Burushaski shepherd vocabulary of Indo-European origin.
Article Type: Essay
Subject: Burushaski language (Etymology)
Vocabulary (Etymology)
Indo-European languages
Shepherds
Author: Casule, Ilija
Pub Date: 01/01/2009
Publication: Name: Acta Orientalia Publisher: Hermes Academic Publishing Audience: Academic Format: Magazine/Journal Subject: Social sciences Copyright: COPYRIGHT 2009 Hermes Academic Publishing ISSN: 0001-6438
Issue: Date: Annual, 2009 Source Volume: 70
Geographic: Geographic Scope: Pakistan Geographic Code: 9PAKI Pakistan
Accession Number: 300652470
Full Text: Abstract

The etymological analysis of Burushaski shepherd vocabulary shows that almost all the pastoral terms in this language are of Indo-European origin (some thirty independently of Indie and Iranian), with a significant proportion showing close correlations with the Paleobalkanic substratal shepherd terms. Considering the conservatism of this semantic field and together with the precise, specific and systematic phonological, grammatical and lexical correspondences, this is further evidence for the Indo-European and Balkan origin of Burushaski.

Keywords: Burushaski, etymology, shepherd vocabulary, Indo-European, ancient Balkan languages, Albanian and Rumanian substratum.

1. Introduction

Burushaski, still considered a language-isolate, (1) is spoken by around 90,000 people (Berger 1990: 567) in the Karakoram area in North-West Pakistan at the junction of three linguistic families-the Indo-European (Indo-Aryan and Iranian), the Sino-Tibetan and the Turkic. Its dialectal differentiation is minor. There are three very closely related dialects: Hunza and Nager with minimal differences, and the Yasin dialect, which exhibits differential traits, but is still mutually intelligible with the former two.

The earliest sketchy descriptions and word lists of Burushaski are from the mid to late nineteenth century (e.g. Cunningham 1854; Hay ward 1871; Biddulph 1880; Leitner 1889). The limited dialectal differentiation and the lack of older attestations make the internal historical reconstruction extremely difficult.

The most notable and authoritative modem description is Berger's (1998) three-volume work on the Hunza-Nager dialect. Still very relevant is Lorimer's earlier ground-breaking three-volume work on Hunza-Nager (1935-1938) and Yasin (1962) Burushaski. Edel'manKlimov's (1970) analysis (revised and summarised in Edel'man (1997) is valuable in the quality of the grammatical description. Willson's (1999) compact basic Burushaski vocabulary is also very useful. Fundamental for the study of Yasin Burushaski are Berger's (1974), Tiffou-Pesot's (1989), Tiffou-Morin's (1989) and Zarubin's (1927) works.

Berger's (2008) posthumously published synthesis is instrumental in establishing aspects of the historical phonology and morphology of Burushaski and its internal reconstruction.

In our comparative historical work (Casule 1998, 2003a, 2003b, 2004), we have uncovered regular phonological correspondences between Burushaski and Indo-European (outside of Indie and Iranian) in over 550 lexical correspondences, largely in core vocabulary and in compact semantic fields (the correspondences [over 70 of them] in the names of body parts can be found in Casule (2003a). Most importantly, and highly significant in determining genetic relationship, we have found very significant and fundamental derivational and grammatical (2) correspondences (in the personal pronouns, the whole system of demonstratives, nominal case and plural endings and derivational suffixes, the numerals, verbal prefixes, suffixes and endings, the complete non-finite verbal system, all of the adjectival suffixes, adverbs, postpositions, etc.) 3 We concluded that Burushaski displays characteristics of a language which could have had an early relationship or contact in its history with the Southern (Aegean) branch of Indo-European on the one hand (see esp. Casule [2004], on the possible correlation with Phrygian) and with the Northern/Western Indo-European group on the other.

For easier reference, we reproduce Berger's table of the phonological system of Hz Ng Burushaski, which is essentially valid for the Yasin dialect as well (Yasin Burushaski does not have the phoneme ch [Tiffou-Pesot 1989: 7-9]):

[TABLE 1 OMITTED]

Notes:

1. All five vowels can be phonetically long, but for prosodical reasons Berger marks them as double (two component) vowels, in order to mark the position of the stress.

2. Retroflex consonants are marked with an underdot.

3. w and y are allophones of u and i.

4. c = ts in Lorimer and c in Tiffou-Pesot (1989).

5. n = [n] or [ng] [nk].

6. y is a retroflex, articulated somewhere between a "r grasseeye and a [gamma] or rather a fricative r with the tongue in a retroflex position" (Morgenstierne 1945: 68-9).

We summarise below the established phonological correspondences between Burushaski and Indo-European (for the exemplification, analysis, and the internal variation, refer to Casule (2003b: 24-42):

In this paper we look closely at the Burushaski shepherd vocabulary and analyse the lexical correspondences with Indo-European, particularly outside of Indie and Iranian. We first look at twenty correlations in names of 'small cattle (goats and sheep)' (sections 2.2.1-2.2.4) and at eleven related shepherd terms ('shepherd', 'byre', 'grazing ground', 'wool', 'flute') (section 2.2.5).

So far, all of these terms have been deemed to be autochthonous Burushaski words by Berger and Lorimer and other scholars, with the exception of Bur dagar Ng 'ram', ex. [13], which has been tentatively linked by Berger to Indo-Aryan.

We have labelled the validity of the etymologies, in the sense of original Burushaski words of Indo-European (non-Indo-Iranian) origin, as follows: C = certain, VP = very probable, P = probable, T = tentative. Although such assessment can be subjective, some of the well known criteria applied are: the systematic character of the phonological correspondences, drawing on internal reconstruction where possible, consistency with the phonological processes, alternations, changes and adaptations typical of Burushaski, minimal to no semantic latitude, specific semantics, antiquity in IE, improbability of borrowing, consistent correlations with particular IE groupings and in other semantic fields, avoidance of "root" etymologies, links involving a derivational cluster (e.g. noun + underlying verb, noun + underlying adjective, several nouns + underlying verb, adjective + underlying noun; noun, adjective + underlying verb etc.), correlation of longer phonological segments, etc.

The sources for the IE material are Gottlieb (1931), Pokorny (1959), Watkins (2000), Mallory-Adams (1997), Mallory-Adams (2006) and Buck (1949). It is important to note that the Burushaski material has already been sifted carefully for Persian, Urdu and Indo-Aryan loanwords by Berger, Lorimer, Morgenstierne, Zarubin, Edel'man, Klimov, Varma, Tiffou, Buddruss, Tikkanen and other scholars who have studied the language--their findings are conveniently incorporated in Berger (1998). The main source we have used for further comparison with Indo-Aryan is Turner (1966)5 and with Persian, Steingass (1999) [1892].

In all our etymologies we have excluded all cases where there is a full match (phonological or semantic) with Indo-Aryan. It could well be that in some cases we may be identifying as Sanskritisms the native Burushaski Indo-European vocabulary. Thus, for example, in Berger (1998) there are 45 Burushaski stems not found in the surrounding Indo-Aryan languages which he relates directly to Turner's Old Indian entries, sometimes with no intervening attestations within Indo-Aryan, and with little (or inconsistent) historical change. The phonological rules for Sanskrit borrowings into Burushaski cannot be less systematic just because of the close geographical and historical proximity of the two languages.

2. Etymological analysis of Burushaski pastoral terms of IE origin

2.1. Indo-Aryan and Iranian borrowings

Burushaski has only a handful of borrowings in words denoting sheep and goats. We list first those already identified: Bur war 'full-grown ram (entire)', ruse war 'Ovis poli ram' from Wakhi war, war, Lorimer also points tentatively to Khw werkalu 'ram' (L371) (B 464); Bur rus 'Pamir wild sheep, Ovis Poli' < Wakhi ris, Sh ruus (B 366); Bur buc 'he-goat, two or three years old, not castrated' (B 60) (L 86) (possibly from Wkh but registered only by Lorimer); Bur don 'big herd' (also in Sh) < T 6717).

To these we can add Bur baqta Hz Ng 'a type of big sheep', NH 'fat tailed sheep', Ys baxta, (Sh bakhta, Khw baxta) (B 38) from Persian bakhta 'a ram three or four years old; a fat tail' (Steingass 159) and Bur chageni NH 'black goat' (B 95) [and possibly Bur chigir 'goat' (B 76)] < Skt chaga- 'he-goat' (< IE *(s)kegos- 'sheep, goat', M-A2 140).

Bur also has mees 'bag of skin, (...) holds 24-40 lbs of grain' (in Sh kaladu)' which is not found in the surrounding Indo-Aryan or Iranian languages (L 286). Berger (B 286) points to T 10343, i.e., Skt *maisya- 'ovine', also T 10334 mesa 'ram, sheep; fleece, skin'. It is curious that the meaning of the Burushaski word corresponds closer esp. to Lith malsas 'bag', Lett maiss 'sack' or OS1 mexu 'sack, skin, fur' (M-A2 140) than to the Indo-Aryan developments--nowhere in Indo-Aryan, according to Turner, do we find the meaning 'sack' or 'bag'. (6)

Further borrowings from Indo-Aryan are: Bur uriin 'male young sheep, castrated' (B 457), noted as a loanword from Shina, where we have both uran and uriin (< T 2349: OInd urana- 'ram, sheep, young animal') from an IE *ur[h.sub.1]en- iamb' (M-A2 140) and Bur Ys bran 'ram' (LYs 50) (BYs 135) possibly from Khw bran 'same'.

2.2. Burushaski autochthonous pastoral vocabulary of Indo-European origin

2.2.1. Sheep and goats jointly

The overwhelming majority (80-90%) of the Burushaski terms denoting 'sheep' or 'goats' can be traced to Indo-European--that is the case with 9 of the noted loanwords from Indo-Aryan or Iranian (Persian) and the 20 autochthonous IE terms outside of the Indo-Iranian sphere which we discuss in this subsection.

The basic Bur word for 'small cattle, sheep and goats' has a firm Indo-European etymology:

[11 [C] Bur huyes (sg and pi) 'small cattle (i.e., sheep and goats)' (B 209) (L 211) (Will 66: 'a sheep; a goat; a flock'), in Ys also: hui s (T-P 140). We should note the difference between Burushaski and neighbouring Shina ei 'sheep' and Wakhi yobc 'ewe', which eliminates them as a source of borrowing. Burushaski also has [2[ [C] huyoo 'wool-bearing animal, sheep' (L 208) (B 209) (the latter < *huyyoo with a suffix -yo-, see the discussion in [27]).

There is a direct and remarkable correspondence with IE *[h.sub.2]ouis (7) (gen. *[h.sub.2]ouios) 'sheep (Ovis aries)': OIr oi , Lat ovis, ON oer, OEng eowu, OHG ou ~ ouwi 'sheep', Lith avis, Lett avis, PS1 *ov1ca (e.g. Med ovca) < *owi-ka, Luw hawa/i-, Lycian xawa, OInd avih, all: 'sheep', (Pali avi 'ram', Shina ei 'sheep', etc.), Arm hoviw 'shepherd', Gk o(w)is 'sheep', Wakhi yobc 'ewe', TochB eye (<*[h.sub.2]oueis) 'sheep', pi. awi 'ewes' (M-A 510) (G 461) (IEW 784-*oui-s).

