ܓܘܫܦܐ gwšpʾ and ܓܘܼܫܦܲܩܬܵܐ gwšpqtʾ gušpaqtā (fem.) γαυσάπης gausapa; a rug
Allotropes: ܟܘܼܫܦܵܐ kwšpʾ kušpā, ܟܘܼܫܦܬܵܐ kwšptʾ kušptā (fem.) outer garment; ܓܘܼܫܡܩܵܐ gwšmqʾ gušmqā fillet, head-band
This word is problematic: first, it is not certain that all the above-mentioned allotropes are actually allotropes of the same word: e.g., only PS Suppl. claims that kwšpʾ is an allotrope of gwšpʾ, whereas LS 350b does not even mention the hypothesis of an Iranian origin of kwšpʾ. Further difficulties arise on semantic grounds: if the relationship between the meanings "rug", "cloak"and "carpet" is more or less clear, more dubious is then the semantic connection with the meaning of gwšmqʾ, namely "fillet, head-band". Another problem is the chronology: kwšpʾ is attested very early and also appears in Talm. Aram. kwšpʾ "rug", unlike gwšpʾ, which appears only in later lexicographers; nevertheless, gwšpʾ seems connected with Gr. γαύσαπος or γαυσάπης (Strabo 5, 1, 12: γαύσαποι ed. Lasserre and Kramer; γαύσαπαι ed. Meineke) the name of a rough hair-cloth, and Lat. gausapa, ae (also gausape, is), whose meanings are: 1) “étoffe épaisse et à long poils, introduite à Rome vers l'époque d'Auguste; vêtements, lingerie faits avec cette étoffe; 2) perruque” (Ernout – Meillet 1959, 268).
In my opinion, we have to distinguish two basic forms: gwšpʾ (with the allotrope kwšpʾ: further cases of allotropes involving devoicing or sonorization of the same plosive in § 11.3.1), meaning "rough hair-cloth", and gwšmqʾ, meaning "head-band, ear-cover". These two forms perhaps were originally different and only later formally merged, at least as regards the initial sequence gwš- (see below). As regards the etyma, it seems that gwšpʾ/kwšpʾ is related in some way to the OP word that Hinz (1975, 106), following Mayrhofer (OnP 8.733), reconstructs as *gaušapa- "Rinder haltend", "keeping/holding cattle" (see also Gershevitch 1969b, 198: OP *gau-šaba- "he who assembles cattle"). The etymon of this OP word is disputed, but it seems to imply a reference to cattle: perhaps the cloth named gausapa was made of cow's hair, or was used for herdsmen's garments? In any case, the first element of the word would be OIr. *gaw- cow, bull, IE *gweH3us cow. A different etymon has been proposed by Sims-Williams (2000, 200): this scholar traces back the corresponding Bactrian word κωσοβο "blanket" to OIr. *kaučapa- and compares OInd. kaucapa-ka- woollen blanket (Turner 1966, addenda et corrigenda ad no. 3490 *kōjava- fleecy cloth; OInd. kaucapa-ka-, attested in the Arthaśastra, is not mentioned in Monier-Williams 1899). However, many details remain unclear. If we accept the etymon suggested by Sims-Williams, it remains to explain the initial voiced plosive in Gr. γαύσαπος and in Syr. gwšpʾ. The allotropy gwšpʾ/kwšpʾ is easy explainable with similar cases of allotropes of Syriac words involving devoicing or sonorization of the same plosive (see § 11.3.1): but, if we believe that kwšpʾ comes from OIr. *kaučapa- and that gwšpʾ shows a secundary Syriac sonorization, we are forced to conclude that Gr. γαύσαπος was borrowed from Syriac, and this conclusion is weakened by the fact that the Gr. word is attested since Strabo (1st cent. BC). On the other hand, if we hypothesise a case of reborrowing (kwšpʾ from OIr. *kaučapa- and gwšpʾ from Gr. γαύσαπος), the relationships between the OIr. and the Gr. forms remains unclear. Moreover, the origin of Gr. γαύσαπος is disputed: it is certainly a loanword, but some scholars (e.g. Frisk I 292) believe that the word was borrowed not from Iranian, but from some Indo-European Balcanic language.
As regards gwšmqʾ "head-band, ear-cover", it seems that the word, especially on account of its meaning, is a loanword from MP *gōš-pēč, cf. MP gōš ear (CPD 37) and MP pēčīdan, pēč- twist, entwine (ibid. 68); NP gōš-pēč a cap covering the ears against cold (Steingass 1104): if the etymon is tenable, we need to suppose that -m-is an accidental mistake for -p-. Note also that in gwšmqʾ the first element represents an Iranian form connected with OP gauša-, Av. gaošā- "ear", whereas the formally identical initial sequence in gwšpʾ/kwšpʾ represents OIr. *gaw- cow, bull (or some other Indo-Iranian word, if we accept the etymon proposed by Sims-Williams) ● gwšpʾ, gwšpqtʾ BA 2824; BB 476, 21; kwšpʾ, kwšpqtʾ am 4, 338, 6; Bh ad Jos 4, 10; 2Rg 8, 15; BA 4677; BB 759, 5; 884, 1; 1078, 4; gwšmqʾ MAMQ 91u; BB 476, 15 ◆ LS 137a, b; 350b; Lagarde GA 27, 66; PS 1846; PS Suppl. 71; 173; Duval index pers. 217