*lóh₃u̯-/lh₃u-´

Hittite lāḫu-^(ḫi) and CLuw. lūwa-⁽ⁱ⁾ serve as the basis for the reconstruction of a PAnat. root verb of the ḫi-conjugation, i.e. *lóh₃u̯-/lh₃u-´ ‘to pour’ (Melchert 2011a:129f.). As opposed to the previous literature who reconstructed the root with *h₂ (thus Oettinger 1979a:424, Melchert 1994a:72, Jasanoff 2003a:143, EDHIL:512; cf. HED L:23f.), Melchert 2011a:128-130 argues cogently for *h₃, preserved as voiced/lenis -ḫ-.

As per Melchert 2011a:127f., CLuw. ēlḫā(i)-^(di) ‘to wash (off)’ speaks for the idea that the u in the root must have been seen as an enlargement.Contra HEG L-N:5 and HED L:82, who do not attribute it to the same root as Hitt. lāḫu-^(ḫi). To explain the morphology of ēlḫā(i)-^(di), Melchert 1994a:241, 263, 277f. suggests a lengthened-grade iterative type as in Lat. cēlāre ‘to hide’, back-formed from the zero-grade *l̥h₃-, i.e. *ēlh₃-eh₂-i̯e/o-. Lehrman 1998a:147 likewise assumes a zero-grade, but with a prothetic initial vowel i-, which is not convincing. Sasseville 2021a:87 casts doubt on the PIE reconstruction of such a type involving reanalysis and pleads instead for an inner-Luwian denominative verb in *-eh₂-i̯e/o- derived from a vṛddhi-formation *ēlh₃-o- (→ *ēlh₃-eh₂- c.) with Schwebe-ablaut built to the root *leh₃- (without u-extension). Unless this root variant turns out to be attested in another Indo-European language, it is not necessary to posit its existence beyond Anatolian and, therefore, no initial *h₁ needs to be ascribed for structural reasons.

For the reconstruction of a Proto-Anatolian reduplicated formation, see CLuw. lilūwa-⁽ⁱ⁾.

[D.S.]

The semantic dichotomy of ‘pour (a liquid)’ in the primary verbal stems in Anatolian (i.e., Hitt. lāḫu-^(ḫi) ‘pour; flow’, CLuw. lūwa-⁽ⁱ⁾ ‘pour’, Hittite lilḫuwa-^(ḫi) ‘pour’, CLuw. lilūwa-⁽ⁱ⁾ ‘pour’) and ‘wash, bathe’ in most of the remaining branches in which the root is attested is striking. In turn, the verb ‘wash’ in Hittite (ārri ‘washes’) and Tocharian (A yärnās-) is derived from a root *h₁erh_x- (see LIV²:239) or *h₁erh₁- (see PAnat *h₁órh₁-/h₁r̥h₁-´), which is interpreted by Melchert as reflecting a chronology: “First, there was *leh₃w- ‘to pour’ liquid (only) beside *h₁erH- ‘to wash’ (Hitt. ārr-, TochA yär-). In Core Indo-European the latter was lost, and *leh₃w- (in the form *leuh₃-) was specialized to ‘to wash’” (Melchert 2011a:130f.). While it is certainly true that a meaning ‘wash’ can develop from ‘pour (a liquid)’ (cf. Ved. plutá- ‘flooded’ and Gk. πλυτός ‘washed’), this change may be considered so trivial that it should not be used to distinguish Anatolian and Tocharian from Core Indo-European. And indeed, Melchert himself (Melchert 2011a:127) assumes a meaning ‘to wash (away)’ for CLuw. lā(ḫ)uni-/lā(ḫ)unāi- and ‘to wash’ for CLuw. ēlḫā(i)-^(di) and ililḫā- (the latter two derived from the unextended root *leh₃-; see below), showing that the semantic change ‘pour’ > ‘wash’ also (independently?) happened in Anatolian. In any event, Tocharian AB lu- means ‘rub (off), wipe away’, which is easier derivable from a meaning ‘wash’ than from ‘pour’.

Based on the widespread o-vocalism in the majority of derivatives (see above), the root for ‘wash, bathe’ had traditionally been reconstructed as *lou̯-, *lou̯ə- (e.g., in IEW:692; including “hitt. la-ḫu-uz-zi (lauzi) [sic] ‘gießt aus’”). Subsequently, the identification of the root-final schwa indogermanicum as *h₃ allowed for a modified root shape *leu̯h₃- (thus, e.g., LIV²:418), better in line with commonly accepted PIE root phonotactics. Support for *leu̯h₃- came from Mycenaean, where re-wo-to-ro-ko-wo /lewotro-kʰowos/ might preserve an unmetathesized version of alphabetical Greek λοετρο-χόος ‘pouring water into the bath’. The Anatolian material, however, was not reconcilable with *leu̯h₃-, which is why LIV²:401 assigned Hitt. lāḫu-^(ḫi) ‘pour; flow’, CLuw. lūwa-⁽ⁱ⁾ ‘pour’, Hittite lilḫuwa-^(ḫi) ‘pour’, CLuw. lilūwa-⁽ⁱ⁾ ‘pour’ (and Latin lāma ‘pool, puddle’but see Melchert 2011a:128) to a separate root *leh₂- ‘pour’.