Berger (B 209) cites a form hunyes 'sheep, goats' in Hayward (1871), not registered by anyone else as such in his time or later, which indicates it is most likely an error, especially as the group ny would need to have been preserved in Yasin. It could however be explained by derivation from Indo-European as well, from an IE *[h.sub.2]oui-no-s, as in Baltic *awinas > Lith avinas, Lett avins, auns, OPruss awins and OS1 ovlnu all: 'ram' (Vasmer 111:113, "from an old Indo-European augmentative") and as adjective in Latin ovlnus 'ovine'.

The exact semantic and phonological correlation (the preservation of the laryngeal, houi- > hui, -s) with Proto-Indo-European is remarkable in this word attested in eleven of the main groups, which shows once again the "resilience" of basic shepherd vocabulary.

The other basic Burushaski word for 'sheep, goats' jointly can be linked tentatively to Indo-European:

[3] [PI Bur adds 'sheep and/or goat; small cattle' (L 8) (B 11).

We suggest a possible link with IE *h"eigs- 'goat' (M-A 229) (in Wat 1, without a laryngeal: *aig-): Alb edh 'kid', Gk aiks '(she-) goat', Arm eye '(she-) goat', Av izaena '(goat) hide'. Perhaps in some way related, and closest to the Burushaski form, note also Phrg aseis 'he-goat' (Haas 1966:158).

In Burushaski we would have the usual changes ai > a and gs > ks > s and we would have expected a form *a$as or *asas. It is indicative, however, that we can find examples where we have a : s variation in Burushaski e.g. masis : macis (B 283) or asuus vs Sh acuus (B 23). Influence from forms like OInd ajd 'he-goat' or Av aza 'he-goat' (< IE *[h.sub.a]egos 'he-goat') cannot be excluded entirely. In Burushaski, considering the ending -as, maybe < IE *[h.sub.a]egios, as e.g. Lith ozy s 'he-goat' (IEW 7). (8)

The next Burushaski word has a firm Indo-European etymology and refers both to 'goatskin' and 'sheepskin':

[4] [C] Bur bat 'goatskin, sheepskin; animal hide; a shape, a form, a likeness' (Will 21) (B 44). Further, Berger derives from *u-bat > Bur -wat (9) 'skin, body; person' and wat 'bark, a peel, rind' (B 466) (Will 74). Note also Bur batd 'bald; hornless' which Berger relates to the same stem (compares it with the semantics of 'skinhead') (B 44), where we have an exact phonetic match with the IE form below, with -a < -[eh.sub.2] in the auslaut.

Most likely related, reflecting the p : b alternation in IE below, is Bur phatako Ng 'bald', which Berger (B 327) links with bata 'bald' (B 44).

The Burushaski words correspond directly with IE *baite[h.sub.2] -'goatskin, cloak' (OEng pad 'coat', OHG pfeit 'garment', Goth paida 'tunic, shirt', Gk baite 'shepherd's or peasant's coat of skins; tent of skins', maybe also Alb petk 'clothes, garment' (IEW 92-3: *baita ~ *paita--considered in Greek possibly from a Thracian source) (M-A 109-110: "may be but need not be a borrowing from a non-IE source").

In the following very complex example the reference is to 'kid' or

'lamb':

[5] [C] sumphalikis, with -ph- in L and B, in L also sum(p)falikis, but with -p- according to the forms given by NH: sumpal, sumalkis, sumapalikis '(L very) young female kid or lamb (10), (L 318, also 'a derogatory term for effeminate man') (B 385). Berger derives it from sum 'female animal' (B 384) + the second component that does not occur independently. [We (Casule 2003b: 31 and 2009) have traced sum 'female animal' and suman 'male animal' together with Bur -isamanum 'first-bom (son, daughter, young animal)' (L 47) from IE *sem 'one' with a semantic development as in Slavic, e.g. Russ samec 'the male animal' and samka 'the female animal' (Buck 139-40).]

Burushaski also has sepalkis 'sheep with fine wool' (L 326) (B 394) (Leit sepalis q. in Berger 2008: 19.3) where se 'wool' [28], and sekis 'woolly' (B 392), also the compound biskepalkis 'wool-carrying, wool-producing animal', Ng also biskepalis" (
The Burushaski stem -p(h)al- (with e>a in an unstressed syllable, see B 1:2.5 and below), can be related to IE *pel[h.sub.2]- 'bear young' (found in the West Central IE area): Alb pjell 'give birth to, produce', pjelle 'child', pele 'mare', Gk polos 'foal', Arm ul 'kid, young of deer or gazelle', Eng foal (M-A2 192). Burushaski would have lost the underlying verb (like Gk, Arm or Eng), yet note that in biskepalkis 'wool-carrying, wool-producing animal' the semantic element of 'bear, produce' comes to the fore. The Burushaski semantics of 'kid' or 'lamb' fits neatly with the other semantic developments. Particularly interesting is the correspondence between Burushaski and Albanian, considering the large number of correlations in shepherd vocabulary between the ancient Balkan languages and Burushaski.

We can analyse su mphalikis into four components: sum + phal + i + kis : < *sum-pel([h.sub.2])-yo-kirs (12) or rather < *sum-pel([h.sub.2])-yo + *kirs : sum- 'female animal' + -phal- 'young [newborn = 'very young'] animal [kid, lamb] + a suffix -i- possibly from the relational suffix -yo- discussed in [27] [for a change -yo- (unstressed o) > -yu- > i, note Berger's analysis (2008: 19.15) of Bur Ys cumanikis < *culmanyu-kis] + -kis < kiso 'interjection for driving away sheep and goats' (NH 'sheep' in children's talk) (B 245) < *ki'rso, of Indo- European origin, < IE *ker(s)- 'horned; sheep, ram, etc' see [10-12].

The semantic structure of the compound noun would thus be 'female' + 'very young, newborn' + 'sheep' (or 'horned small cattle'): lit. 'female animal+very young+of+sheep/goat', i.e., 'very young female lamb or kid'.

The form sumpal could be a backformation (as per Berger 2008: 19.3), or perhaps proof that the ending -kis is not actually a suffix here, but a noun.

Berger (ibid) considers -kis in these examples to be the Burushaski adjectival, but also nominal suffix (B I: 19.6), which is probable--identifying it in our analysis as the final morpheme would not change significantly our etymological analysis.

Berger (2008: 131-2) offers a different etymological solution. He derives -pal- from belis (L also belis), Ys beles 'ewe (which has had young)' (B 48) [14], which in turn he derives from a hypothetical adj. *belkis and ultimately from bei- 'to put on clothes' (a stem of IE origin in our analysis < IE *uel-, see the discussion in [27]. He doesn't seem to account for the -i- in su mphalikis or sumapalikis, and the postulated change -mb- > -mp- is difficult to substantiate.

His appears to be a weak etymology on several accounts: if the original word means 'young female kid or lamb', there would be no need to indicate especially 'female sheep' twice (by using sum-) and in the other compound nouns it refers to 'sheep' in general. The suggested change -kis > -is would be most unusual, if not impossible. Moreover, the Ys form wel- 'to put on clothes', points to an older form with w- for Hz, Ng bei- 'same' (although Berger [2008: 69] suggests the Yasin verb may derive < *u-bel-). And while a direct semantic derivation from 'to put on clothes' to 'wool' is common, much less plausible is a direct link from 'to put on clothes' to 'young sheep or kid'.

It may be possible that -pal- and belis are actually related (see [14]), yet with a change p > b rather than b > p. The semantic derivation is different to Berger's - from a semantics of 'borne' > 'kid' or 'lamb' and 'bearing young' > 'ewe, which has borne young' (similar to the correlation 'foal' and 'mare' in the Albanian examples above). This provides a coherent frame for all the noted examples.

Berger goes on to stipulate that the form Ys beskaret, Hz Ng baskarat, Cunn. Leit. "bashkar" 'wether, ram (over 2 years old, castrated)' (L 72) (B 42) [8] should be derived from a "highly reduced first component," i.e., traces bes- < *belikis or *belis or *belis which is highly unlikely (we would also expect compensatory lengthening of the preceding vowel), and for such a radical change there are no other examples.

The Burushaski words are not found as such in any of the neighbouring languages, but there are some developments in Shina and Domaaki on the one hand and Balti on the other, that need to be addressed and explained in order to rule out any borrowing.

As for Shina, Lorimer (in personal communication to Turner (noted in T 8125.2) cites Sh palo 'young animal' > Dom palo 'same' (not correlating it with Burushaski). These words most certainly would have been noticed and rejected by Berger who makes excellent and very precise use of Turner. They are at odds semantically with the other derivatives in Indo-Aryan: Kai palo(i) 'cattle-shed', Bshk palo 'same', Panj pali 'herdsman, shepherd' < OInd (reconstructed) *pal 'flock' < OInd (also reconstructed) *pala 'protection', ultimately from OInd paid 'protector; herdsman', palaka 'guardian' (T 8125). The meanings 'herdsman, protector, guardian; cattle-shed' as per Turner and (implicitly) Lorimer, are not recorded for Sh, Dom or Khw, and only in Pk pala 'keeper', Ksh pal 'shepherd', Ku palsi 'shepherd'. Note also OInd avipala 'shepherd' (T 893). The direct contact and the many borrowings between Burushaski and Shina could suggest the Burushaski words above are loanwords from Indo-Aryan. Yet, as we have indicated in the introduction, close proximity and language contact cannot absolve us of a systematic explanation.

The problems with considering su mphalikis a Shina loanword into Burushaski are:

1. Semantically, the Shina word does not refer specifically to 'kid' or iamb' or 'sheep', but to 'young of animals' (in general). Within Indo-Aryan this meaning is kept only in the two languages, Shina and Domaaki which in their intensive contact with Burushaski have borrowed readily from it. Moreover, the semantic shift in regard to Old Indian or Indo-Aryan is on the borderline of acceptability (a direct semantic derivation of the word for 'young animal' from the word for 'shepherd' seems to us impossible, so too from 'herd'). Within the Indo-Aryan developments above, both Shina and Domaaki stand out as the only ones with a shift to 'young animal' [even in T 8100: para 'bringing across' which Turner gives as the ultimate source for OInd pa la (above) there are no derivatives denoting any kind of small cattle or other animals].

2. Phonologically, when Burushaski borrows Shina (or for that matter Urdu) words with a long vowel, e.g. Sh rajaaki (T 10694) > Bur rajaaki (B 362); Sh jaalo (Pers jalah, Khw jalo) > Bur jdalo (B 219); Sh da ado (T 6261) > Bur daado (B 108); Sh daar (T 6793) > Bur daar (B 109), etc., the vowel length is retained, which is not the case with the Burushaski pastoral term above.

If Burushaski had borrowed the stem from Shina, it would have retained the -o in sumpal as well, since according to Berger (B I: 209), it has actually borrowed this suffix from Shina.

