Remerciez-le!

Remerciez @Admin pour avoir partagé cet document gratuitement, de la manière la plus simple, en partageant sur les réseaux sociaux.

1 A Look over Lat. umerus 29th WeCIEC, UCLA, November 3–4, 2017 Stefan Hoefler

1 A Look over Lat. umerus 29th WeCIEC, UCLA, November 3–4, 2017 Stefan Hoefler hoefler@fas.harvard.edu §1 The dilemma The PIE reconstruction of the word for ‘shoulder’: *(hx)omeso-A *(hx)ómso- Lat. umerus, -ī m. Umbr. loc. sg. uze, onseE Ved. áṁsa- m. Gk. ὦµος m.B PGmc. *amsa- (Goth. acc. pl. amsans) Arm. owsC Toch. A es, B āntseD Umbr. loc. sg. uze, onseE A Walde/Hofmann II:815; IEW:778; de Vaan 2008:640; Weiss 2009:140. B *-ómsV- > Att.-Ion. -ῶµ- (not †-οῦµ-) is the product of regular sound change, as in κῶµος m. ‘revel’, Ved. śáṁsa- m. ‘praise’ < *k ̑ ómso- (cf. Durante apud Peters 1980:307 n. 253; LIV2:326 s.v. *k ̑ eNs- n. 1; Hackstein 2002:190f.). The expected Aeol. outcome -οµµ- can be seen in ἐπ-οµµάδιος ‘on the shoulders’ (v.l. in Theocr. 29.29). C The development *o > *u before a nasal is regular; cf. cownr ‘knee’ < *g ̑ ónu and Schmitt 2007:49f. D The Toch. words (cf. the seminal treatment in Hilmarsson 1989:6ff.) seem to presuppose a *(hx)ōmso- (or *h2emso-, see below). However, a dual *(hx)omsoh1 > *omsō(+nº) > PT *ænsā(+næ) (with *ºnæ as the dual marker) would have constituted a possible environment for the phenomenon known as ā-umlaut, by which PT *ænsā(+næ) > *ānsā(+næ), whence TB āntsne (verse), aṃtsane (prose), TA esäṃ, from which the singular TB āntse, TA es was backformed / analogically remodeled. I am unsure, however, why the umlaut-causing *ā (< *ōC-) was shortened to *-ä- in TB; in TA, a threesyllabic PT *ānsānæ would have undergone “Vokalbalance” to esäṃ anyway. Incidentally, the same development can be assumed for *podyoh1 ‘(the two) feet’ (cf. Hilmarsson 1989:13) > PT *pæyyā(+næ) > *pāyyā(+næ) >> *pāyyä(+næ) > TB paine, TA peṃ. In any case, an analogical explanation of the first vowel of āntse is also possible, cf. Peters 1980:307f. note 253. I plan to take a closer look over the Toch. ‘shoulder’ on another occasion. E Umbr. loc. sg. uze (IIb 27,28), onse (VIb 50) ‘in umero’ can apparently continue either *(hx)omes(e)i or *(hx)oms(e)i (cf. Meiser 1986:163). See also below note 10. N.B. Hitt. gen. sg. a-na-aš-ša-aš=ša-aš (KUB 35.148 iii 24) ‘lower part of the back (?)’ does not belong here (cf. Kloekhorst 2008 s.v.) Topic of this talk: how to account for *(hx)omeso- (Lat. umerus) next to *(hx)ómso- (all others)?1 1 I am indebted to Francesco Burroni, Jay Jasanoff, Martin Peters, Jeremy Rau, and Zachary Rothstein-Dowden for precious comments and advice, but no endorsement of any specific ideas is hereby implied. 2 §2 Two unattractive solutions 2.1 Two different lexemes *(hx)omeso- and *(hx)ómso-? • This is, obviously, not very attractive. 