27 January 2016

The bear and the marten (updated)

The common IE word for 'bear', found in most branches (Hittite hartagga-, Celtic *arto-, Greek árktos, Armenian arǯ, Avestan arša-, Latin ursus, Sanskrit ŕkṣa-, etc) is usually reconstructed as  *h₂ərətk´o-, where the "thorny cluster" tk´ (þ in older reconstructions) has various reflexes t, s, š, kt, kṣ.

However, I regard the Hittite segment -gga- as a suffix like the one found in Turkic qarsaq 'steppe fox' < Altaic *karsi 'marten'. This would leave us with a protoform *h₂ərəC-ko-, where C would represent a sibilant (possibly palato-alveolar) affricate like the one of Caucasian *χHVr[tɕ’]V 'marten; otter'1 (a Nakh-Dargwa isogloss). Probably also Yeniseian *χa(ʔ)s (~ k-) 'badger' belongs here.

Ignoring external data, some Indo-Europeanists have proposed a link between 'bear' and an IE verb 'to destroy'. Interestingly enough, this meaning is represented in Caucasian by *HarGG(w)V, which I'd link to Altaic *jàrgi (~ -o) 'wild beast of prey' and possibly also to Arabic ʕurāʒ- 'hyena(s)', proposed by Nostraticists as cognates of IE 'bear'.
1 Using his etymological instinct, Bengtson links this to Basque hartz 'bear', an IE loanword, most likely from Celtic.

03 January 2016

Etruscan φersu 'masked character (in games)' (updated)

Latin persōna 'theatre mask' is an Etruscan loanword from an unattested *φersu-na, in turn related to φersu 'masked character (in games)'. This word can be analized as a derivative *φers-u from *φers 'husk', an agricultural term with correspondence in the Greek theonym Perséphonē (Etruscan Φersipnei, Latin Proserpina), attested on several Attic vases from the 5th century BC as Persóphatta, P(h)erséphatta, Pherréphatta.

Rudolf Wachter analyzes this is a compound whose second member would be derived from I*-gʷhn-t-jā < *gʷhen- 'to beat, to kill', and the first one would be linked to Sanskrit pará- 'sheaf (of corn)', Young Avestan parša 'sheaf', thus literally meaning 'sheaf-beater', i.e. 'threshing maiden'1. Michael Weiss (in a personal communication to Wachter) also links Latin porrum and Greek práson 'leek' to the Indo-Iranian word.

However, like most Indo-Iranian lexicon related to agriculture, this happens to be a substrate loanword from the language spoken by BMAC people2. In fact, as noticed by De Vaan, the meaning 'sheaf, bundle' can hardly be reconciliated with either 'leek' or 'ear of corn'3.

In my opinion, a better cognate would be Kartvelian *purtś 'husk, foliage' (Georgian purcel 'leaf, foliage', Megrel purča 'chaff, husk', Laz purča 'sweet corn ear', purčumale 'a k. of weed'), *prtś-wn- 'to husk, to scale'4, linked by Alexei Kassian to Hittite paršdu 'leaf, foliage5.

This reminds me of Slavic *proso- 'millet' < *ps-o- 'ear of corn', regarded by Georg Holzer6 as a substrate loanword from a language he calls Temematic (Temematisch in German) on account of its proposed sound correspondences with PIE and where *r̥ ro. According to him, the source would be IE *bhar(e)s 'a k. of cereal (milletbarleyspelt)' (Latin far, farris), a remnant of the languages spoken by the Neolithic farmers who colonized Europe from the Near East.
1 R.S.P. Beekes (2010): Etymological Dictionary of Greek, p. 1178-1179.
2 M. Witzel (2003): Linguistic Evidence for Cultural Exchange in Prehistoric Western Central Asia, in Sino-Platonic Papers 129, p. 33.
3 M. De Vaan (2008): Etymological Dictionary of Latin and the Other Italic Languages, pp. 481-482. 
4 A. Kassian (2009): Anatolian lexical isolates and their external Nostratic cognates, in Orientalia et Classica, §48.
5 Wrongly translated by some authors as 'sprout, sprig'. See A. Kloekhorst (2008): Etymological Dictionary of Hittite, pp. 645-646. 
6 G. Holzer (1989): Entlehnungen aus einer bisher unbekannten indogermanischen Sprach in Urslavischen und Urbaltischen, §2. See also F. Kortland (2010): An Indo-European substratum in Slavic?, in Studies in Germanic, Indo-European and Indo-Uralic, pp. 73-81.