In Burushaski words borrowed from Shina, we have not been able to find a change p- > ph-, whereas in the older Indo-European layer this is a regular change (Casule 2003b: 34).

3. The only available source for the Shina word seems to be Lorimer, i.e., his example comes most probably from the Burushaski speaking or adjacent areas, and a strong argument can be made that it is actually a loanword from Burushaski into Shina and Domaaki.

We should comment further on the Balti (Burushaski's Tibetan neighbouring language) word bal, (Tib bal, also Nep bal) all 'wool' (RYTEDD online - in all examples it means 'wool' and not 'sheep' or 'kid'). It is very difficult, if not impossible, to correlate with it Bur su mphalikis - where the only basic meaning is 'female kid or lamb' or the compound words with -p(h)al- where a first element meaning 'wool' needs to be added, as in: sepalkis 'sheep with fine wool', where se 'wool', and the compound biskepalkis 'wool-carrying, wool-giving animal' where biske means 'hair (of animals), fur', as the stem does not refer to 'wool'. If the basic meaning of the second component of these compound nouns were 'wool', there would be no need to mark that twice. Furthermore, the phonological change mb > mp(h) would be highly unusual. Moreover, if there is a link with [14] belis 'ewe (which has had young)', which is probable, then the discrepancy in the stem vowel cannot be explained. All of this rules out Balti as a source for the Burushaski words as well.

All the adduced evidence in our etymological explanations makes this a firm correspondence.

[6] [VP] Bur du 'kid (up to one year old of either sex)', NH also du do (L 139) (B 123).

It can be correlated internally with a cluster of Burushaski words which we have derived (Casule 2003a: 36-37) from IE *dhe(i)- 'to breastfeed' (IEW 242), also with the -1- formant, i.e. < IE *dhh,ilcha- 'teat, breast', *dhe[h.sub.1]lus 'nourishing, suckling' (M-A 82, who consider this form a northwestern and late IE word for 'teat, breast').

The stem is strongly represented in Burushaski and cannot be traced phonologically or semantically to Indie or Iranian: (1) Ys -dil (BYs 142), Hz Ng -ndil 'breast, chest' (L 276) (B 302) (cp. with Mir deil 'teat', OHG tila 'woman's breast', OEng delu 'breast, teat', ON dilkr 'lamb') and further with (2) Ys dulas 'boy, young lad' (BYs 142), Hz Ng hiles 'boy, lad, youth (unmarried); child, infant' (L 201-202) (B 198) (the last form is considered secondary by Berger (2008: 55), derived from the form with d-) (cp. with Lett dels 'son', Alb djale 'boy, young man, son' (reconstructed by Cabej 1976 [apud Desnickaja 1984], from OAlb *delas 'a boy'), Lat filius 'son'; (3) Bur dinar 'buttermilk' (: e.g., Alb dhalle 'buttermilk', also Alb dele 'sheep', and further Illyr dalm- 'sheep' (IEW 242), and possibly (4) duno 'teat of udder' (B 125) (cf. from the same stem Olr denaid 'teat' (< *dhi-na-ti), dinu iamb'). For a discussion of the alternation i:u/_l,r, manifested here, see Berger (2008: 2.10).

In light of all these examples, it is very likely that Bur du belongs to the same cluster. (13)

2.2.2. Sheep

2.2.2.1. Sheep (specific)

[7] [T] Bur sopan 'sheep's meat' ('Schaffleisch oder Buchweizenteig im Schafmagen gekocht') (also in Sh), sopan boin 'a shepherd festival' (B 397).

There is a very tentative possibility for comparison with OHG scaf, Grm Schaf, OEng sceap, Eng sheep, all: 'sheep' under an older hypothesis derived from IE *sqeb- (Gottlieb 1931: 19), i.e., < IE *skeb(h)-, also *skab(h)-, *skob(h)- and *ske p-, *sko p-, *ska p-(IEW 930-3) 'base of words with various technical meanings such as 'to cut', 'to scrape', 'to hack' > 'creature' (Wat 77) : Goth gaskapjan 'to create', OHG scaffon 'form, cause', OEng gesceap 'form, creation', sceppan 'to form', OEng sceafan 'to scrape, pare away', Eng shave, Grmc *skopo 'thing cut out', 'container', Gk skaphe 'boat' ('thing cut out'), Lat capo 'castrated cock', scabo 'to cut', scapula 'shoulder blade', capulare 'to cut', OSl kopati 'dig', skobli 'scraping knife' etc. (IEW 930-33). Another etymology of Eng sheep etc. would see it as a dissimilated form of IE *(s)Regos- 'sheep, goat', e.g. Skt chaga- 'he-goat' [which is most likely the source of Bur chageni NH 'black goat' (B 95) and possibly of Bur chigi'r 'goat' (B 76)], Osset sney 'she-goat', OEng hecen 'kid' (M-A2 140).

In the Burushaski example we have both the meaning of 'sheep' and '(cut up) sheep's meat'. In Burushaski: sk- > s, and we would have the o-grade of the IE stem.

Another more realistic interpretation could seek a correlation with Pers suban > Turk coban 'shepherd', Srb and Croat coban, MGk tsopanis, Alb coban and Rum cioban 'same' (Alinei 2003: 51, who considers the Balkan developments not of Ottoman provenience, but of greater antiquity). The semantics in this second etymology only corresponds partially and we would need to seek a modern Turkic or Iranian source for the Burushaski word.

2.2.2.2. Ram

[8] [C] Bur Ys beskaret, Hz Ng baskarat, Cunn. Leit. "bashkar" 'wether, ram (over 2 years old, castrated)' (L 72) (B 42).

And further, [9] [C] Ys (DC) beske, also biske, Hz Ng biske 'hair (of animals), fur', biske siqa (14) 'Verbascum thapsus Linn., = kleineblutige Wollblume, an Quellen', and the first component in biskepalkis 'wool-carrying, wool-giving animal' (L 84) (B 56) and perhaps bisqar adj. 'raw, rough wool' (B 56-7). The forms ending in -e would point to a genitive form from *bes-ka or *bis-ka (-e is the Burushaski genitive and ergative ending). The forms with -s- are certainly older, i.e. sk > sk. The derivation of Ys beskaret could be from *bes-ka-r(u)-eti (< 'wool-giving').

This is an important direct correspondence with Alb bashke 'sheep's wool', Rum basca 'same', considered of substratal Paleobalkanic origin by Brancus (1983: 40-1), who analyses -ke as a suffix. Several etymologies have been put forward for the Albanian and Rumanian words, e.g. 1. From a Thrac *baska, *vaska 'wool', derived with the suffix -ka < IE *ues- 'to clothe'. 2. Russu (apud Brancus (41) (less likely) < IE *bhasko- 'band, bundle' (e.g., AncMcd baskioi, Lat fascis 'bundle').

Correlatable to IE *wes- 'to clothe' (Wat 101), Burushaski also has the verb -wasi- 'put s-thing in or on', (fory-sg objects bisi-) 'put on; wear; don (clothing); fix, attach; fit; throw; suspend, hang (etc.)' (Will 121) (B 56), which can be linked with the first of the proposed Paleobalkanic etymologies, as IE u- > Bur b- (15), thus beskaret < *wes-karet and beske < *wes-ka < *wes-i-ka.

Lorimer (L 35 and 372) provides forms of the verb with an -estem vowel, and gives both -wasias and -wesias and the pp. only as n-wesin (with many of his examples also with -e- as the stem vowel, (16) which could explain the change e > i in the form bisa-) which points to an older form with -e- consistently corresponding with IE -e-, thus not derivable from OInd vaste 'to put on clothes' or Av vaste 'same' (IEW 1173). Bur -s- < s+i (for this change, see further Berger 2008: 19.19). For the Burushaski stem ending -i-, the Alb vesh 'to dress', from the same stem, is very indicative, as Orel (1998: 50115 502) derives it from a PAlb *wesja < IE *wosejeti (or *wesejeti ?) (with the IE caus. suffix *-ei-), also Goth wasjan 'dress', Hitt wassezzu, Skt vasayati (RV) 'clothes' (T 11600, in all of IA continued only in Si vasanava 'clothe, cover') (M-A 109). The Bur -i- might otherwise be from the IE Mo- formations (see the discussion in Casule 2003b: 78), the most important and productive present suffix of late IE, as the Burushaski present stem regularly involves yodation, whereby s > s (see e.g., E-K 30; 60-61).

In relation to the suffix *-k(a) posited above, there are indications of an old, no longer productive Burushaski nominal suffix -k(a) (if not a form of the suffix -ko).

We note examples with the same derivational structure, like Hz Ng bisqa 'saliva' : Ys mesqe, musqe 'same' (L 83) (B 56), with the b:m alternation, most likely from the zero-grade of IE *meus-, *meug'damp; slimy, slippery, with derivatives referring to various wet or slimy substances' (Wat 55) (IEW 744-5) + the suffix -ka. The bare stem is possibly to be found in Bur -mus 'snot, nasal mucus' (B 296).

For this -k(a) suffix, perhaps most relevant to the analysis here (with an unreduced suffix -ka ?) is Bur tharkd and thari (< ther) 'strip of dirt on the face' (B 438) from Bur ther, Ng therk 'dirty, soiled' (L 351) (B 439) (see Berger 2008: 124). Berger points also to Bur gask 'thick rope for tying loads, for swings' (B 149) (L 163) :
The correspondence with the Paleobalkanic developments is remarkable.

Berger (2008: 132) (see in [5)) derives Ys heskaret, Hz Ng baskarat from a "highly reduced first component", i.e., traces bes<*belikis, or *belis or *belis which is not plausible. Furthermore, that does not account for the obvious connection with beske, biske.

His suggestion, however, that the component *-karet can be correlated with karcelo 'ram' [10| is not without merit. In this case we could have had a haplology (18) *bes-ka-karet > beskdret. See also the tentative discussion in [13].

[10] [CI Bur kareelo 'ram' (also in Sh) (B 242) (L 229: karlelu). -6elo is a Burushaski nom. and adj. suffix: nams 'greed, greediness' > nams-eeio, Ng namsiilo 'greedy', also char 'stone', char-eelo 'climber' (B I: 210), perhaps relatable to IE *-lo-, secondary suffix, forming diminutives, nom. and adj. suffix: Latin -ellus dim. suffix, -ilis adj. suffix, OEng *-ling dim. and nom. suffix (Wat 50).

And further from the same stem: [11] [C| Bur karu 'male ibex' (BYs 157) and [12] [C] krizi and krozo 'small cattle (i.e., sheep and goats)' (BYs 157), ki'ro NH 'sheep (in children's talk)' (B 245) and 2khor 'Schafskorper, dem die Eigenwide herausgenommen sind' (B 256). Note also khar 'a call to a sheep or goat' (B 252), and khirga 'a call to a sheep' (B 255, who relates the first component khir-tentatively to kiro 'sheep'), and further kiso (< *kirso, for the change rs > s in Hz Ng, see the examples and discussion in Berger 2008: 3.26) 'an interjection for driving away sheep and goats' (NH: 'sheep' in children's talk) (B 245). Under one interpretation, suggested by Berger (2008), the component -kar-et in Ys beskaret, Hz Ng baskarat, Cunn. Leit. "bashkar" 'wether, ram (over 2 years old, castrated)' (B 42) would also belong to the same derivational cluster (in our analysis < *beska-karet, see [8]).

