to the cohesion of the root and the addition which originally formed the present stem, e.g. √man (think) from ma, gan (be born) from ga, bhandh (bind) from bhadh, etc.; the nasal originally characterizes the present stem only. Hence we divide roots into primary and secondary. It is an invariable rule that Indo-European roots are monosyllabic. There is no distinction, as regards form, between the so-called verbal-roots (roots conveying a conception) and the so-called pronominal-roots (roots expressing relativity); the roots i, ka, ta, ya, e.g. are pronominal- as well as verbal-roots (i demonstrative, go; ka interrogative, be sharp; ta demonstrative, stretch; cf. Beitr. zur vergl. sprachforschung, ii. p. 92 sqq. 'wurzeln auf a im Indogermanischen,' by A. Schleicher. We get the root in its fundamental-form when we have taken from a given word all sounds expressing relativity and their possible influence upon the radical sounds (a process generally easy, but sometimes scarcely possible), and reduced the root-vowel to its fundamental-vowel, whenever it appears in a raised form in the word, e.g. da is the root of da-dă-mi (I give), vak of vāk-s (speech), div of daiv-a-s (shining, heavenly, God), dyu= div of dyau-s (heaven), su (bear, beget) of su-nu-s (son), ta of ta-m (him), etc. Indo-European roots may be formed in any way provided they are monosyllabic. The following are the sound-combinations of the root: 1. Vowel, that is, accurately speaking, spiritus lenis+vowel, e.g. a (dem. pron.), i (go), u (Sk. enjoy oneself, favour, Sclav., Lith., and Lat. ind-uere, ex-uere). 2. Consonant+vowel, e.g. da (give), bhi (fear), bhu (be). 3. Vowel+consonant, e.g. ad (eat), idh (kindle), us (burn). 4. Consonant+vowel+consonant, e.g. pat (fly, fall), vid (see), bhugh (bend). 5. Two consonants + vowel, e.g. sta (stand), pri (love), kru (hear). 6. Vowel+two consonants, e.g. ardh (wax), ark (shine; cele- § 81. brate). 7. Two consonants + vowel + consonant, e.g. star (strew), stigh (ascend). 8. Consonant + vowel+two consonants, e.g. dark (see), vart (turn). 9. Two consonants + vowel + two (scandere). consonants, e.g. skand In the case of roots of the form consonant+a+consonant, or a+consonant, even when these forms arise through stepformation from u and i, there occurs frequently a transposition of the sounds, so that a comes to be final, e.g. gan and gna (know, be born); mar and mra (die); ghar and ghra (shine, be yellow or green); par and pra (fill); ak and ka (be sharp); i, raised to ai, and ya (go); hu raised to hau, hav, and hva (call), etc. Likewise div and dyu (shine). Note 1.-Transposition of consonants such as is assumed by Alb. Kühn, über Wurzelvariation durch Metathesis, Bonn, 1868, e.g. in vid (see) and div (shine), Sk. paç (bind), origl. pak, and Lat. cap (take), etc., I cannot consider proved. Note 2.-In a complete grammar of Indo-European this chapter ought to contain a full list not only of those roots which can be proved to be Indo-European, but also of those which are peculiar to individual divisions or families (fundamental languages) of the Indo-European language. Note 3.-Hindu grammar, which is in this respect still followed by many European philologists, assumes no verbal-roots in a. Hindu grammarians mark roots ending in a either (1) in the raised form (e.g. dhā put, dā give, instead of dha, da), or (2) give to the root final n and y, which arise from the formation of the present stem, and therefore originally belonged to a suffix (e.g. ģan be born, instead of ga; hve cry, instead of hva, hu; rāi bark, instead of ra), or (3) mark them with ō, which obviously does no more than indicate certain peculiarities of these roots in forming their tense-stems, since in really existing forms ō nowhere appears (e.g. çō sharpen, for ça, aç). Cf. Beitr. ii. 92 sqq. |