*=k^u̯e

Hitt. =(k)ku ‘now, even, and’, CLuw. =ku(wa) ‘in turn’, Pal. =(k)ku ‘and, also’, Lyc. B =ke ‘and’, Car. =q ‘and’ and Lyd. =k ‘and’ are etymologically equated (cf.EDHIL:483, HEG I-K:599). The particle *=k^u̯e is also found in Hittite takku ‘if, when’ (cf. EDHIL:816) and nikku ‘not?’ (cf. EDHIL:601f., HEG I-K: 598, 600). In most languages, the particle was phonologically reduced due to its enclitic status. The final vowel in *k^u̯e can only be seen in Lycian B =ke and perhaps CLuw. =ku(wa). Otherwise, the vowel was dropped leaving only the labio-velar *k^u̯ (Hitt. =(k)ku, Pal. =ku). The labial element was further eliminated in Lyd. =k. In sum, the particle *=k^u̯e underwent a great amount of erosion in Anatolian.
A coordinating function between sentences and/or nominal/verbal phrases emerges from the whole Anatolian data, although the Proto-Anatolian particle *=k^u̯e ‘and’ seems to have acquired additional functions in the individual Anatolian languages (Carruba 1985b:83). However, a precise study of their functions in most of the concerned languages is hindered by the lack of understanding of the contexts or by their fragmentary states.

[D.S.]

PAnat. *=k^u̯e exhibits three different functions, all of which are also reconstructable for PIE:

1. additive conjunction PIE 1.*=k^u̯e ‘and’ (LIPP II:689–702), attested in Anat., IIr., Gk., Phryg., (?)Messap., Ital., Celt., Germ.
2. conditional conjunction PIE 2.*=k^u̯e ‘if’ (LIPP II:702–706), attested in Hitt., Ved.As this lexeme possibly does not belong here (cf. the comment sub Transmission with ref. to Opfermann 2019a:192–194), it will not be investigated in this lemma.
3. focus particle PIE 3.*=k^u̯e ‘±even’ (LIPP II:442–446), mainly in univerbations, attested in Anat., IIr., (?)Gk., Arm., (?)Phryg., Ital., Celt., Germ.The use as a postfix to derive (universal free choice) indefinites from simple indefinites (cf. Core IE *k^u̯o/i‑=k^u̯e ‘any, every’) can be explained from its function as a (scalar) focus particle, cf. Opfermann 2019a:197–199 with refs.

Contra LIPP II:446, the conjunction PIE 1.*=k^u̯e ‘and’ may be connected with the particle 3.*=k^u̯e ‘±even’ (cf. already Grdr² II.3:1004–1008),In this case, Dunkel’s etymology of PIE 1.*=k^u̯e ‘and’ (< *=sk^u̯e ← *sk^u̯‑o ‘(s)he follows’, an injunctive middle form of PIE *sek^u̯- ‘to attach’, cf. LIPP II:701f.) has to be rejected.and it may furthermore be proposed that the former developed out of the latter – contra, e.g., LHS:475 (cf. LIPP II:445f. with further lit.).This development is substantiated by the fact that PIE 1.*=k^u̯e originally connected opposites or parts of a whole (cf. Gonda 1954a:189–192), cf., e.g., Lat. dies noctes=que “day and night” ≈ “at daytime, even at night” (Plaut.Most.22+).
Under this scenario, the particle PIE 3.*=k^u̯e ‘±even’ would be primary. It may plausibly be traced back to an indefinite instr. sg. PIE *k^u̯e‑h₁ ‘somehow’ of the pronominal stem PIE *k^u̯o/i- (as per LIPP II:446). For a semantic parallel, cf. the emphatic particle PAnat. *=ma < ‘somehow’.Another, but less likely, possibility is an endingless loc. sg. PIE *k^u̯e ‘somewhere’.
According to Berenguer Sánchez 2000a:482, the Anatolian evidence shows that the change from the particle PIE 3.*=k^u̯e ‘±even’ to the conjunction PIE 1.*=k^u̯e ‘and’, viz. its lexicalization was not finished completely at the time when Anatolian split off from the parent language. Thus, the evidence attests not only additive use but also emphatic function, cf., e.g., Lyc. B =ke=pe (above), which emphasizes change of subject, CLuw. =ku(wa) ‘in turn’ etc.
To sum up, the lexeme under investigation may be reconstructed as PIE 1.*=k^u̯e ‘and’ and traced back to 3.*=k^u̯e ‘±even’ << *k^u̯e‑h₁ ‘somehow’.

