*sElH-i-

Čop 1956a:25-28 suggests based on Hitt. šalli- and Luw. šalḫitt(i)- an etymological connection with the PIE root *su̯elH- (cf. OHG. swellan). To bring further support to Čop’s etymology, Brosch 2011a:60-64 argues that the labial glide in the initial sequence *su̯- is lost before o (followed by Rieken 2010b:655 and DCL:235). However, this is contradicted by other roots such as the ones of CLuw. šuwāru- < *su̯ór-u- and of Lyd. sfa- (< *su̯eh₂- ← *su̯o- ‘one’s own’); see under the respective lemmata.Oettinger 1979a:549 n. 13 also casts doubt on the etymological connection, since *su̯- is otherwise preserved in other roots (adducing the lexeme šuwāi- c. supposedly meaning ‘bird (?)’, but now considered outdated CHD Š:539). Therefore, the etymology with Skt. sárva-, Gk. ὅλος, Lat. sollus/saluus ‘all, complete, intact’ and OIr. slán ‘complete, sane’ is preferable on phonological grounds, i.e. *selH- (see HEG S:773f. with further literature, followed by EDHIL:710, EDLIL:537 and 572, Yakubovich 2010d:235f., Melchert 2014e:261), although the identity of the laryngeal as *h₂ or *h₃ remains open from the Anatolian perspective.

As pointed out by Oettinger 1979a:550 and Melchert 1994a:55, the preservation of the sequence -lḫ- in Luwian vis-à-vis Hittite -ll- is due to the position of the accent and the different ablaut grades of the root vowel, e.g. sélH-i- (or *sólH-i-) > *salli- vs. *sl̥H-éi̯- > *salḫi-. One should also compare the derivational pattern reconstructed by Rieken 2005a:57f., who suggests that the i-adjective was built on a thematic noun, i.e. *solH-o- → *solH-i-/-oi̯-. For the present reconstruction, we leave the ablaut grade open. One should also notice that an o-stem adjective is probably reflected in Hitt. šalla-, see the comment below under Hittite šalli-.

Beside the reconstructable i-stem based on Hitt. šalli- ‘great’ and perhaps CLuw. *šalḫa(/i)-, it is possible to reconstruct a stem in *-eh₂- continued in Lyc. B sla- (and Lyc. A *hla-) ‘praise, offering’, i.e. a virtual *sElH-eh₂-. Lycian does not preserve the sequence **-lx-, rendering indeterminable which variant was generalized; cf. Lyc. mla- c. For the connection of the Lycian with the Hittite material, see HEG S:771 with further literature.It is also noteworthy that a verbal stem in -āi-/-i(ya)-^(ḫi) derived from the same root might be attested in Hittite (see CHD Š:83), although its interpretation is still uncertain.

[D.S.]

The adjectival formations attested in all non-Anatolian languages can easily be united under a meaning ‘healthy, safe, whole’ (whence ‘all, every’ and ‘entire, complete’ in much the same way as English whole, Danish hel in these meanings from Proto-Germanic *hailaz ‘whole, safe, sound’; compare Gothic hails, German heil ‘whole, safe, sound’). It is more difficult, however, to combine this evidence with the meaning ‘great, big’ attested in the Anatolian languages. Of the three thinkable options (viz. (1) the Anatolian meaning ‘great, big’ is the old one and the meaning ‘healthy, safe, whole’ of the other languages is a [common?] innovation, (2) ‘healthy, safe, whole’ is the old meaning and ‘great, big’ is the innovation, (3) both ‘great, big’ and ‘healthy, safe, whole’ are innovations and go back to a different meaning of the underlying root *selh_2/3-), the second one (2) (in lack of any supporting evidence for (3)) is perhaps the most plausible one, as it is more easily imaginable that ‘healthy, safe, whole’ develops into ‘robust, strong’ and ‘great, big’ (compare Russian zdoróvyj ‘healthy’ and ‘strong, robust’ and ‘big, large’) than ‘great, big’ in turn changing into ‘healthy, safe, whole’ (for which I cannot offer any semantic parallels). For potential remnants of the semantics ‘healthy, whole, robust’ in Hitt. šalla-, see below under Hittite šalli-.

