Celtic *marko- 'horse' (updated)

According to Gamkrelidze-Ivanov¹, Celtic *marko- 'horse' (Old Irish marc, Middle Welsh march, Gaulish márkan [acc.])² and Germanic *márxa-/*margá- 'horse, mare' (Old Icelandic marr 'horse', merr 'mare', Old High German marah- 'horse', meriha 'mare') is an Asian Wanderwort related to horseback riding in the Eurasian steppes before the first millenium BC.

The closest forms would be North Caucasian *far-nē 'horse, mare', Sinitic *mrāʔ , Tibeto-Burman *mrāŋ 'horse', with Schwebeablaut. These words³ look as different formations from a lexeme found in Altaic *mórV 'horse' (Mongolian *mori, Tungusic *murin, Korean *màr). One of these derivations, with a nasal suffix (*mori-n), found in Mongol and Tungusic, would be the origin of the Tibeto-Burman and North Caucasian word. The other would be a diminutive *mor-qa reflected in the Sinitic and Celto-Germanic words.

 

However, Sanskrit mṛga 'deer, antelope' would point to a possible origin of this Wanderwort in the domestication of wild horses by the Botai culture of northern Kazakhstan (3,500-3,000 BCE). 

________________________________________________________________________

¹ T. Gamkrelidze & V. Ivanov (1995): Indo-European and the Indo-Europeans, p. 472-473.

² R. Matasović (2009): Etymological Dictionary of Proto-Celtic, p. 257.

³ For which Sergei Starostin reconstructs a Sino-Tibetan protoform *mrāH / *mrāŋ.

Greek parthénos 'virgin, unmarried girl' (updated)

Goddess Britomartis

Greek parthénos 'virgin, unmarried girl' is an interesting word with no clear Indo-European etymology, although the German Indo-Europeanist Gert Klingenschmitt¹ has proposed a derivation from *pr̥-sténos '(having) protunding breasts', a compound of *pr̥- 'before' and *sténos 'breast' (Avestan fštāna-) < IE *psten(o)- parallel to Avestan ərəduua-fšnī- 'having firm breasts'.

Despite being accepted by some specialists², to me this proposal is phonetical and semantically implausible, so I regard this word as a kinship term with parallels in Old Prussian mārtin, mārtan 'bride', Lithuanian martì ‘bride, daughter-in-law’, Latvian mā̀rša 'brother's wife', Crimean Gothic marzus ‘wedding’, apparently a Wanderwort found in Eteocretan *marti-/*marpi- 'virgin/maiden' (cfr. the Goddess Britomartis 'Sweet Virgin/Maiden'), Etruscan marθ 'bride' (in papac marθc svlisva 'the wills of the grandfather and the bride')³, and ultimately related to East Caucasian (Nakh-Daghestanian) *bHaddɮi (~ -ǝ-) 'young (of animals)'⁴, with cognates in Proto-Altaic *bāla 'child, young' and PIE *pelH- 'foal' (Greek pôlos 'foal', Armenian ul 'kid, young of deer or gazelle', Albanian pelë 'mare')⁷

In my opinion, this is the likely source of Sanskrit pr̥̥thuka- 'calf, young of an animal', Armenian ort^c 'young of cattle or deer' and Greek pórtis 'calf, young heifer', pórtaks 'calf'⁵, linked by Thomas Burrow⁶ to parthénos. The semantic shift from 'heifer' to 'virgin, unmarried girl' would be explained in the context of a pastoralist society like those of Kurgan people. Also related would be Germanic *farzá-, *farzḗn, *fársō(n) 'bull, ox' and Slavic *porsъ 'bull'.

______________________________________________________________

¹ G. Klingenschmitt (1974): 'Grieschisch παρθένος', in Antiquititae Indogermanicae. Gedenkschrift für Hermann Güntert, pp. 273-278. 
² R.S.P. Beekes (2010): Etymological Dictionary of Greek, p. 1153. See also X. Delamarre (2008): Gauloises Ardasina, Titiluxsa, Uxesina, grec parthénos, avestique ərəduuafšnī-. Une dénomination indo-européenne de la jeune femme: 'celle qui ha les seins hauts'.
³ A. Morandi (1987): La tomba degli Scudi di Tarquinia [Contributo epigrafico per l'esegesi dei soggetti], in Mélanges d'École française de Rome. Antiquité, p. 104.
⁴ Diakonoff-Starostin link to this Hurrian pōra-(m)mi, Urartian porā 'slave'. See I. Diakonoff & S. Starostin (1986): Hurro-Urartian as an Eastern Caucasian Language, §2.
⁵ Also related are Old Church Slavonic za-prъtъkъ 'wind egg', Czech s-pratek 'premature calf' (Pokorný).
⁶ T. Burrow (1955): The Sanskrit Language, p. 71. 
⁷ J. Mallory & D. Adams (2006): The Oxford Introduction to PIE and the PIE world, p. 192.

Furnée and the Pre-Greek substrate (updated)

In order to characterize Pre-Greek loanwords in his own Etymological Dictionary of Greek (2010)¹, the Dutch Indo-Europeanist Robert Beekes used his compatriot Edvard Furnee's PhD disertation Die wichtigsten konsonantischen Erscheinungen des Vorgriechischen (1972), a pioneer work later followed by Vorgriechisch-Kartvelisches. Studiem zum ostmediterranen Substrat nebst einem Versuch zu einer neuen pelasgischen Theorie (1979) and Paläokartvelisch-pelasgische Einflüsse in den indogermanischen Sprachen. Nachgewiesen anhand der spätindogermanisch-griechischen Reflexe urkartvelischer Sibilanten und Affrikanten (1986).