The wide representation of the stem (with apophony): -kar- : -khar- : khir- : kir- : khor in Burushaski (10 words), argues strongly in favour of considering Burushaski the original source of these words, some of which were subsequently borrowed into Shina (note further Shina Chilas kara 'a ram') and maybe in Wkh kar 'sheep with short ears' (q. in Zoller 2005: 103).

There is a direct correspondence with Indo-European derivatives from IE *Rer- 'horn' [with its many ablaut grades, as *kr-, *kera-, *kera-u-, *kera-i- ] : *ker[h.sub.2] (s), e.g., Gk keras 'horn', TochB karse 'stag' [< *'horned one'] or *R6ru 'horn', e.g. Lat cervus 'stag', Lith karve 'cow', Rus korova 'cow', Gk korudos 'crested lark', Av srva- 'horn; claw, talon' and further ON hrutr 'ram', Gk karnos 'sheep', krios 'ram', OInd srrigam 'horn', Hitt kar(a)war 'horns' (Gottlieb 16) (IEW 574) (G 345-6) (M-A2 137).

Thus, we would have: Bur kareelo < IE *ker-eelo (in Burushaski unstressed e>a), Bur karu < IE *kr[h.sub.2]-u- (or *kera-u- or *ker-u), and Bur [sup.2]khor < o-grade *Kor-. Bur krozo can be derived from IE *Rors- (like Gk Ion korse 'head') > *korz- (with a noted change s>z/r_) (19) + Bur plural morpheme -o and krizi < *kirzi < *kir(s)< *ker(s)- < *ker[h.sub.2](s)- (M-A 2: 137) or from < IE *kras- (like Gk Att kraspedon 'fringe') (IEW 575). Under one interpretation, Bur har 'ox' (B 191) could be ultimately derived from the same Indo-European stem, as there are instances where Bur h- seems to derive from k- (as noted by B 381; Morgenstierne 1945: 74; E-K 25, 29; Casule 2003b: 42) (see also [27]).

Further support for this firm etymology can be found in Bur karooyo (in Sh karoowo) Ng 'with curved horns' (B 242) (20) < IE krios with the same suffix *-yo- as in [2| huyoo and [27] luuyo (see the discussion therein). (21) Alternatively, it may be from IE *kerawo-s or *kr wo-s 'horned'. [In Pokorny (IEW 577), derivations from this precise form are registered only in Lat, Cymr, OPrus, Alb, Lith and SI.] The direct full parallel of the form of the Shina word with IE *kerawo-s 'horned' is remarkable. It is possible it may reflect an older Burushaski form.

[13] [T] Bur da gar Ng 'ram' (B 128). Berger suggests a tentative connection (which he finds dubious himself) with Skt dan gar a 'cattle, horned cattle' (T 5526?) (everywhere in IA, according to Turner it only means 'cattle'), which is possible, although the Burushaski word does not seem to be found in its Indo-Aryan neighbours, has a very specific semantics and Skt -ng- > Bur -m- or -n-

It can be compared to derivatives from IE *digh- 'goat' (IEW 222), i.e., from a form of this stem *[deigheh.sub.a] -, as e.g. Alb dhi 'shegoat' (< *deigheha-), and further: Arm tik 'leather skin', OEng ticcen 'kid', ?Gk di'za 'she-goat', Grm Ziege 'she-goat', Wkh ti [lambda] 'goat call', Ishk dec 'goatskin bag' (M-A 229) and possibly the Phrg gloss attagos 'he-goat' (D-N 95) (N 137). We cannot exclude the possibility that two stems could have merged here.

The final -ar in dagar could be an old suffix (-ru ?) in Burushaski, as it is found in a number of other animal names, e.g. thugar 'he-goat' (B 442) (a word that could be historically related), tulpar 'a very quick horse' (L 432), tatdr 'tailless rat' (B 445), butar 'male kid (the animal) (under one year old)', chinddr 'bull' (B 65), culdar Ys 'bull' (B 77), chiatar 'young goat' (B 76), chiar 'young bull, steer' (L 364), and maybe bisqar adj. 'raw, rough wool' (B 56-7) and bashkar 'wether, ram (over 2 years old, castrated' (L 72) < Ys (DC) beske, also biske, Hz Ng biske 'hair (of animals), fur'(see [9-10]).

On the other hand, we could be dealing with a compound word: *dag- 'goat' + *kar- 'horned' (see the discussion under [8] beskaret and [10] kareelo).

2.2.2.3. Female sheep, ewe

[14][VP] Bur belis (L also belis), Ys beles 'ewe (which has had young)' (L 75) (B 48).

Considering the m:b alternation in Burushaski (B I: 82) (for numerous examples see Casule 2003b: 28), we could correlate this shepherd term with IE *(s)melo- 'small animal', e.g. Arm mal 'sheep', Gk melon 'sheep and goats' (Wat 80). In M-A2 (142) the stem is given as *(s)mehil- 'small animal' with further examples: Olr mil '(small) animal)', Ukr mal 'small cattle (as collective stock)' (Gottlieb 18), ON smale 'small cattle', NDutch maal 'young cow', Eng small, OS1 malu 'small'--with a West-Central distribution in IE.

Another etymological interpretation, assuming p>b (22) (Berger 2008: 3.11) would be to derive it from *pel[h.sub.2-] 'bear young' (found in the West Central IE area): Alb pjell 'give birth to, produce', pele 'mare', Gk polos 'foal', Arm ul 'kid, young of deer or gazelle', Eng foal (M-A2 192) and thus possibly related to [5] sumphalikis etc., 'young female kid or lamb' (refer to the extensive discussion under [5].)

It is significant that we appear to have in this example, and in [15], the IE Nom. sg suffix -is (for further examples and discussion, see Casule 2003b: 71 and the fifteen names of body parts retaining this IE suffix in Casule 2003a). There is another possibility of interpreting it as the IE nominal suffix -es, for which see under [21].

[15][C] Bur meenis 'female sheep over one year old which has not had young' (L 264) (B 285).

There is a very strong probability that it may be related to an ancient European cultural word of very wide diffusion, particularly in the Balkans. Semantically, the Burushaski word is closest and almost identical to Rum iriinzare 'female sheep (for milking)' and mlnzarar 'shepherd of female (milking) sheep', mlnzar 'one year old lamb', mlnzat 'calf up to two years old' and Brancus (97-100) considers it a certain substratal (autochthonous) word in the Balkans. And further: Rum mlnz, mlndzu 'foal (up to one year old)', Alb (Tosk) mes, m'ezi, (Gheg) maz 'foal (up to two-three years old)', also mezat 'bullock', Messap Menzanas 'epithet of Jupiter' (to whom the Messapians sacrificed horses), Lat mannus 'small horse' (considered from a Paleobalkanic source in Rum, Alb and Lat), Grm (Bavarian) manz, menz 'sterilis uacca', minzekalb 'iuvenca', Itl manzo 'bullock', manza 'young cow, sterile cow' [= 'cow that hasn't given birth'], (in Basque mando 'mule'), Mir menn 'kid, young of an animal', MWels myn 'young goat, kid', usually derived from the IE stem *mend-, *mond- 'to suck, to feed young animals' (IEW 729).

In Burushaski we would have *mendis > meenis, with loss of -d-and compensatory lengthening of -e-. For the loss of -d- in the group -nd-, consider e.g. Bur hanik which B 189 (not found in the neighbouring IA languages) derives directly from OInd handika (T 14050) or Ys gendes : Hz Ng genis 'gold' (B 175). Both the specific semantics of 'young animal' and 'not having young' are present in Burushaski and it matches directly the Indo-European and especially the Rumanian developments.

2.2.3. Goat

2.2.3.1. He-goat

[16][VP] Bur buqheni NH 'goat with distinctive features on the head' (B 63).

Compare with IE *bhugos 'buck, he-goat': OIr boc, Wels bwch, OEng bucca, Eng buck, OHG bok (Gmc < *bhugn6-), Arm buc 'lamb', Av buza 'goat, he-goat', OInd bukka- 'goat, he-goat' (T 9312) (M-A 229) (Gottlieb 19, also includes Gypsy buzni). It may be an old loanword from Old Indian, yet everywhere in Indo-Aryan the form has -o- , and the word is not found in the surrounding languages. The -ni ending appears to indicate a closer correlation with the Germanic developments. The -qh-, instead of the usual reflex g (also not corresponding to Indo-Aryan -k-), may be due to expressive reasons, although there is a possibility that the older form would have been *bugni- > *buqhni- > buqheni (g > q(h)/_C). The same ending -eni (not found anywhere in T 9312) is also found in chageni NH 'black goat' (B 95). For the alternation q : g in intervocalic position see Varma (1941: 141).

|17]|C] Bur butdr 'male kid (the animal) (under one year old)' (L 90 also butar) and ga butar "Krahenzicklein", schwachliches Junges (von Ziege, Steinbock)' [ga 'crow'](L 90) (B 65).

We derive it from IE *uet-ru- (23) (*uetero) (< *uet- 'year', e.g. Gk etos 'year') similar to OEng wether 'wether', Goth wiprus 'one year old Iamb', OIcl vedr, OHG widar 'wether' < Grmc *wethruz perhaps 'yearling', with other suffixes: Lat vitulus 'calf, yearling' (Wat 101), OInd vatsd 'year; yearling, calf, Alb vitsh 'calf (IEW 1175). Note the close, specific semantic correspondence between the Burushaski meaning 'under one year old' and IE 'yearling'. It may derive from a suffixed zero-grade form *ut-ero- or *ut-ru. It could also have been influenced by forms like buc 'he-goat, two or three years old, not castrated' (B 60) (L 86) (possibly from Wkh bud registered by Lorimer), buqheni NH 'goat with certain distinctive features on the head' (B 63), budoko '(very young) foal; small horse, pony' (L 86) (B 61), bum 'ibex, markhor' (L 87) (B 62-3), busooso (24) 'calf (B 65) or bud (also buja) 'cow' (B 60) (this last word could be correlated with derivations from an Indo-European onomatopoetic *bu- like ON baula 'cow', Russ byku 'bull', Lat bubalus 'buffalo' (Gottlieb 1931:26).

A very strong indication that we may be dealing with a derivative from *uet-ru- are the following Burushaski forms, where Berger identifies a suffix -atar: as in budtar 'young cow, heifer' < bua 'cow' (L 86) (B 60) or chiatar 'young goat' (B 76). From a diachronic point of view, these could be compound nouns and the second component, (which clearly means 'young' ~ 'one year old' perhaps as in butar above) in these words could in fact be a noun: *bua- + *wet-ru- > *bua- + wat-ur > budtar. [For -w- in intervocalic position see e.g. alto-altar 'forty' (B 16) vs L altowaltar : altuwaltar : altualtar 'forty' (L 3).]

