*melit-

The etymological connection between Hitt. milit(t)- n. ‘honey’, CLuw. mallit- n. ‘honey’ is undisputable, cf. HEG L-M:207f., HED M:153, EDHIL:580f. The weight of the reconstruction of the Proto-Anatolian stem *mélid- is given to the attested Cuneiform Luwian stem mallit-, where the fortition of *-l- (> -ll-) after accented /e/ according to Čop’s law (cf. Melchert 1994a:74) as well as the lenition of the stem final *-t- (> -d-) between unaccented vowels indicate that the stress was on the root (but. cf. now Lyc. A *mltt-, showing the outcome of a different stem).

The Palaic stem *malit- must also have undergone the lenition *-t- > -d- between two unaccented vowels, which is taken to be a Proto-Anatolian development. Melchert 1994a:221 interprets the first vowel -a- in Palaic malitannaš, which cannot reflect *e directly, as mute, i.e. /mlidannas/. If so, we must assume on account of the Proto-Anatolian lenition of the alveolar in *mélid-, which is expected to yield Pal. *melid-, that the syncope cannot be original. Eichner 2010a:48 posits a denominative *mlitāi- as the base of the abstract noun *malitātar, in order to account for the mute vowel -a-, and takes the lenition (*malittātar- → *malitātar) to be analogical to the unattested Palaic word ‘honey’ (PIE *mélit-).

The attested adjective milittu- in Hittite indicates that the stress was on the suffix, since the fortis *-t- does not lenite and the root syllable has an i-vocalism (Rieken 1999a:155f.), i.e. PAnat. back-projection *mlít-u- ‘sweet’. The Hittite attested base noun mili(t)t- and its derived verb mili(t)tešš- show inconsistencies in the spelling of the alveolar. The fortis spelling is probably due to the influence of the adjective m(a)littu- ‘sweet’ (cf. Melchert 1994a:140f.).Kloekhorst (EDHIL:580f.) argues unconvincingly that the stress was on the root and that an i-umlaut (i < *é) in Hittite nevertheless occurred. The theory with Hoffman’s suffix (*-h₃on-) explaining the lenition in the Palaic word malitannaš through the sound change */Th₃/→ */D/ should be abandoned, since it is explicable as an inner-Anatolian development (for the literature on this theory see Mayrhofer 1986:144).

[D.S.]

PIE *mélit- means ‘honey’, cf. IEW:723f., EIEC:271, Watkins 2011a:55.

A. Word formation

Morphologically, PIE *mélit- n. is comparable with PIE *(h₁)álb^hit- n. ‘barley groats’ (Gk. ἄλφι, pl. ἄλφιτ-α, Alb. elb, -i ‘barley’), Ved. *hárit- n. ‘golden-yellow color’ (presupposable as derivational base of Ved. hárita- ‘golden yellow’ [YAv. zairita- ‘id.’], hárit-vant- ‘id.’, and harít- ‘id.’, f. ‘golden-yellow mare’), and Hitt. šeppitt- n. ‘a sort of barley’ (cf. Rieken 1999a:158f., EDHIL:744f.). A further analysis of these forms is difficult; at least one of them, however, allows a relatively clear explanation (Schaffner 2014a:375): sbst. Ved. *hári-t- n. ‘golden-yellow color’ < *g̑^hélh₃i-t- : adj. Ved. hári- ‘yellowish, pale’ (Av. zairi- ‘yellow’) < *g̑^hélh₃i- (EWAia II:805f.). And accordingly PIE sbst. *méli-t- n. ‘the yellow one’ : adj. *meli- ‘yellow’ (or sbst. *meli- ‘yellow color’?) as basis of PCelt. *meli-no- ‘yellow’; cf. Ved. hári- ‘yellowish, pale’ : hari-ṇá- m. ‘yellowish animal, gazelle’ (RV+). The neuter stems in *-i-t- would thus have been derived either from underlying i-stem adjectives by means of an individualizing suffix *-et-/*-t- (as seen in the type Gk. γυμνής, γυμνητ- ‘soldier with light weapons, infantryman’ : Gk. γυμνός ‘naked, unclothed’, cf. Schaffner 2014a:375 with further examples) or from underlying i-stem abstract nouns by means of an instrumental ending *-t (Pinault 1980a:32f., Peters 2002a:371 n. 64, Widmer 2005a:196f.); cf., e.g., Av. huzāmi- ‘good birth’ → instr. *huzāmi-t ‘with good birth’ → adj. huzāmi-t- ‘being with good birth, having good birth’, Gk. χάρι- ‘charm, grace’ → instr. *χάρι-τ ‘with charm/grace’ → Xάρι-τ- ‘provided with charm/grace; Charites’ (Widmer 2005a:197).