2.2 Regular development of *(hx)ómso- > Lat. umerus? This assumption implies the insertion of an anaptyctic vowel in the intervocalic sequence *-ms-. • *(hx)omso- > *(hx)omVso- > umerus (so tentatively Weiss 2009:171). The sound law seems somewhat counter-intuitive since no other sonorant + *s cluster in Latin shows the development of an anaptyctic vowel. • It is usually assumed that intervocalic *-ms- gave Lat. -ns- (cf. Leumann 1977:212). Examples for PIE *-ms- are, however, scarce. But cf. Lat. cēnsēre ‘think, suppose’ (if from *√k ̑ ems LIV2:326); Lat. mēnsa ‘table; meal; altar’ (if from *mē ̆ ms-o- ‘meaty’).2 Putative other examples for *-ms- > *-mVs- could be numerus m. ‘number’ (if *nomso-) and cumerum n. ‘chest, box’ (if *komso-), but these are more easily otherwise explained. §3 A third alternative – the prehistories of Lat. cumerum and numerus • The third strategy to solve the umerus dilemma would be to look at the two other words that synchronically rhyme with umerus: o numerus, -ī m. ‘number’ (Naev.+). o cumerum, -ī n., also cumera, -ae f. ‘chest, box, basket’ (Varro+). 3.1 Lat. cumerum n., cumera f. ‘chest, box, basket’ Walde/Hofmann I:306 set up *komeso- and compare Ved. kaṁsá- m. ‘a drinking vessel (made of metal)’ (AV+; *komsó-), und camasá- m. ‘a drinking vessel (made of wood)’ (RV+; *kemesó-). The prehistory of these is, however, far from clear. Semantically, a connection with Ved. √cami ‘sip, slurp’ (LIV²:389f as *√ku ̯ em ‘(hinunter)schlucken, einsaugen’) is plausible, a fortiori in view of Ved. camū́- f. ‘bowl, wooden vessel (for soma)’ (cf. EWAia I:286). But this is unrewarding for Lat. cumera. An alternative account of Lat. cumera will be mentioned below. 2 Note that a post-syncope sequence of *-ms- (or *-m-s- with morpheme boundary) develops to -m(p)s- (as in sūmō ‘I take’, perf. sūmpsī; the town name Temesa [Ov.], Tempsa [Plin.], cf. Weiss 2009:171; or *sm ̥́ h2seh2 ‘scooped mass, heap’ > *samasā > Lat. sam(p)sa f. ‘olive pomace; mass of crushed olives (before and after the pressing)’ [Plaut., Col.; Ital., Cat. sansa], cf. Höfler 2017a:21); in other instances the sequence develops to -ns- (as in intrinsecus, altrinsecus with adverbial *-im as in utrimque, and also in com- + sº as in cōnsul, etc.; cf. Leumann 1977:212). 3 3.2 Lat. numerus m. ‘number’ The word is usually reconstructed as *nomeso- (cf. Walde/Hofmann II:186f; IEW:763f.; de Vaan 2008:419), more recently also as *nomh1-r ̥ (cf. LIV² Addenda s.v. *nemh1-). The latter can be discarded easily, the former requires further attention. • An alleged *nomeso- could be interpreted as a derivative of an s-stem from the root *√nemh1 ‘distribute’ (with *h1 because of Gk. νέµεσις f. ‘retribution’, perf. νενέµη-; cf. also the Latv. accentuation of nem̃ t ‘to take’; cf. Peters 1980:141 n. 94; Peters 1997 [2002]:114 n. 32; Nikolaev 2010 [2011]:84ff.; Malzahn 2010:685; LIV² Addenda s.v. *nemh1-). But in my dissertation on possessive derivatives of neuter s-stems in Latin, Greek, Vedic, Anatolian, and the Celtic languages (Höfler 2017b), I found NO examples of derivatives of this shape. • A type with o-grade root and e-grade suffix did not exist; an alleged *nomh1-es-o- would, thus, have never been a well-formed PIE derivative. • A type with e-grade root and zero grade suffix, on the other hand, is attested throughout all language branches. o An oft-cited example of such a formation is Ved. vatsá- m. ‘yearling, calf’ which shows the e-grade root and zero-grade suffix of the s-stem *wét-os n. ‘year’ and continues the substantivized form of an adjective *wet-s-ó- ‘having a year’.3 I want to propose, therefore, to reconstruct a similarly formed *nemh1-s-o-, implying a development *nemaso- > *nomaso- > *numaso- > Lat. numerus. But of course, I need to justify both the formal (§4) and the phonological side (§5) of this reconstruction.4 §4 A formal account of *nemh1-s-o- The oldest layer of s-stem-derivatives via the possessive suffix *-ó- seems to have had zero-grade in both root and suffix of the s-stem base word (cf. also Vine 2016:136): • *h1réwdh-os n. ‘redness’ (Gk. ἔρευϑος n. Hp., Lat. rōbur n. ‘heartwood’ Enn./Cato+) → *h1rudh-s-ó- ‘having redness’ (Lat. russus ‘red(-haired)’ Enn.+) • *léwk-os n. ‘light’ (Av. raocah- n. ‘light, day’, Ved. rókas- n.) → *luk-s-ó- ‘having light’ (Ved. rukṣá- ‘shining, radiant’ RV 6.3.7). • *ksér-os n. ‘dryness’ (Lat. serēnus 3 ‘clear (of the weather)’ < *kser-es-no-) → *ksr ̥ -s-ó- ‘having dryness’ (Gk. ξηρός ‘dry, dried’ Att.-Ion.). 3 A different analysis of Ved. vatsá- is proposed by Vine (2009:161). 4 Note that PIE also had the possibility to substantivize adjectives by introducing an o-grade vorne im Wort (cf. Schindler 1980:390 note 17 [“Substantivierung durch o-stufige Vr ̥ ddhi”]; Neri 2013:197f; Neri 2016:24 et passim), which could encourage a reconstruction of numerus as *nomh1so-, itself an o-substantivization of adjectival *nemh1só-. But as far as I can see, Latin shows only inherited o-substantivizations of such adjectives (i.e. this word formation process was not productive in the precursor of Latin), viz. (1) Lat. collus m. ‘neck’, PGmc. *halsa- < *kwól(h1)so- ‘the turner’; and (2) Lat. uespa f. ‘wasp’, Lith. vapsà f., russ. osá f. < *(h1/2)wobh(h1)seh2 ‘the weaver’. We would have to assume, then, that *nomh1so- was a PIE formation, in which case the Saussure effect would have deleted the laryngeal (as arguably in *kwól(h1)so-), but then a *nomso- would not have resulted in numerus (see above 2.2) 4 A presumably younger, but still PIE layer shows full-grade in the root of the base word: • *téwh2-(o)s n. ‘power’ (OAv. tǝuuiš- n. ‘violence’) → *tewh2-s-ó- ‘having power’ (Ved. taviṣá- ‘strong’, Lyd. tawśa- ‘id.’) • *pék ̑ -os n. ‘wool’ (Gk. πέκος n.; also Lat. pecus, oris n. ‘farm animals’) → *pek ̑ -s-ó- ‘having wool’ (Lat. pexus 3 ‘woolly’5) • *wéyd-os n. ‘seeing, knowing’ (Ved. védas- n. ‘knowledge’) → *weyd-s-ó- ‘having knowledge’ (PGmc. *wīsa- ‘wise’ > OIc. víss, OE wīs) In a parallel manner we can set up the following derivational chain: • *némh1-os n.‘distribution’ (YAv. nəmah- n. ‘a loan’) → *nemh1-s-ó- ‘having distribution, distributed’ Just as with adj. *wet-s-ó- ‘having a year’ → subst. Ved. vatsá- ‘the one having a year’ > ‘calf, yearling’, one has to assume a “simple” substantivization of *nemh1-s-ó- ‘distributed’ to *nemh1so- ‘the distributed thing’ > Lat. numerus ‘number, category’. Formally, the derivation seems to work. §5 The development of *-ema- > Lat. (*)-uma- A development *-ema- > Lat. (*)-uma- might seem adventurous at first, but there are some not so bad examples that speak in uploads/Geographie/ a-look-over-lat-umerus-pdf.pdf