An alternative connection can be sought with IE *pou- : *pau- : *pu- (25) 'small, little; young (of animals)', esp. with the t-formant: OInd putrd 'son, child', Lat putus, putillus 'boy', pullus 'young, young (of animals)', Ose puklo- 'child', Balt-Sl *puta 'bird', Lith putytis 'young animal, young bird' (IEW 842-3). Phonetically a correlation is possible with Old Indian, although putrd in T 8265 everywhere in Indo-Aryan only has the meaning of 'son', and the semantics of the Burushaski word is more closely related to the Latin, Slavic and Baltic developments (yet note however Skt [pota-.sup.1] 'young of animal or plant', (e.g., H poti' 'young female of any animal') (T 8399). The Burushaski word is not found in the surrounding IndoAryan or Iranian languages and has a retroflex -t.

Both changes (and alternations): w- > b- and p- > b- are well attested in Burushaski (see Casule 2003b: 33, 40; and [9] and [14] in this paper).

2.2.3.2. She-goat

[18][C] Bur Ys hdlkit, Hz Ng elgit 'she-goat over one year old, which has not yet borne young' (L 38) (B 138).

It can be derived from IE *el-, *ol- 'red, brown (in names of trees and animals)' (IEW 302^4), in Wat 23, also without a laryngeal, or rather *[h.sub.1]elu- 'dull red' (in IE also: 'yellow; white; reddish, golden' (M-A 481) with the k-formant (Wat 23: *ol-ki) (in Gottlieb 14: 'deerlike [horned] animal'), as in OHG elho 'elk', ON elgr, OEng eolh 'elk'. Eng elk < Grmc *algiz 'elk', Russ losi, Cz, PI los 'stag', Pam rus 'wild mountain sheep' [borrowed into Bur rus* 'Pamir wild sheep, Ovis Poli' possibly through Wakhi (B 366)], also with an -n- suffix, e.g., OCS jeleni 'hart', Lith elnis 'deer', Olr elit < *eln-ti-s 'deer', Arm eln 'hart', OPrus alne 'animal', and from *l-on-bho-s > ON, Goth lamb 'sheep' (IEW 302-4). Some newer etymological interpretations (Adams 1985, apud M-A 177-178) reject the link of the animal name with the colour term, and propose a form *[h.sub.x]lkeis or *[h.sub.x]olkis elk'.

The derivation of the Burushaski word would be < IE *[h.sub.1]el-k-ti-. It is highly significant that we have the basic colour form in Bur hal 'fox; yellowish, reddish' (B 186) (a < e in unstressed position, consider the pi. form hal jo) (see Casule 2003b: 46-7). The Hunza and Nager form maintains the e-vocalism after the loss of the laryngeal, and it appears that in Yasin we have a form influenced either by the colour term or by some semantic analogy to Bur halk- 'to give birth', halkis 'womb', Ys Bur halkis 'pregnant cow' (B 188). This correlation is suggested by Berger (2008: 74), but that in itself does not appear to explain the Hunza and Nager forms, nor the suffix -it, or the change in the position of the stress. The vacillation in assigning a laryngeal to this Indo-European stem is perhaps reflected in the dual dialectal outcomes in Burushaski, with and without a laryngeal.

Another very tentative possibility is to analyse Hz, Ng elgit as the primary form, i.e., as a compound word: *el- + *git, where the second component in turn could be correlated with IE *gh(a)id-o- (Watkins 28, who following Pokorny stipulates that this stem could be possibly related [by metathesis] to *digh- [for which see [14]], Lat haedus, Goth gaits 'goat', OEng gat 'she-goat', Eng goat [in IEW 409: *ghaid-o 'goat', who also notes Alb qith 'goat', Mir cit 'sheep', ON kid 'young of animals'], Gottlieb (1931: 16).

A conflation of two forms cannot be ruled out.

Burushaski also has [19][P] halden 'male goat (full grown, and in the case of domestic goats, castrated); a male ibex' (Wil 63). The first component would very likely be related to the examples above and the second could be the Burushaski word den 'year; age' (B 118) as in jimden 'any time in the future' (B 227). The semantics would be 'over one year' ~ 'of age'. [Another etymological interpretation could be < *[h.sub.1]el-t-en, by metathesis < IE *[h.sub.1]eln-ti- (?)]

2.2.4. Young of animals

The following term is used to refer to young animals in general, including lambs, kids, calves, kittens, chickens etc.

[20] [C] Bur -sk, NH Bur -sko, Ys -is 'young (of animals), young one; also jokingly for children', e.g., (belise) alta iskumuts bie 'there are two young ones (lambs)', or bus* isk 'kitten' < bus 'cat' (L 316) (B 380). It has the force of a diminutive.

There is a direct connection with IE *-i-sk 'formant of adjectives and noun diminutives' (Illic-Svityc 1976 I: 204, who indicates that the -i- is probably from the i-stems, a continuant from many old root stems), in Wat (36) IE *-isko, compound adj. suffix, forming relative adjectives, denoting origin in Slavic, found also in Germanic and Thracian (for the latter, see Illyes 1988: 212). We have discussed the Burushaski adj. suffix -ki, -ski (e.g., Burusaski < Burusin (B 491) and the related -ko-, -kus, -kus with identical functions in Casule 2003b: 71-2, also Bur -urn adj. and participial suffix < IE -enko, -nko- (composite suffix with -ko as the second component [Wat 36]). This example shows a full systematic derivational correspondence.

2.2.5. Other shepherd vocabulary

2.2.5.1. Shepherd

[21] [VP] huyeltarc 'shepherd, herdsman' (L 211) (LYs 131). B (209) also notes huyeltarci 'pasture for sheep and goats' and analyses it as [1] huyes + ?, the second component also in Hz Ng bualtarc, Ys bietarc 'cowherd' (B 60). For the second component, there doesn't seem to be a veritable internal derivation, although Lorimer (questioned by Berger) points to the verb Hz Ng Ys --ltir- 'show, point out, indicate', Ys also -ltar- , in DC Ys also Ys-lter- (B 269), which is semantically imprecise, and phonologically we would have expected a form *huyeltirc. The ending -c is a suffix, as in daruc 'hunter' < daru 'hunting' (B 116) or maruc 'gold washer' < maru 'washing river sand' (B 282), durac messenger' (B 125). Berger (2008: 144) derives through internal reconstruction the suffix -c < -es (also -is). It most likely corresponds to IE nominal suffix -es in animate stems with possessive semantics, as in IE *[kerh.sub.1]-es- 'grain' > *kerhi-es- '(the one) associated with or possessing grain' (Fortson 2004: 112). Note perhaps further -es, -is in [14] Bur belis (L also belis), Ys beles 'ewe (which has had young)' (L 75) (B 48) if from IE *[pelh.sub.2]- 'bear young' (found in the West Central IE area): Alb pjell 'give birth to, produce', pele 'mare', Gk polos 'foal', Arm ul 'kid, young of deer or gazelle', Eng foal (M-A2 192), thus: 'bear young' > 'the bearer of young'. The same analysis can be applied to [15] Bur meenis 'female sheep over one year old which has not had young' (L 264) (B 285), which we derived from the IE stem *mend-, *mond- 'to suck, to feed young animals' (IEW 729).

We find a very important close parallel with the Thracian element trair, tral, tra in personal names like e.g. Trai-bithus, Traii-centus, Aulou-tralis, Auro-tra, Aulu-tra, Muca-tra for which Tomaschek (1980 [1893-1894] II, 2:38f, 102) proposes the meaning 'shepherd, herdsman', from an IE stem *tra- 'to feed, hold, protect, shield, shelter, look after, care for' (not found as such in Pokorny) and points to SI trajati 'durare' and Skt Agni-tra as possibilities for comparison.

The stem may be linked derivationally with [24] Bur tark 'byre, hut for animals' (L 346) (B 422), with a basic shared semantics of 'holding, sheltering, enclosing' and can be segmented as *huyes+tra(l)+-c 'one who shelters, looks after the sheep'.

Another etymological possibility within Indo-European is the stem "*[terh.sub.2]- oldest form with variant [metathesized] form *[treh.sub.2] -, colored to *[trah.sub.2]-, contracted to *tra-, also zero-grade form *tr(a)- 'to cross over, pass through, overcome' [for derivatives in Bur from this stem that follow typically non-Indo-Iranian forms see Casule 2003b: 66- 7]: e.g. Skt tirati, tarati 'he crosses over', trayati 'protects, shelters' (widespread in IA but without the pastoral element), Irn *thraya- 'to protect', Lat intrare 'enter', Hitt tarhzi 'defeats' (Wat 91) (M-A2 395), Av tar(v)- 'to overcome', H tarna 'save' (G 629-30). It is the current preferred etymology also for SI trajati 'endure, last; wait' mentioned by Tomaschek above, which some (G 634-5) link further with PS1 *tSrjati 'drive', e.g. Med tera 'to drive (out) (esp. cattle, small and big); to force, make; to chase, pursue', Srb terati, Croat tjerati (without Baltic parallels).

Overall, the correlation and connection with Thracian and with [23] tark 'byre for animals' with a Paleobalkanic etymology, appears to be the most promising one.

[22] [P] Bur [dak.sup.1] 'driving animals', dak etas 'drive an animal; follow up tracks', dak otas hir 'shepherd' (lit. "driving animals man") (L 108) (Will 46). Berger (B 128) unjustifiably gives it together with dak2 'to hammer; smithery; sharpening tools', dakaalo 'blacksmith' (which only and specifically means 'to hammer'), whereas Lorimer and Willson correctly give the two forms as homonyms. Zoller (220) gives for Indus Kohistani dag gho 'to drive (cattle)'. First word no separate meaning, with parallels only in WPah, where it means 'head of cattle', and points to T 5524a, which has the meaning 'cattle', and not 'driving animals'.

Can be compared with IE *deik-, *deig- 'to show, pronounce solemnly, also in derivatives referring to the directing of words or objects', with semantics in IE of 'throw', 'justice, right', 'show, teach', 'sign, mark', 'sample, pattern', 'direction', 'region', 'meadow, pasture', 'rule, cannon, measure' (Wat 14) (M-A 158), e.g., Gk dikein 'to throw', deiknunai 'to show', OInd di's- 'direction', ON tegr 'strip of land', OHG zeiga 'directions', Av daesa 'direction, region', OEng tik ~ tik 'meadow, pasture', and from the variant *deig- OEng taecan 'show, teach', ta(c)en 'sign, mark', Goth taikns 'sign', Lat index 'indicator, forefinger' (< 'pointer') (M-A 159).

In Bur ei > a and k > k (see Casule 2003b: 31-2, 38). The Burushaski semantics of 'to drive' = 'direct (show) the animals, to follow up tracks (signs/ fits well within the IE developments.

2.2.5.2. Byre for animals, goat house, sheep house [23] [C] Bur tark 'byre, hut for animals', also tarkari 'stable', Ng 'stable for horses' (Berger notes that the latter word has an older pi. ending -an of tark) (B 422) (L 346).