B. Inflection

Three proposals for the inflection of the original PIE paradigm have been made in the literature, which can be briefly summarized as follows:

1. Heteroclitic neuter i/n-stem: nom./acc. sg. *mél-i(t), gen. sg. *mel-n-és, dat. sg. *mel-n-éi̯, etc. For the type, cf. Ved. ásthi, gen. sg. asthnás n. ‘bone’ (Benveniste 1935a:7ff., IEW:723f.).
2. Ablauting paradigm: nom./acc. sg. *mélit, gen. sg. *mlit-és, dat. sg. *mlit-éi̯, etc. (Steinbauer 1989a:86f., EDHIL:580, Tremblay 2010a:250).
3. Non-ablauting paradigm with constant root accent: nom./acc. sg. *mélit, gen. sg. *mélit-os, dat. sg. *mélit-ei̯, etc. (Klingenschmitt 2005a:349).

There are arguments in favor of and against each of these proposals, which makes it difficult to prefer one of them.

C. Root etymology

Etymologically, PIE *mélit- can be connected with PCelt. *meli-no- adj. ‘yellow’ in MW melyn, OBret. milin, OCorn. milin, Corn. melyn (cf. IEW:724, Klingenschmitt 2005a:349, EDPC:263). This implies a basic meaning *‘the yellow one’ (or: *‘the one provided with yellow color’) for PIE *mélit-. As a semantic parallel, cf. PGerm. *χunanga- m. ‘honey’ (ON hunang, OHG hona(n)g) < *kn̥h₂enk̑ó- ‘the yellow one’ : Gk. Att. κνηκός, Dor. κνᾱκός ‘yellow, brownish yellow’ < pre-PGk. *kn̥h₂kó- (EIEC:271, EDPG:255f.). Alternatively, PIE *mélit- could have been a collective formation *mél-i n. ‘complex of dark color’ (with later t-extension), which was based on the PIE root *mel- ‘dark in color, dirty’, cf. Gk. μέλας ‘black’, Ved. mála- n. ‘dirt, mud’ (Oettinger 1999a:207f., Oettinger 2002c:363 n. 2). As for the semantics, Oettinger 1999a:207f. argues that filled honeycombs are dark in their natural state and that a ripe honeycomb of wild bees is a great complex of single honeycombs.

[S.S.], [T.S.]

Gk. μέλι, gen. μέλιτος n. ‘honey’ (Il.+)Cf., e.g., GEW II:200f., EDG:925.

Lat. mel, gen. mellis n. ‘honey’ (Pl.+)Cf., e.g., EDLIL:370, WH II:61f.

Arm. mełr, gen. mełow (u-stem) ‘honey’Arm. mełr is normally interpreted as a continuation of PIE *mélit-, with secondary u-stem inflection due to a contamination with the (almost) synonymous stem PIE *méd^hu- n. ‘honey, honey wine’ (NIL:467f. s.v.) in the prehistory of Armenian (cf. IEW:723, GEW II:201, Matzinger 2005a:59 n. 256, Praust 2005a:154 n. 36 with lit.). There is, however, also the possibility that Arm. mełr directly continues PIE *méd^hu-. If one accepts the sound law PIE *-Vd^hV- > PArm. *-VδV- > Arm. -VrV- (proposed by Jasanoff 1979b:144ff., cf. also Praust 2005a:150–154), then pre-PArm. *med^hu-r (with secondary r-extension in the nom./acc. sg. of neuter u-stems, cf. Arm. cownr ‘knee’ < *g̑onu-r, Arm. artawsr ‘tear’ < *drak̑u-r; see Klingenschmitt 1982a:18 n. 11) could develop via PArm. *meδur to *merur, which – with dissimilation r–r to ł–r (cf. *erbayr > Arm. ełbayr ‘brother’ and *vaδar [loanword from NW-Iranian *vaδar-, cf. YAv. vadar- ‘sharp weapon’] > *varar > Arm. vałr ‘short sword, dagger’) and loss of u in the final syllable – resulted in Arm. mełr (Cowgill apud Matzinger 2005a:59 n. 256, Praust 2005a:154).

OIr. mil f. (i-stem) ‘honey’, OW, MW mel, OBret., MBret., MoBret. mel, OCorn. melCf., e.g., de Bernardo Stempel 1999a:173, EDPC:263.

Goth. miliþ n. ‘honey’ (acc. sg. miliþ haiþiwisk Mc 1.6); OE mildēaw, meledēaw n./m. ‘honeydew’, OS milidou n. ‘id.’, OHG militou n. ‘id.’Cf., e.g., Casaretto 2004a:435f., EDPG:363, EWAhd VI:419–421.

Alb. mjaltë ‘honey’Cf., e.g., Demiraj 1997a:270f.