Reconsidering the fate of medial *h₃ in Hittite and Luwian, Melchert 2011a (followed by LIV2+ s.v.) successfully combined the evidence under a novel root *leh₃u̯- ‘pour’. The zero grade *lh₃u-C would have regularly metathesized to *luh₃-C, which explains both CLuw. lūwa-⁽ⁱ⁾ ‘pour’ (qua generalized weak stem) and the new full grade in non-Anatolian Indo-European *leu̯h₃- (cf. Myc. /lewotro-kʰowos/). Non-Anatolian Indo-European *lou̯- may continue *loh₃u- < *leh₃u- independently (cf. LIV2+ s.v. *leh₃u̯- note 1).

It is possible that the *-u̯- in the coda of the phonotactically rather unusual *leh₃u̯- is the remnant of a present stem-forming suffix *-u- that was incorporated into the root. On u-presents see LIV²:15f., Jasanoff 2003a:141–43. CLuw. ēlḫā(i)-^(di) ‘to wash (off)’ might preserve a trace of the bare root *leh₃- (see above under Reconstruction *lóh₃u̯-/lh₃u-´ ‘to pour’).

[S.H.]

Hitt. lāḫu-^(ḫi) ‘pour; flow’, CLuw. lūwa-⁽ⁱ⁾ ‘pour’, Hittite lilḫuwa-^(ḫi) ‘pour’, CLuw. lilūwa-⁽ⁱ⁾ ‘pour’, lā(ḫ)uni-/lā(ḫ)unāi- ‘wash (away)’, ēlḫā(i)-^(di) ‘wash (off)’, ililḫā- ‘wash (off)’see above CLuw. lūwa-⁽ⁱ⁾ ‘pour’, CLuw. lilūwa-⁽ⁱ⁾ ‘pour’, CLuw. lā(ḫ)uni-/lā(ḫ)unāi-, CLuw. ēlḫā(i)-^(di), ililḫā-.

Gk. λοέω ‘wash (the body), bathe’, λούσω, ἔλουσα, λέλουμαι and λούω, -ομαι ‘id.’, λοετρόν, λουτρόν ‘bath’, Myc. re-wo-to-ro-ko-wo /lewotrokʰowos/ (: λοετρο-χόος ‘pouring water into the bath’, subst. ‘the slave who did this’), re-wo-te-re-jo /lewotrejos/, epithet of bathtubs.Cf., e.g., GEW II:138f. s.v. λούω, EDG:872f. s.v. λούω, DMic II:249f. s.vv. re-wo-te-re-jo, re-wo-to-ro-ko-wo, LIV2+ s.v. *leh₃u̯-

Lat. lauō, -ere ‘wash’ and lauō, -āre ‘wash (oneself), bathe’, perf. lāuī, sup. lauātum, ppp. lautus ‘washed; splendid, elegant’ (vulgar lōtus), ab-luō, -ere ‘wash off’, l(au)ātrīna ‘lavatory’, l(au)ābrum ‘bathtub’, pollūbrum ‘wash basin’, etc.; Umbr. vutu /wo:tu/ ‘shall wash’ (< *lou̯etōd).Cf., e.g., WH I:773–5 s.v. lavō, EDLIL:330f. s.v. lavō, Meiser 1986a:155, LIV2+ s.v. *leh₃u̯-

Arm. loganam ‘bathe’, lowanam ‘wash’.Cf., e.g., Klingenschmitt 1982a:117, EDAIL:310f. s.v. loganam, 316 s.v. luanam, LIV2+ s.v. *leh₃u̯-

OIr. loathar, lóthar, n. ‘trough, vat, tub’, MBret. lovazr, NBret. laouer ‘trough’, Gaul. lautro (glossed as balneo ‘bath’ in the Endlicher’s glossary), perhaps OIr. lúaith i, f. ‘ashes, dust’, Welsh lludw ‘ash’, MBret. ludu, MCorn. lusow ‘id.’ (if < *‘substance for washing’).Cf., e.g., Delamarre 2003a:198 s.v. lautron, EDPC:245 s.v. lowatro-, 246 s.v. *lowtu-

PGmc. *lauþra- n. ‘lather’ (ON lauðr n. ‘lather, froth, foam’, OE lēaðor E lather, OS lōthar n. ‘detergent’), PGmc. *laugō- ‘(water for) bathing, washing’ > North Gmc. ‘bath, wash’ (ON, Icel., Norweg. laug, Swed. lög), > West Gmc. ‘lye’ (OE lēaġ, E lye, OHG louge, G Lauge).Cf., e.g., de Vries 1977a:347 s.v. laug, EDPG:329 s.v. *lauþra-

B lu- ‘rub, wipe away’ (pret. lyāwa, pret. ptc. lyelyuwormeṃ), A lu- ‘rub (off)’ (impv. pälwār).Cf., e.g., Malzahn 2010a:852f.

Old Gheg subj. laa ‘wash’, Alb. laj ‘I wash’.Cf., e.g., Schumacher & Matzinger 2013a:980, LIV2+ s.v. *leh₃u̯-