In these later works (thoroughly ignored by Beekes²), Furnee develops the theory Pre-Greek/Pelasgian was a Paleo-Kartvelian (an older stage than Proto-Kartvelian) language. For example, Greek kháos 'abyss, chasm'³ < Pre-Greek *kʰáw- can be linked to PK *qew-/*qaw- 'ravine' (Old Georgian qev-i, qev-n-i, Georgian xev-i, Megrel xab-o)⁴, where PK q ~ Pre-Greek kʰ'⁵.

In my opinion, these and other correspondences between Pre-Greek and Kartvelian would be evidence of language contact rather than a genetic relationship⁶, as they include borrowings into the opposite direction. For example, although PK *tor (Georgian tor-i, torn-i 'harness, armour')⁷ corresponds to Greek thôraks (Ionic thôrēks, Mycenean TO-RA-KE) 'cuirass; trunk, chest'⁸ < Pre-Greek *tʰur-āk-, it's clear the source can't be Paleo-Kartvelian.

This would be an Eurasiatic root found in Latin corium 'skin, hide, leather', cortex 'bark, shell', Sanskrit cárman-, Avestan čarǝman- < PIE *s-ker- 'bark, skin'⁹, PK *tser- (Laz cara) 'rough infertile earth', Proto-Uralic *tɕarV 'hard', Proto-Altaic *tSarV 'snow, snow crust', Proto-Dravidian *tSara- 'rough on surface', and whose protoform is reconstructed by Dolgopolsky as *tɕ(ˀ)arV/*tsˀarV 'hard/firm, hardened crust, hard surface' (ND 311). The initial sibilant affricate was centumized into a velar stop¹⁰ (palatalized in Indo-Iranian) in IE but it gave a dental stop in Pre-Greek.
__________________________________________ 
¹ Which is part of the so-called Indo-European Etymological Dictionaries Online (IEDO) project, a series of etymological dictionaries of IE languages edited by Alexander Lubotsky from Leiden University. Until some years ago, preliminary versions of them were available online for free under the label Indo-European Etymological Dictionary (IEED), but unfortunately now they are of restricted access, as copyright is owned by publisher Brill. 
² He regards Pre-Greek as a uniform non-IE language "with no recognizable cognates", thus ignoring most recent research on the subject. 
³ On the other hand, the related kháũnos 'slack, porous; vain, frivolous' < Pre-Greek *kʰáw-n-o- derives from a suffixed variant of the same root. Beekes' allusion to a possible parallel in Germanic and Baltic 'palate' is totally ungrounded. See R.S.P. Beekes (2010), op. cit., p. 1614.
⁴ E.J. Furnée (1979), op. cit., §57.
⁵ Notice, however, the possible link (already observed in Varro's De lingua latina) between the Greek word and Latin cavus 'hollow' < PIE *kouH-o- by Thurneysen-Havet's Law. See L. Horton-Smith (1895): Establishment and Extension of the Law of Thurneysen and Havet, The American Journal of Philology, 16, pp. 444-467.
⁶ In fact, the Kartvelianist and Indo-Europeanist Thomas Gamkrelidze accepted some of Furnée's correspondences, relating them to the argonauts. See T.V. Gamkrelidze & V.I. Ivanov (1995): Indo-European and the Indo-Europeans, p. 799-804.

⁷ E.J. Furnée (1979), op. cit., §16.

⁸ R.S.P. Beekes (2010), op. cit., p. 569.
⁹ Please notice Greek kôrykos 'leather sack' is unrelated. See R.S.P. Beekes (2010), op. cit., p. 816.
¹⁰ Not *sk- as thought by Nostraticists, because *s- (called "s-mobile" in the literature) is a fossilized prefix.

Basque sagu 'mouse'

Basque sagu, sat- 'mouse' is a word found in compounds such as satandera (B, G), satandre (S, R) 'weasel' (and(e)re 'lady'), satitsu (G, HN) 'shrew' (itsu 'blind'), sat(h)or 'mole' (second member unknown) and saguzar (B, G) 'bat' (zahar 'old'), and which has cognates in Afrasian: Akkadian šikkû 'mongoose', Amharic ǝkokko (aškokko, ǝškokko) 'rock hyrax', Ge'ez sˁǝkˁ^wǝnkˁ^wǝn, sˁǝkˁ^wǝskˁ^wǝn 'field mouse; a kind of lizard'¹, Central Chadic *Sakw- 'squirrel; dog'², Omotic *sakw/*sikw 'chameleon; bat', Kartvelian *tsiqˀw- 'squirrel, mouse' (Georgian ciqˀwi 'squirrel'), Dravidian *tśikk- 'mouse', Old Turkish čekün 'young of marmot', Tungus čekše 'tarbagan (Marmota sibirica)'. Also related are Etruscan seχ³, Thracian súkhis and IE *dhug-h₂ter 'daughter'.

Although Bengtson links the Basque word to Caucasian *tsa:rggwɨ:(~ -ǝ:,-a:) 'weasel, marten' (see here), from which Starostin thinks the Kartvelian word was borrowed, in my opinion this is a genuine Eurasiatic root which in some languages (Kartvelian⁴, Dravidian, Altaic⁵) conflated with a phonetically similar word 'small'⁶, probably for tabooistic reasons. 

______________________________________________

¹ Quoted by Dolgoposky (ND 310a). See also W. Leslau (2010): Concise Dictionary of Ge'ez, p. 226.

² Where S is an unspecified sibilant.
³ Also spelled as sec, śec, śeχ. 

⁴ Kartvelian *tśˀukˀ- (Georgian cˀukˀ-an-a 'very small', Megrel čˀukˀ- 'mouse').

⁵ Altaic *tʃhjà:khe(~ -u) 'small'.

⁶ For which Dolgopolsky (ND 334) reconstructs a separated root, which he considers to be phonosymbolic.