suffix appears occasionally as -ar- by dissimilation when an -l- sound has already occurred in the word; hence palmā-ri-s for *palmă-li-s. In Latin moreover many words appear with the -li- suffix which have -lo- in other languages: cp. oua-λó-s, Lat. simi-li-s. -mi- appears in a few words 0-μ-s (rt. *0e- of tí-On-μɩ), þñ-μı-s, Lat. ver-mi-s1. -ni- is very rare in Greek; cp. Kλó-vi-s, Lat. clu-ni-s with an unexplained difference in the root-syllable, Lat. com-mu-ni-s, ig-ni-s and some others. om-ni-s probably represents *op-ni-s3. -u- stems. 371. (2) The suffix -u- was employed originally to make both substantives and adjectives. It is not used as a secondary suffix. The feminine was made in -i- (-iē-), and in Latin all the adjectives have become -i-stems (§ 367). In compound adjectives a trace of the original stem sometimes remains, as in acu-pediu-s connected with wkú-s, and in genu-ini (sc. dentes) 'cheek-teeth,' cp. yévv-s. -u-stems are of all genders, and the root-syllable appears in different grades. For the relation in Greek between -v- and -ev- stems see $ 365. The suffix -u- appears also both Variations in as long and as short; πήχυς but ὀφρύς. The form of the genitive in Greek -u- stems seems to vary according to the quantity of the -v-; hence rýxεos (replaced in Attic by πήχεως) but οφρύος. The Attic forms xews aσTews are analogical. Homer has only the genitive in -ceos, which is preserved in Attic in the adjectives-déos etc. In Latin many -u- stems vary -u-stems. 1 An attempt has been made recently to treat these forms as an amalgamation of suffixes (Meringer, Beiträge, p. 3). 2 Bréal's view, that the plural omnes is homines in the weak grade and with the aspirate lost, is improbable. -tu- stems. in the dative and ablative plural between -u- and -iforms, the syllable being unaccented. The relation between yóvě and Lat. genu is difficult to explain'. 372. Of the suffixes composed of a consonant and -u-, -tu- is the most important. It is comparatively rare in Greek, but is widely developed in Latin in the form -ātu- to make abstract substantives, especially in the sense of function or office; consulatus, principatus etc. The infinitive forms called supines are cases of -tu- substantives formed from verb stems (§ 529). The ordinary Latin substantives in -tuare all masculine; the corresponding Greek forms such as Bpw-Tú-s, ¿d-n-rú-s etc. are all feminine. The neuter forms ão-Tu, pî-ru have no parallel in Latin. Forms in -tu- rarely occur from the same roots in Greek and Latin. Compare however i-rv-s (= FI-Tv-s), Lat. vi-tu-s; ap-rú-s, Lat. ar-tu-s. 373. Brugmann cites as other -u-suffixes -nu- (MyOther -u- suf- vú-s, Lat. pi-nu-s), -ru- (dák-pv, dakpü-μa, Lat. lacri-ma for *dacru-ma) and -lu fixes. (On-Au-s from dhe 'suck,' Lat. fe-l-are). -i-(-ie-) stems. 374. (3) The suffix -- and -i- was largely used to form feminines from existing masculine stems. The original form of the suffix and the relations between the -- and -i- forms are by no means clear, and though much has been written on the 1 Johannes Schmidt (Pluralbildungen, p. 50) contends that final short -u was dropped in Latin like final short -i, and that the long -ū is introduced later by using the collective plural instead of the singular. 2 The reading dacrumis for lacrumis in Ennius' epitaph nemo me dacrumis decoret has no ancient authority, but is an emendation made by Bergk. subject in recent years no certain conclusion has as yet been reached. The suffix appears in the nominative in Sanskrit as -i (dēvī‘goddess' fem. to deva-s, Lat. dīvu-s, Indo-G. *deiuo-s), but in Greek as -tă: deîa, Oeрáπaiva, οὖσα, δότειρα, ἀλήθεια representing respectively *ήδεια, *θεραπν-ια, *sont-ia, *δοτερ-ια, *ἀληθεσια. In Latin it appears in the great majority of the forms of the fifth declension ac-ie-s, spec-ie-s etc. But here the restoration of the original form is complicated (1) by the fact that these stems have