Can be compared conclusively with a Balkano-Carpathian word believed to originate from one of the ancient Balkan languages (Thracian?): Alb thark, cark 'byre for animals', Rum tare, Arum tarku 'winter byre for sheep; fence around stack' (Neroznak 1978:207). It is also found in PI Slk Ukr Hung Mold (Bernstejn 1988: 139), borrowed into Gk as tsarkos. Rasmussen (1999: 648-9) correlates the Albanian word internally with Alb thur 'embrace; fence in' and further with Lith tvarka 'Haltung, Fassung, Ordnung', tvora 'fence, hedge, borderwalP < Lith tverti 'embrace, enclose, fence in', ultimately from IE *tuer-2 , *tur- *tuera- 'to grab, enclose' (IEW 1101). In Albanian, *tu- > th- [[theta]], the intermediate stage, according to Rasmussen (649) being [ts] ("with which one may compare the development of *tu > s(s) in Greek"), and who goes on to say that "the details and the antiquity of the Albanian word are proved by its appearance as a loanword in Rumanian (...) and Modern Greek (...)".

There are numerous examples of a t: th alternation in Burushaski, e.g. Hz altar : Ng Hz althar (B 16), Ys ti'san : Hz Ng this an (B 440), tanii : thanif (B 419), Ys tarkon : Hz Ng tharkoon (B 438), tar del- : thar del- (B 421), etc.--we have noted seventeen such examples in Berger and a further sixteen examples in Lorimer.

In light of this, [23a][VP1 Bur thaark (in Sh thraako) 'walled enclosure (constructed around a shrine, or a saint's grave)' (L 337) (B 435) may well belong here, esp. as Lorimer gives it as thark in the dictionary entry, but as tark in the example. We have already noted (Casule 2004: 84) the possibility that the Phrygian word terkos 'burial plot' from the inscriptions (Orel 1997: 461, without etymology) could be correlated with the Burushaski words.

In Burushaski we would have tark < *tuar-k(a) < *tuer-k(a) (note the variation ue : ua , as in [15]), and t(h)aark < *tuerh2-ka (with the lengthening caused by the lost laryngeal). For the suffix *-k(a) see [8] and [9].

The exact match with a word belonging to the most ancient Balkan layer of shepherd vocabulary is very important.

[24]|P] Bur Ys bac 'goat house, sheep house' (LYs 398).

There is a close correspondence with a Balkan word considered of ancient substratal autochthonous origin and with Carpathian distribution (Neroznak 1978: 204, Alinei 2003: 51): Rum, Megrum baci 'older shepherd, cheese-maker', Arum baciu, bagiu, Alb bac 'same', Med bac 'shepherd in charge of a summer mountain pasture', bacilo 'pen, enclosure in the mountains where milking is done and cheese made', also Srb and Croat bac'shepherd', bacija 'pen for sheep', PI bacza 'Tatra mountain shepherd', 'head of the young shepherds', reg. Cz baca 'shepherd', Hung bacs, bacsa, bacso 'same' (Skok 1974/1: 85). Skok postulates a Thracian origin.

Even though the correlation with Burushaski involves a slight semantic change, this is an interesting correspondence.

2.2.5.3. Grazing ground

[25|[VP] Bur run'alpine pasture, open grazing ground on hills, grassland' (also in Sh) (L 305: Hz ruun) (B 366) (Will 100).

Can be compared with IE *reua- 'to open, space', suffixed zero- grade form *ru-mo-: OEng, OHG rum 'space', OProv run 'ship's hold; space' suffixed form *reu(a)-es : Lat rus 'open land, the country [countryside, open fields]' (Wat 71), in M-A 534: IE *[reuh.sub.x es] (< *[reuh.sub.x] - 'be open') and further OIr roi 'field, open land', Av ravah- 'space', Goth rums 'open space', TochAB ru- '(be) open', OCS1 ravlnu 'level'. Not found in Indo-Aryan.

We note also Bur rat 'smooth, flat, level' (also in Sh) (L 302) (B 364, who gives Balti rat, rad with the same meaning). In Burushaski, if it is not a Tibetan word (and not found in RYTEDD) it could have developed from the stem above < *reu-to, with the semantics as in the Slavic example.

In Bur run, Hz ruuh < *ru- (with the length preserved in Hz) + Bur adj. suffix -un (B I: 5.1) which we have derived from the IE composite adj. suffix *-nko- with the meaning of 'open' (adj.) at the core of the Burushaski word, i.e. a substantivised adjective = 'the open'. It could on the other hand, be a plural form with the pi. ending -n as e.g. bada sg. 'sole; step, pace' : pi. badan (for examples and discussion, see Berger 2008: 11.4), leaving the stem as ru-, subsequently with a double plural form (on the Bur double plural, see B I: 3.32), i.e., runanc.

Another indication that the stem could be ru- is possibly the Ng pi. ruanc vs Hz runanc (Berger 2008:57, who apud Anderson 1997: 1035, gives it however as an example of dissimilation of nasals). Only two more examples of a dissimilation h-n > [ensemble vide]-n, are cited by Berger, and in both cases we cannot exclude the possibility that the -n in the singular is in origin from an adjectival or a plural ending.

[26][VP] Bur ter 'summer grazing ground up in the mountains, mountain pasture (with some cultivation and farming)' [note the specific semantic correspondence with the Rumanian and Latin examples below] [in contrast to [25] run]; 'in plant names "wild"' (L 350) (B 425). Note further the Burushaski toponym Baltar 'village near Chalt above the Baltar glacier' which Berger analyses as ? + ter 'ground' (B 489).

There is a possibility of comparison with Lat terra 'earth, land, country, the earth', derived by some etymologists from IE *ter'spread out, extend' and further to OIr, Wels, Bret tir 'land, country' (Tucker 241) (E-M 687-8). The Rumanian Jarlna 'field under cultivation' (not found in Albanian) was considered by Russu (apud Illyes 232) to be a Balkan substratal remnant, although Neroznak (1978: 204) proposes a Latin origin for it, i.e. from Lat terrena 'plough field, arable land'. Another etymological analysis links tentatively the Latin word to IE *ters- 'dry': Lat torrere 'dry up, parch', Gk tersomai 'dry up', Skt trs- 'be thirsty', Eng thirst (Buck 17) (G 639) (Wat 91).

In regard to the possible wider Indo-European links in support of the 'spread, extend' proposal we can point to Bur tirkan '(of spilled oil, wound) to spread' (B 427), tarau manas 'to spill over, overflow (of things on a big scale)' (L 338). In connection with the 'dry' hypothesis note Bur tir man- 'to dry (of woman's hair)' (B 446).

2.2.5.4. Wool

See also the discussion under |9] Ys beske, also biske, Hz Ng biskd 'hair (of animals), fur; wool' and the first component in biskepalkis 'wool-carrying, wool-giving animal' (L 84) (B 56) and perhaps also bisqar adj. 'raw, rough wool' (B 56-57).

[27][VP] Bur luuyo 'a tuft of wool; impurity in the wool; small balls of scraped wool, lint' (B 271).

It can be correlated with IE *uj[h.sub.2][neh.sub.a]- 'wool' (with different enlargements also 'hair; ear of corn; forest') distributed in nine different groups, e.g., Lat lana, Gk lenos, Skt urna-, OEng wul(l), NE wool, NWels gwlan, Lith vllna, Med volna, Hitt hula na-, Av varana-, all: 'wool' (M-A2 177-8) (Wat 98). The Indo-European word is considered a likely derivation from IE *uel- 'to tear, pull' (Wat 98) (IEW 1144), *uel[h.sub.2]- (M-A 150) e.g. Lat vellere 'to tear, pull', villus 'shaggy hair, wool', with derivatives in IE meaning 'wound', 'blood', 'fight', 'captured', 'die', etc.

It is very indicative that Burushaski has the verb waalas 'to get lost, to go astray; to be separated from work or family; to die' (Will 73) (possibly < *u-baalas) (< IE *u [elh.sub.2]- : e > aa caused by the laryngeal). Burushaski also has bei-, Ys wel- 'to put s-thing on, to don, to wear' (B 47) (L 77, 385) (Will 22) that could be traced to IE *uel- 'tear' with a semantic development as in PS1 *ob-velkti 'to put on clothes' < *velkti 'to pull, drag' (related stem: IEW 1144-5) (for the semantics also note Lith drapanos 'clothes', Skt drapi 'cloak, mantle' < IE *der- 'tear' (Wat 16). The Burushaski meaning of 'scraped wool' might reveal the underlying meaning of 'tear'.

We would derive Bur luuyo < *ulun-yo, with a suffix -yo- which we can correlate with the IE relational adj. suffix *-yo- , also 'of or belonging to' (Wat 103), also *-ijo- (Fortson 120-121) (and retained in huyoo 'wool-bearing animal, sheep' (L 208) (B 209), and further ma may o 'endearing term for 'mother' < mama, ma mo 'mother' (B 277), sooyo 'a type of vegetable' < Shina soo (B 397), karooyo 'with curved horns' (B 242) (see under (101), etc- In Burushaski we have evidence of a change ny > y: e.g. Ys nyam : Hz Ng uy am 'sweet, tasty' (B 460), Ys nya : Hz Ng ya 'bear' (B 467) etc. For the loss of the initial u- in front of -1 note e.g., the doublet Ng ulan- 'can, to be able' : Ys lan- 'same' (B 454) or Hz Ng -ulgis 'nest' : Ys -lkis 'same' (B 454). u- is a positional variant of w- /_l or C e.g., Ys welji: Hz Ng -ulji 'dream' (B 454). A metathesis ul > lu with lengthening cannot be ruled out.

Perhaps related to luuyo, from the same Indo-European stem above, is Bur iilikis 'ear of corn, just as it appears' (B 211), with *ifrom the pronominal prefix, and with the alternation i:u/_l : *ul-yokis > *i-ul-yo-kis > iilikis (for the change yo > i in the inlaut see in [5]). The semantic development is the same as in other Indo-European derivatives, e.g. Corn gwels 'grass', Lith vdltis 'panicle of oats, of corn', OPrus wolti 'ear of corn', Srb vlat 'same', Ukr volou 'panicle' (IEW 1139) which strengthens the etymological analysis.

[28] [VP1 Bur se 'wool' (L 326) (B 393).

A comparison is possible with developments from IE *kes- 'to comb, scratch, itch': Mir cir , Hitt kiss- ~ kisa(i)- 'to comb', kiske/a 'comb', Luw kis 'to comb', Lith kasa 'braid', OCS kosa 'braid, hair', cesati 'to scratch, to comb', and specifically in regard to wool, as Gk ksaino 'scrape, comb [hair or wool], full [cloth]', or flax, as in OEng heordan [pi] 'hards [of flax], tow' (M-A2 233), with differing semantics in OInd kacchu 'itching' or ksura 'knife, thorny plant' [ > Bur as chur 'big knife' (B 107], NPers sor 'salty' (IEW 585-586) (G 175)'(Trubacev 1974, IV: 86).

The Burushaski form would be from an extended zero-grade: ks+ -e(s) - directly correctable semantically with Gk ksasma 'crumpled wool' (< *ks-en-), also kseo 'grate, scratch' (*ks-es-5) (IEW 586) as well as with the Slavic and Baltic examples above.

