The Common Slavic name for horse *koń', with a probably older, yet geographically more limited variant *komoń', has so far no generally accepted etymology. Given the great importance of this animal in the prehistory and the early history of the Indo-European and other peoples of Eurasia, this sets a problem not only for linguists, but also for historians and archeologists. The PIE word for horse, *ekuos, attested among all other branches of IE linguistic family, originally must have been common to the Slavs, as it was to their Baltic, Iranian and German neighbors, but at a later moment - which is hard to determine precisely, although we can assign it to a time before the disintegration of Slavic linguistic unity around the middle of the first millennium A. D. - for this inherited designation the new one *ko(mo)ń' was substituted, either as a lexical innovation made by the Slavs themselves or as a borrowing from another language. Under the entries *komoń' and *koń' of the Moscow diction...ary, where their continuations in Slavic languages are respectively listed, O. N. Trubachev gives a survey of previous etymological proposals and rejects all of them in favor of his own explanations. According to him, *komoń' is a Slavic onomatopoeic creation imitative of neigh, while *koń' is a loan-word, going back to Celtic *konkos/kankos 'horse' (originally 'springer'?) through an intermediate form *konk', which was presumably understood as a diminutive in -'k'' and consequently shortened. Apart from the facts that the word in question is scarcely attested in comparison with two others Celtic designations for horse, *equo- (> Olr. ech, Gall, epo-) and *marka-, and that in Slavic mouth it should have been reflected as *kok'', and not as *kon'k'', the very separation of both forms, *koń' and *komoń', seems unmethodical. With more reason Gamkrelidze and Ivanov recur to the old proposal connecting *koń' via *komoń' ( **kobn-?) with *kobyla and further with Lat.-Gall. caballus, OTur. keväl, Pers. kaval etc., and trace all these forms back to a migratory term of unknown origin and etymology. Nevertheless, other possibilities of interpreting *komoń' (presumed to be an earlier form, which, through haplology, resulted in *koń') are not to be excluded; here a new one is envisaged, analyzing it on a level one may conventionally call Proto-Slavic, which was, as I have argued elsewhere (Studia Etymologica Brunensia 2, Prague 2003, 267 sq.), characterized by the presence of inherited PIE lexems that in a later, Common Slavic stage fell out of use, as well as by the vitality of archaic morphological proceedings, especially the nominal composition. Hence the word *komoń' can be analyzed as a compound *ko-moń', the second element being PIE *mon- 'nape, neck, mane', cf. OInd. mányā- 'nape, the edge of horse's ear', Lat. monile 'necklace, mane', Olr. muin- 'neck', OHG. mana, NHG. Mähne 'mane', E. mane etc., whereas the first element *ko- may be identified as the pronominal stem intensifying or (pejoratively) varying the meaning of the word to which it is prefixed, thus *kuo-monio- 'marked by mane', or 'less-maned', in this case if compared with another species of horses (possibly a domestic one with the European wild horse, the long-maned tarpan that until the 20th century inhabited the south-Russian steppes). There is, however, a difficulty in supposing for Proto-Slavic *moni- the meaning 'mane' in face of Common Slavic *monisto 'necklace', where it obviously designates 'neck' (the second element °sto is, rather than understood as a rare suffix, to be identified with Gk. detós, OInd. ditá- PIE *dH1tó- 'tied', with the regular loss of interconsonantal laryngeal, cf. *d''kter- 'daughter' = Gk. thygátēr, OInd. duhitár- PIE *dhughH2ter-, and with d > s before a dental). Besides, Slavic has another word for 'mane' with a good PIE pedigree, *griva. If departing from Proto-Slavic *moni- in the sense of 'neck, nape', an alternative interpretation of the first component in *ko-moń' > can be proposed, relating it to Gk. pókos m.,pékos n. 'fleece', Arm. asr, asu 'id.', thus PS1 *pko-monio- 'fleece-necked', i. e. 'maned', the p being regularly lost before a consonant in the later Slavic; for the zero-degree in the first syllable of a compound cf. Gk. kteís, ktenós 'comb' vs. Lat. pecten 'id.' and OInd. ksu°, Av. fšu° as a composition form of *paśu-, Av. *pasu- PIE *peku- '(fleeced) cattle', derived from the same root *pei-, and for the centum-treatment of PIE *k in Balto-Slavic other important terms of cattle breading may be cited: Slavic *korva 'cow' and Baltic peku- 'cattle', both belonging to an ancient stratum of vocabulary, the latter reflecting the very word PIE *pek-u-. Such a description of horse as 'fleece-necked' may have originated as a poetic epithet, used primarily with some common designation for the animal, most probably along with the continuation of PIE *ekuos, and eventually in place of it. Not unlike *ekuos furnishing a part of several formulas traceable back to the PIE poetic language (*ekuo- ōkú- / ŕgró- 'swift horse', etc.), *ko(mo)ń' must have belonged to the formulaic repertoire of Common Slavic heroic poetry, in view of correspondences between the Serbian and Russian epics, such as ORuss. br'zyj *kon' 'swift horse' (4 instances in the 12th century poem of Igor's campaign), dobr(yj) *kon' 'good horse' (in the biliny) : Serb. brz(i) konj / dobar konj (from the beginning of the tradition, cf. the long-verse poem Bogišić 49°, recorded in 16th century, where both are alternatively used).