assumed a final -s on the analogy of such stems as are included in the third declension, ab-ies etc.; and (2) because a number of such words. have byeforms in -ia, the regular representation of original -ia, cp. luxur-ie-s and luxur-ia etc. But as the suffix -o- seems to stand in ablaut relation to the suffix -i-, so ia- may possibly like -ie- have a weak grade of the form -- Forms with long -i- in Latin are found only when another suffix follows, as in vic-tri-x fem. to vic-tor; cp. do-Týp and dó-Teipa. Some suppose that -ă in the Greek nominative may have come from the accusative form -av and supplanted the older -i-', others consider a the older form, et adhuc sub judice lis est. In the adjectives Latin has added -s to the feminine forms, which thus become confused with other -i- stems. Thus suavi-s is properly the etymological equivalent of deia, although it comes to be treated as an -i-stem and used as such in all genders (§ 367). -0- and -a. 375. (4, 5) The -o- and -a- stems cannot be separated, the - forms having been used as feminines to the -o- stems from the proethnic stems. period (§ 291), although in all probability the suffix -à had originally nothing to do with gender. These suffixes 1 Brugm. Grundr. 11. § 109. are more frequent than any others. The -o-suffix is, indeed, so widely extended that the question has often been raised whether it ought not more properly to be treated as part of the root than as a suffix. And, as has already been mentioned, there seems to be no consonant suffix which has not an -o-form by the side of it, and even root nouns have parallel -o-forms. According to Torp's theory' the forms with -o- are the earlier. Thus from an original *pédo-s (cp. Skt. padá-m neut.) there came a form *péds, Lat. pēs with a "sentence-doublet" *pod-s Doric ws; from an original *légo-s (cp. Gk. λóyo-s) *leg-s, Lat. lex; from an original *bhéro-s (Skt. -bhará-, Gk. -þópo-s) *bhér-s, Gk. púp; from participial forms *dhe-to-s, *bhéuto-s came *dhét-s, bheut-s, Gk. Ons, 'free labourer,' pus man.' Torp attributes this change to the influence of accent, and almost alone amongst philologists constructs a scheme of original declensions consistent with the theory he propounds. One of these declensions may be given as typical of all-that of the stem found in Attic ἄρσην, Ionic ἔρσην”. 1 Den Græske Nominalflexion, pp. 1-18, (see § 344, note). 2 Torp, op. cit. p. 14. The same theory with certain modifications is held by other writers, and is the foundation of the article by Streitberg already mentioned (Die Entstehung der Dehnstufe, I. F. 111. pp. 305-416). 376. Apart from the distinction between -o- and -a-stems to indicate gender, a distinction Uses of -o- and which as we have seen (§ 293) is not fully -a-stems. preserved in the classical languages, the most common values of -o-stems are (1) as class names (common nouns), (2) as adjectives; the most common of -ā-stems as root abstracts. (1) olk-o-s : vic-u-s (§ 176 n.) : -wick (borrowed from Latin). pny-ó-s fag-u-s ¿vy-b-v jug-u-m φυγή : fug-a (2) vé-o-s : nov-u-s (§ 180) νέ-ο-ν : nov-u-m vé-a : nov-a : new 377. The combinations of -o- with a consonant may be taken in the same order as the consonant stems. Original -bho- is found developed to a small extent in Skt. and Greek, much more widely in Letto-Slavonic. With the possible exception -bho- stems. of mor-bu-s' it is not found in Latin. In Skt. and Greek this suffix is mostly confined to names of animals; Gk. a-po-s (where a = n), epipo-s, kidάon 'fox'.' Comκιδάφη pare however kóλa-po-s 'weal,' κрóтα-po-s 'temples,' Kopv-p'top' and the adjective apyv-po-s 'bright' with a byeform apyú-pe-os. -to-stems. 378. The suffix -to- is very common, especially in participial formations. In English, -ed as the suffix of the weak past participle is of this origin. 1 Brugmann, Grundr. 11. § 78. 2 For this adaptation of the suffix cp. Bloomfield, A. J. P. XII. p. 24 f. |