It is assumed above (sub Semantic reconstruction) that the additive/completive conjunction PIE 1.*=k^u̯e ‘and’ developed from the particle PIE 3.*=k^u̯e ‘±even’ as early as in PIE times and, according to Berenguer Sánchez 2000a:482, presumably around the split-off of the Anatolian branch. Furthermore, it is proposed that PIE 3.*=k^u̯e ‘±even’ goes back to an instr. sg. form PIE *k^u̯e‑h₁ ‘somehow’. The unexpected short vowel (not †=k^u̯ē < *=k^u̯e‑h₁) which Dunkel (in LIPP II:446) cannot explain may be due to irregular loss of the laryngeal in posttonic/clitic position.
Therefore, PAnat. *=k^u̯e ‘and; even’ can ultimately be traced back to instr. sg. PIE *k^u̯e‑h₁ ‘somehow’.

The enclitic additive conjunction PIE 1.*=k^u̯e ‘and’ coordinates not only nouns (and verbs) but also clauses in many IE languages (cf. above on Anatolian); however, it is most likely that its primary function was the connection of noun phrases, cf. LIPP II:699f.On Old Avestan, cf. d’Amato 2015a:6.This observation corresponds to the assumption that PIE 1.*=k^u̯e ‘and’ developed out of the focus particle PIE 3.*=k^u̯e ‘±even’, which is attached primarily to (pro‑/nominal) phrases rather than to complete sentences.
Syntagmatically, it can be observed that PIE 1.*=k^u̯e is normally placed either after every constituent of the phraseCf. Gk. πατὴρ ἀνδρῶν=τε θεῶν=τε “father of men and gods” (Il.1.544+).or after the last one,Cf. Lat. dies noctes=que “day and night” (Plaut.Most.22+).cf. LIPP II:694–699.

The enclitic conditional conjunction PIE 2.*=k^u̯e ‘if’, contra LIPP II:705f., most likely did not develop from PIE 1.*=k^u̯e ‘and’, cf. Opfermann 2019a:192–194. Opfermann points to the mismatch between the attested word order in the relevant additive syntagms, viz. *A(=k^u̯e) B=k^u̯e “A and B”, and conditional sentences, viz. [protasis=k^u̯e]–[apodosis] “if [protasis] then [apodosis]”.This order of clauses is not only attested in the context of conditional 2.*=k^u̯e, but also typologically widespread, cf. Opfermann 2019a:192f. and the references ibid. n. 12.Furthermore, he bases himself on the fact that PIE 1.*=k^u̯e ‘and’ originally was used to connect (pro‑/nominal) phrases and not sentences (cf. above).

[A.O.]

PIE additive conj. 1.*=k^u̯e ‘and’, cf. the transmission of the individual branches/languages below
LIPP II:689–702, IEW:635f., EIEC:20, Opfermann 2019a:191f.

(?)PIE conditional conj. 2.*=k^u̯e ‘if’,Since this function is not attested in the minor Anatolian languages, its etymology will not be discussed here. Cf. the most recent, yet short treatment of PIE 2.*=k^u̯e ‘if’ by Opfermann 2019a:192–194 (< ‘when’ either < loc. sg. *k^u̯e ‘somewhere’ or ← instr. sg. *k^u̯e‑h₁ ‘somehow’) who separates PIE 1.*=k^u̯e ‘and’ and 2.*=k^u̯e ‘if’.cf.
Ved. =ca ‘if’ (EWAia I:520f., Grassmann 1996a:428f.)Goth. ni=h ‘if not’ is also used as (negated) conditional conjunction and is thus taken as evidence for PIE 2.*=k^u̯e ‘if’. Lat. (abs)=que is usually adduced here (cf., e.g., LIPP II:703 with n. 1) but actually does not belong here, cf. Cabañeros 2016a:678.
LIPP II:702–706, Opfermann 2019a:192–194