The color of the root-final laryngeal in *selh_2/3- cannot be determined beyond doubt. In most of the o-grade formations listed above, the laryngeal was lost regularly without a trace according to the so-called “Saussure effect” (see Nussbaum 1997a). The formations that do show a trace of the laryngeal, on the other hand, are not informative. The preform of Lat. salvus ‘safe, secure’, Umbr. sal(u)uo-, Osc. salavs, σαλαϝς, for example, can be either *salau̯os (< *sl̥h₂-eu̯o- or *sĺ̥h₂-u̯o- or *sĺ̥h₃-u̯o-) or *salou̯os (< *sl̥h₃-eu̯o-), see Nussbaum 1997a:186f., and for OIr. slán both *sl̥h₂-no- and sl̥h₃-no- are possible. It is only the Hesychian glosses ὁλοεῖται· ὑγιαίνει ‘is sound, healthy’ and ὀλοόν· (…) ἀγαθόν ‘good’ (both obelized by Latte Hsch. ο 598, 617, 621) that might be seen as the continuants of an adjective *holou̯ós from *sl̥h₃-eu̯ó- that would be formally comparable to ταναός ‘outstretched, tall’ < *tn̥h₂-eu̯ó-. Though to account for the o-vocalism in the root of *holou̯ós instead of *halou̯ós < *sl̥h₃-eu̯ó- one would have to invoke an analogical influence from the related adjective Gk. ὅλος, Ion. οὖλος  < *hólu̯os < *sól(h_2/3)u̯o-, which does not make this account superior to its obvious alternative, viz. to analyze *holou̯ós as a remodeled *holeu̯ós and thus a by-form in *-eu̯ó- of *hólu̯os (ὅλος, Ion. οὖλος) of the type seen in κενεός ‘empty’ < *ken-eu̯ó- besides κενός, Ion. κεινός ‘id.’ < *ken-u̯ó-.

The i-stem *solh_2/3-i- (see above under Reconstruction PAnat. *sElH-i- ‘big, great, full-grown’) recurs in Latin solidus ‘solid, whole’ (as per Nussbaum 1999a:401f.). It is best explained as an o/e-acrostatic abstract from which a proterokinetic i-stem adjective (Hitt. šalli-, CLuw. šalḫa(/i)-) was internally derived.

[S.H.]

Ved. sárva- ‘whole, entire, all, every’.Cf., e.g., EWAia II:711 s.v. sárva-

YAv. hauruua- ‘undamaged, whole’, OP haruva- ‘whole, entire’, Khotanese harmä ‘every’.Cf., e.g., EWAia II:711 s.v. sárva-; Bailey 1979:469 s.v. harmä

Gk. Att. ὅλος, Ion. οὖλος (< *sólu̯os) ‘whole, entire, complete’Cf., e.g., GEW II:381 s.v. ὅλος; EDG:1072 s.v. ὅλος

Lat. salvus ‘safe, secure’, Umbr. sal(u)uo-, Osc. salavs, σαλαϝς, lat. salūs, -ūtis f. ‘safety, health’, salūber, salūbris ‘healthy’, soli-dus, soldus ‘solid, whole’, OLat. sollus ‘whole, unbroken, complete’, Osc. sullus, Paelign. solois (dat.pl.), etc.Cf., e.g., EDLIL:537 s.v. salvus, 571 s.v. solidus, 571f. s.v. sollus; Untermann 2000a:652–54 s.v. O.mc.u. salavs, 714f. s.v. O.p. sullus; Nussbaum 1997a

Arm. ołǰ  ‘sound, whole’Cf., e.g., Olsen 1999a:26, EDAIL:531

OIr. slán ‘safe, whole, sane’; perhaps OIr. (h)uile, MW, MBret. (h)oll ‘all, whole, every’Cf., e.g., Schrijver 1995a:323; Nussbaum 1997a:189; Matasović 2009:345

Toch. A salu, B solme ‘complete(ly), altogether’Cf., e.g., Adams 1999a s.v. solme

Alb. gjallë ‘alive, living’, ngjall ‘to heal, revive’Cf., e.g., Demiraj 1997a:182f. s.v. gjallë