It is important that Burushaski has also Ys hesk, Hz Ng hisk 'comb, loom, wrist' (L 204) (B 200) (for the suffix -k(a) see in [81), which would be a derivation from the e-grade. There are a small number of examples in Burushaski where h < k (see the note and references in [10]), which would explain the h- instead of the prevalent k-. It could be the result of dissimilation.

The following two words belong more to textiles vocabulary, but they are also indirectly related to the pastoral terminology:

[291 [CI philam 'woollen homespun cloth' (L 290) (B 329) (Will 94).

We can compare it with IE *pel- 'animal skin, hide; cloth; wrap' (Wat 63), in IEW 803 *pel-no < IEW 985-986 *(s)p(h)el- 'tear off, M-A 268-9: *peln-, B-K 60: *p[h]al- / *p[hlel- : especially Lat pellis 'animal skin, hide', ON fjall 'skin', OEng fell 'animal skin, hide; pelt', OHG fel 'same', OPrus pleynis 'meninges', Lith plene 'film (on milk); scab', Lett plene 'membrane', Russ plena 'felt', Gk pelloraphes 'sewing skins together', Gk spolia 'fine wool plucked from the legs of sheep', and further OEng filmen 'film, membrane, foreskin', Gk pelma 'sole of the foot', Russ pleva'membrane', also Lat spolium '(animal) skin, hide', (< *(s)pel- 'tear off) (according to M-A, widespread, though not universal, in late PIE).

In Burushaski we would have a derivation of philam < IE *pilman-< IE neut. *pel-mn- , as in Ys hasuman 'star' < haas 'glowing ambers', hanuman 'alone' < han 'one' + -man, or hoom 'sign, omen' (short form of the suffix) < IE *([h.sub.2])[oh.sub.3]- 'believe' > Lat omen 'sign, omen' (for a discussion of the continuants of the IE suffix *men in Burushaski, see Casule 2003b: 48-9, 58-9). Furthermore, IE p- > Bur ph- (for examples, see Casule 2003b: 34). Perhaps the second component of the compound noun Ng gurpaltiri, gulpaltin, Hz, Ys gupaltin 'trousers (man's or woman's)' (L 173,4) (B 161) can be correlated here, linked by Berger (2008: 146) to -phalt- Hz Ng trs. 'break, break up, hoe, dig a hole, burst in the air' (B 322) (see below).

From the underlying IE verb *(s)p(h)el- 'to split [break in two], cut off, tear off; board' (IEW 985-7) we have: OInd phalati 'breaks, cuts in two' (T 9057), patati 'blows up', Ksh phalun 'to break in two' (and with wide distribution in Indo-Aryan, yet apparently not found in Shina or Khowar) and further Gk sphallo 'trip, knock over, separate, hesitate' etc. For the correlation with Burushaski especially important is the version of this stem with the-t-extension, i.e. < *(s)pel-t- : Grm spalten 'to split, to cleave, to chop', Ir alta(i)n, Cymr Ellyn, Bret aotenn all: 'razor', PS1 *poltiti-'to cleave', e.g. Med poluti 'break in two' (G 493).

Directly correctable to this stem is the Burushaski verb--palt-'to cause to break',-phalt-Hz Ng trs. 'break, to break up, to hoe, to dig a hole, to burst in the air' (L 153, also iphalt-'break a hole in the wall') (B 322).

The extended -t- stem, the absence of a retroflex, the alternation p : ph in the Burushaski words and the fact that the verb is not found in the neighbouring Indo-Aryan languages argue strongly against a loanword from Indo-Aryan or Old Indian and reinforce the etymology of this pastoral and textiles term.

[30] [C] Bur gay 'thread in a warp, i.e. for weaving' (B 175).

It shows a remarkable correspondence with IE *[g.sup.w]heie-, *[g.sup.w]hi-sinew, thread' (IEW 489), in M-A (569): *[g.sup.w][hih.sub.x](-[eh.sub.a])-, e.g., Lith gija 'thread (in a warp), skein, hank (of yarn)', Lett dzija 'thread', OS1 zila 'sinew', Wels giau (pi.) 'nerves, sinew', Lat filum 'thread', Arm "jil 'cord' etc. (not found in Indie or Iranian). The Bur word is semantically precise and specific (identical to Lithuanian), whereas phonetically it could be derived < *[g.sup.w]heie-(ei > a with the laryngeal > y or y in intervocalic position, as e.g., in Bur giy-'go into, enter' (B 155) < IE *[gheh.sub.1]-'let go; go' (Wat 28), or in Bur-yaanis 'so much' (B 468) explained by B < han 'one' (< IE *[h.sub.i]oi-no-s 'one') (see Casule 2009).

2.2.5.5. Flute

[31][PI Bur pureelo 'a (type of) flute' and purururo 'flute player' (B 318). Possibly related in the sense 'flute made of reeds', we note Bur phuruu and phuruuy 'reeds, rushes' (L 294) (B 337, who does not make that correlation). For the alternation 1 : r , see e.g. phulguuy : phurguuy 'feather' (B 335) which we have derived from (as a compound word) < IE *pleus-'feather' + *gour 'body hair, lock of hair' (Casule 2003b: 23). It is possible that the original Burushaski form was *p(h)uleero.

There is a very interesting parallel with a Balkan substratal word: Rum fluier 'flute', Arum fluer, fuleru, Alb flojere, floere, floer 'flute', a typical pastoral term also found in Ukr flojara, PI fujara, Slk fujara, Srb and Croat frula, Hung furulya, MGk flogera, all: 'flute'. There is no firm accepted etymology for this word (Brancus 1983: 75), although some have derived it from Gk floiarion 'bark of tree', and others from Lat flare 'to blow', or as an imitative word.

The semantic correspondence with Burushaski is direct and the metathesis of the liquids would be as in the Serbian, Croatian and Hungarian examples, unless Burushaski displays the original form. Bur does not have [f]-in loanwords it is substituted with ph or p and there is undoubtedly some kind of assimilation in purururo.

3. Discussion and conclusions

Shepherd vocabulary is generally a conservative layer in the lexicon of a language. This can be seen throughout the Indo-European languages and especially very clearly in the case of the Balkan languages. If we look at Poghirc's (1969: 327-356) identification of possible old, substratal words (remnants from the ancient Balkan languages) in Albanian and Rumanian (and their number goes from more widely accepted 122, according to Poghirc, to 210 according to Russu [apud Du Nay 1977: 60-1]), -30% of them belong to shepherd words in sensu stricto, by far the largest compact semantic group, but if related words are added, as pointed out by Du Nay (1977: 71) the number increases to 58% of all the words originating from the substratum.

We find the same pattern in Burushaski. It is highly significant that only in reference to 'sheep' and 'goats' we have identified 20 autochthonous Burushaski words of non-Indo-Iranian Indo-European origin. If we add to this number the 9 loanwords from Indo-Aryan and Persian (some of which could have coincided or merged with the Burushaski autochthonous vocabulary), we end up with 29 words of Indo-European provenience in this very restricted semantic field. This amounts to 90% of this vocabulary in Burushaski. In addition we have a further 11 words of closely related shepherd terms (and upon closer inspection this number may be larger), which brings the total to 40.

In many cases, as in [1| Bur huyes Ys also: huis (sg and pi) 'sheep and goats)' and [2]; [3] Bur adds 'sheep, goat; small cattle'; [10] Bur kareelo 'ram', also [11-12] (with many derivations); ?[13] Bur dagar Ng 'ram'; [14[ Bur belis (L also belis), Ys beles 'ewe (which has had young)'; [16] Bur buqheni NH 'goat with certain distinctive features on the head'; [17] Bur butar (also butar) 'male kid (under one year old)'; [18] Bur Ys halkit, Hz Ng elgit 'she-goat over one year old, which has not yet borne young'; [22] Bur ddk1 'driving animals', dak otas hir 'shepherd'; [25] Bur run (also ruun)'alpine pasture, open grazing ground on hills'; [27] Bur luuyo 'a tuft of wool; small balls of scraped wool'; [28] Bur se 'wool'; [29] philam 'woollen homespun cloth'; and [30] Bur gay 'thread in a warp, i.e. for weaving' we have examples of words of clear Indo-European antiquity, with no semantic latitude and with non-Indo-Iranian phonological traits.

In a few instances, in widely represented stems in Indo-European, we have specific semantic and derivational developments that align Burushaski with some of the IE groups, e.g. in [26] Bur ter 'summer grazing ground up in the mountains, mountain pasture (with some cultivation and farming)', where we find direct semantic correspondence with the Lat, Rum, Wels and Bret derivatives from a widespread IE stem, or under one interpretation with the Paleobalkanic substratum.

In one case: [5] Bur sumpal, sumalkis, su mphalikis, sumapalikis 'young female kid or lamb', we find a specific development within Burushaski connected to a stem found in the West Central IE area, whereby Burushaski has several derivations from it.

The most interesting findings are in the direct and specific semantic, derivational and phonological correlations of Burushaski with the ancient Balkan shepherd vocabulary.

Firstly we have a correspondence with an ancient European cultural word, well attested and identified as substratal in the Balkans, with a specific semantic shift to 'female sheep' both in Rumanian and Burushaski (in the rest of Europe it means 'foal', 'bullock', 'young cow', 'kid', 'young of animal'): [15] Bur meenis 'female sheep over one year old which has not had young'.

We have further Paleobalkanic correlations in: [4] Bur bat 'goatskin, sheepskin; animal hide; a shape, a form, a likeness', which has an attestation in IE that assures antiquity and considered by some a possible loanword from a non-IE source, or from Thracian. Especially strong and convincing is the specific correlation of [8] Bur Ys beskaret, Hz Ng baskarat 'wether, ram (over two years old, castrated)' < [91 Ys beske, also biske, Hz Ng biske 'hair (of animals), fur' with a [Thracian?, Ancient Macedonian?] substratum borrowing in Albanian and Rumanian, with Burushaski also having the probable underlying verb. Very strong, direct and specific is the link between [23] Bur tark 'byre, hut for animals' and Albanian and Rumanian substratal words (+Baltic) and also [23a] thaark 'walled enclosure (constructed around a shrine)' (+Phrygian). There is another Burushaski-Paleobalkanic (and wider Carpathian) correlation in the case of [24] Bur Ys bad 'goat house, sheep house' (LYs 398) and [31] Bur pureelo 'a (type of) flute' (also with Carpathian distribution). In one instance, in the second component of [21] huyeltarc 'shepherd, herdsman' we find a precise correlation with Thracian. Note also the discussion under [6] du 'kid'. (26) Especially important is the fact that the two attested Phrygian glosses for 'goat': aseis and attagos, correspond closely with Bur acas 'sheep, goat' [2] and dagar Ng 'ram' [13]. This means that in 10 [if we include the Phrygian correspondences: 12], out of 31 (-30%) of the Indo-EuropeanBurushaski correspondences we can trace a direct correlation with the Paleobalkanic developments, with very specific and identical semantics and systematic phonological correspondences. Considering the scarcity of Paleobalkanic attestations, this is an outstanding number.