PIE focus particle 3.*=k^u̯e ‘±even’, cf.
Hitt. focus particle =(k)ku ‘even’ in, e.g., apii̯a=kku ‘even then; even there’ (EDHIL:483f., HED 4 K:203, HEG I A–K:47, 598f., HW² 1 A:184f., Opfermann 2019a:198 n. 31)
Core IE indef. pron. *k^u̯o/i‑=k^u̯e ‘every, any’ (LIPP II:443f., Opfermann 2019a), cf. Arm. o=k‘ ‘any(one)’ (EDAIL:522f.), Lat. quis=que ‘everyone’ etc. (WH II:410, EM:559f., EDLIL:506, OLD²:1702), PItal. adv. *k^u̯om=k^u̯e ‘each time, always’ (Lat. indef. postfix =cum=que ‘=ever’, Umbr. adv. pum=pe ‘each time’ [IgT Va.3, 10]) (WH II:152, EM:555, EDLIL:152, OLD²:516, Untermann 2000a:601f.), Goth. ƕaz=uh ‘everyone’, etc. (EDPG:— , GED:198f.)
PIE neg. *ne=k^u̯e ‘not’ (LIPP II:444, Opfermann 2019a:198f.), cf. Hitt. ne/i=kku ‘not’ (EDHIL:601f., HED 7 N:79, HEG II,1 L–P:325f., Opfermann 2019a:198f. with n. 33), PItal. *ne=k^u̯e ‘not’ (Lat. 1.ne=que, ne=c, Osc./Umbr. neip) (WH II:401f., EM:432, EDLIL:403, OLD²:1288, Untermann 2000a:489f.), OIr.(+) na=ch ‘not’ (EDPC:287, LEIA-N:14), Goth. ni=h ‘not’ (EDPG:—, GED:374)
LIPP II:442–446, IEW:636, Opfermann 2019a

PIIr. additive conj. *=ća ‘and’, cf. Ved. =ca,According to LIPP II:689 (with n. 1), Ved. =śca is attested in Vā́yav Índra=śca ‘o Vāyu and Indra!’ etc.OAv. =cā, etc.
EWAia I:520f., Grassmann 1996a:427f., Bartholomae 1904:563–575, LIPP II:691

PIIr. rel. syntagm *Hi̯á- ká‑=ca ‘whoever’, cf. Ved. yá- ká‑=ca ‘(rel.) whoever (RV+); (indef.) anyone (AV+)’,On the scanty and late attestations of OInd. ka/i‑=ca ‘someone’ (cf. Core IE *k^u̯o/i‑=k^u̯e ‘everyone’ above), cf. AiGr III:571.Av. ya- ci‑=cā ‘whoever’According to AiGr III:571, YAv. ciš=ca ‘everyone’ (hapax, HN 2.11) is a recent formation. It may furthermore be doubted that PIIr. *Hi̯á- ká‑=ća ‘whoever’ contains Core IE *k^u̯o/i‑=k^u̯e ‘everyone’ but is rather an extension of the yet free syntagm PIE *h₂i̯ó- k^u̯i- ‘whoever’ (> Gk. ὅστις ‘id.’, → Lat. quisquis ‘whoever [rel.]; anyone [indef.]’, Hitt. ku̯iš ku̯iš ‘whoever’), cf. Opfermann 2018a:126.
EWAia I:520f., Grassmann 1996a:429, AiGr III:572, Bartholomae 1904:425, 426, 574, LIPP II:443 with nn. 5–7

PGk. additive conj. *=k^u̯e ‘and’, cf. Myc. =qe, Gk. =τε (Hom.+)According to LIPP II:689 (with n. 2), Gk. =στε is attested in *Ζεῦ […] Ἠέλιέ=στε (→ Ζεῦ […] Ἠέλιός=τε, Il.3.276f.).
GEW II:862f., GEW III:183, DELG:1098, EDG:1457, LSJ:1763–1765, LSJSuppl:291, LIPP II:691 with n. 13

(?)Gk. adv. τότε, Aeol. τότα, Dor. τόκα ‘at that time, then’ etc.
GEW II:915, DELG:1127, EDG:1497, LSJ:1808, LSJSuppl:294

Phryg. additive conj. =ke, =key, =κε ‘and’
LIPP II:691 with n. 15, 693 with n. 23a

(?)Phryg. rel. pron. ιος=κε ‘whoever’
LIPP II:443 with n. 3

Messap. additive conj. =θi ‘and’
LIPP II:692 with nn. 18f.

PItal. additive conj. *=k^u̯e ‘and’, cf. Lat. =que, Falisc. =cue, Ven. =kve, (?)Osc. p]e (ST Sa 27), Lat. 2.ne=que ‘and not’On non-enclitic Ven. ke ‘and’, cf. PAnat. *k̑ō ‘here(to)’.
WH II:152, 401f., EM:555, EDLIL:506, OLD²:1288f., 1701f., Untermann 2000a:520, LIPP II:691f. with nn. 15a–17, 693

PCelt. additive conj. *=k^u̯e ‘and’, cf. Celtib. kue, OIr. =ch, etc.
EDPC:175f., LEIA-C:89, LIPP II:692 with nn. 20f.

(?)Celtib. rel. pron. os=kuez ‘whoever’
LIPP II:443, 693, 694

Goth.(+) additive conj. =(u)h ‘and’, ja=h ‘id.’, ni=h ‘and not’, OHG no=h ‘id.’
EDPG:—, GED:210, 265f., 374, EWAhd VI:1017–1019, LIPP II:692, 693, Willson 2016a