Of course this does not mean that these words are necessarily Thracian (27) in origin both in the Balkans or in Burushaski, even if their Paleobalkanic correlation appears to be indisputable. Except for [21] and even then with a conjecture, the substratal words are conveniently hypothesised to be Thracian principally because of the ancient spread of Thracian. We have so little preserved from the Paleobalkanic languages that some of these words could be of Phrygian or Ancient Macedonian origin or from languages almost not attested at all, like Paeonian or mostly onomastically attested languages like Illyrian, or maybe even shared by several of these languages.

The presence in Burushaski of a number of shared alternations and changes like betacism w>b (as in Ancient Macedonian, Phrygian, see e.g., Georgiev 1981: 129) or w:b (Thracian) see [9]), or voicing of k(as in variant Latin and Greek transliterations of Thracian onomastics, see Casule 1998: 67-9) or voicing of p-(as perhaps in Thracian, see [4J), (28) the alternation m:b (29) (as in Thracian names--see e.g. [14]) makes this link even more viable.

In this regard, these findings fit in with our earlier general analysis of other Paleobalkanic-Burushaski correspondences in Casule (1998: 13-18, 21-36). (30) Furthermore, as we have shown in detail (Casule 2004), there are other specific semantic fields like ritual, myth, burial and onomastics where we find close and numerous correlations of Burushaski with Phrygian, originally a Paleobalkanic language. (31) In our analysis of names of body parts of Indo-European origin in Burushaski (Casule 2003 a) we also found possible connections with Phrygian in five of these words.

Our analysis of the shepherd vocabulary of Burushaski reveals a language where almost all the numerous words for 'sheep' and 'goats', together with a large number of related terms can be derived from Indo-European, belonging to a community that from antiquity would have been of a markedly pastoral character. It is highly significant that the specific phonological correspondences manifested in this material are consistent with the

other >550 lexical correspondences we have analysed in our previous work. Together with the large number of systematic grammatical correspondences with IE, crucial in asserting a genetic relationship, this is further proof of the Indo-European (outside of Indie and Iranian) origin of Burushaski. The strong and precise links in the Bur shepherd vocabulary with the ancient Balkan substratal lexis advance the hypothesis that Burushaski is the descendant of one of the ancient Balkan languages, very probably Phrygian or a language related to it.

Abbreviations of languages and dialects

Adygh--Adyghian, Alb--Albanian, AncMcd--Ancient Macedonian, Arm--Armenian, Arum--Arumanian, Av--Avestan, Bait--Baltic, Big--Bulgarian, Bret--Breton, Bshk--Bashkafik, Bur--Burushaski, Celt--Celtic, Corn--Cornish, Croat--Croatian, Cymr--Cymric, Cz--Czech, Dom--Domaaki, Eng--English, Finn--Finnish, Gk Greek, Gk Att--Attic Greek, Gk Ion--Ionian Greek, Goth--Gothic, Grm--German, Grmc--Germanic, H--Hindi, Hitt--Hittite, Hung Hungarian, Hz--Hunza dialect of Burushaski, IA--Indo-Aryan, IE Indo-European, Illyr--Illyrian, Ind--Indian, Ind Koh--Indus Kohistani, Ir--Irish, Irn--Iranian, Itl--Italic, Kab--Kabardian, Kal--Kalasha, Khw--Khowar, Ksh--Kashmiri, Ku--Kumaunl, Lat Latin, Lett--Lettish, Lith--Lithuanian, Med--Macedonian, Megrum--Meglenorumanian, MEng--Middle English, Messap Messapian, MGk--Modern Greek, MHG--Middle High German, MIA--Middle Indo-Aryan, Mir--Middle Irish, Mold--Moldavian, Myc--Mycenean Greek, MWels--Middle Welsh, Nep--Nepali, Ng--Nager dialect of Burushaski, NIA--New Indo-Aryan, NPers--New Persian, OCS--Old Church Slavonic, OEng--Old English, OHG Old High German, OInd--Old Indian, ON--Old Norse , OPers Old Persian, OPrus--Old Prussian, OS1--Old Slavic, Osset Ossetian, OSwed--Old Swedish, OWels--Old Welsh, PAK--ProtoAdyghian-Kabardian, PAlb--Proto-Albanian, Panj--Panjabl, Pers Persian, PGrmc--Proto-Germanic, Phrg--Phrygian, PIE--ProtoIndo-European, Pk--Prakrit, PI--Polish, PNC--Proto-North Caucasian, Prov--Provencal, PS1--Proto-Slavic, Rum--Rumanian, Russ--Russian, RV--Rgveda, Sh--Shina, Si--Sinhalese, Skt Sanskrit, SI--Slavic, Slk--Slovak, Srb--Serbian, SSI--South Slavic, Thrac--Thracian, Tib--Tibetan, Toch A, Toch B--Tocharian A, Tocharian B, Turk--Turkish, U--Urdu, Ukr--Ukrainian, Umb Umbrian, Wels--Welsh, WPah--West Paharl, Ys--Yasin dialect of Burushaski

Abbreviations of sources cited

B = Berger, Hermann. 1998.

B-K = Bomhard, A. and Kerns, I. C. 1994.

BYs = Berger, Hermann. 1974.

Cunn = Cunnigham, A. 1854.

DC = Tiffou, E., Morin, Y.C. 1989a.

D-N = Diakonoff, I. M. and Neroznak, V. P. 1985.

E-K = Edel'man, D. I. and Klimov, G. A. 1970.

E-M = Ernout, A. and Meillet, A. 1959.

G = Gluhak, Alemko. 1993.

G-I = Gamkrelidze, T.V. and Ivanov, V.V. 1984.

Hay = Hayward, G. W. 1871.

IEW = Pokorny, Julius. 1959.

L = Lorimer, David L.R. 1938.

LI = Lorimer, David L.R. 1935.

Leit = Leitner, G.W. 1889.

LYs = Lorimer, David L.R. 1962.

M-A = Mallory, J.P. and Adams, D.Q. (eds.). 1997.

M-A2 = Mallory, J.P. and Adams, D.Q. 2006.

RMJ = Recnikna makedonskiot literaturen jazik. (I-III) 1961-1966.

RYTEDD = Ranjung Yeshe Tibetan-English Dharma Dictionary (online).

T = Turner, Ralph L. 1966.

T-M = Tiffou, Etienne and Morin, Yves Charles. 1989.

T-P = Tiffou, Etienne and Pesot, Jurgen. 1989.

Wat = Watkins, Calvert. 2000.

Will = Willson, Stephen R. 1999.

References

Adams, D. Q. 1985. Designations of the Cervidae in Proto-IndoEuropean. Journal of Indo-European Studies 13: 269-282.

Alinei, Mario. 2003. Interdisciplinary and Linguistic Evidence for Palaeolithic Continuity of Indo-European, Uralic and Altaic Populations in Eurasia, with an Excursus on Slavic Ethnogenesis. Quaderni di semantica 26: 1-57. http://www.continuitas.com/interdisciplinary.pdf, accessed February 16, 2009.

Alonso de la Fuente, Jose Andres. 2006. El burusaski, ^una lengua aislada? Revista Espanola de Lingiiistica Vol. 35, no. 2: 551-579.

Bashir, Elena. 2000. A Thematic Survey of Burushaski Research. History of Language 6/1: 1-15.

Bednarczuk, L. 1986. Jezyki indoeuropejskie. Warszawa: Paiistwowe Wydawnictwo Naukowe.

Beekes, Robert S.P. 1995. Comparative Indo-European Linguistics (An Introduction). Amsterdam/Philadelphia: John Benjamins.

Berger, Hermann. 1974. Das Yasin Burushaski (Werchikwar). Wiesbaden: Otto Harrassowitz.

--. 1990. Burushaski. In Encyclopaedia Iranica. Ehsan Yarshater, ed. London and New York: Routledge and Kegan Paul. Vol IV:567-568.

--. 1998. Die Burushaski-Sprache von Hunza und Nager. Teil I. Grammatik. Teil II. Texte. Teil III. Worterbuch. Burushaski Deutsch; Deutsch-Burushaski. Neuindische Studien 13. Wiesbaden: Harrassowitz Verlag.

--. 2008. Beitrdge zur historischen Laut- und Formenlehre des Burushaski. Neuindische Studien 15. Wiesbaden: Harrassowitz Verlag.

Bernstejn, S. B., 1988. Obscekarpatskij dialektologiceskij atlas. Prilozi na Oddelenieto za lingvisticka i literaturna nauka, MANU. 133-141.

Biddulph, J. 1986 [1880]. Tribes of the Hindoo Koosh. Lahore: Ali Kamran Publishers. (Orig. published: Calcutta: Government Printing Office).

Bomhard, A., Kerns, I. C. 1994. The Nostratic Macrofamily: A Study in Distant Linguistic Relationship. New York: Mouton de Gruyter.

Brancus, Gr. 1983. Vocabularul autohton al limbii Romdne. Bucharest: Editura Stiintifica si Enciclopedica.

Buck, C. D. 1949. A Dictionary of Selected Synonyms in the Principal Indo-European Languages. Chicago-London: University of Chicago Press.

Cabej, E. 1976. Studime etimologjike n'efush'e te shqipes. Studime gjuhesore. I. Prishtine.

Casule, Ilija. 1998a. Basic Burushaski Etymologies. (The IndoEuropean and Paleobalkanic Affinities of Burushaski). MunichNewcastle: Lincom Europa.

--. 1998b. Za etimologijata na makedonskiot glagol vrne. Prilozi na Oddelenieto za lingvisticka i literaturna nauka. XXIII, 1-2: 121-136.

--. 2001. Osnovni burusaski etimologii. Skopje: TRI.

--. 2003a. Burushaski Names of Body Parts of Indo-European Origin. Central Asiatic Journal 47, no. 1: 15-74.

--. 2003b. Evidence for the Indo-European Laryngeals in Burushaski and Its Genetic Affiliation with Indo-European. The Journal of Indo-European Studies 31, nos. 1-2: 21-86.

--. 2004. Burushaski-Phrygian Lexical Correspondences in Ritual, Myth, Burial and Onomastics. Central Asiatic Journal 48, no. 1: 50-104.

--. 2009. (in press) Burushaski Numerals of Indo-European Origin. Central Asiatic Journal. 53, no. 2. (22 pp).

Cooper, Gregory R., Trail, Ronald L. 1999. Kalasha Dictionary - with English and Urdu. Islamabad: National Institute of Pakistani Studies, Quaid-i-Azam University/Summer Institute of Linguistics.

Cunningham, A. 1854. Ladak, Physical, Statistical and Historical: With Notes on the Surrounding Country. London [pp. 398-418 a short word list].

Danka, I. R. 1986. Jtjzyki anatolijskie. In Bednarczuk: 275-339.

Desnickaja, A.V. 1984. Sravnitel'noe jazykoznanie i istorija jazykov. Leningrad: Nauka.

Diakonoff, I. M. and V. P. Neroznak. 1985. Phrygian. New York: Caravan Books, Delmar.

Dimitrovski, Todor et al. 1987. Recnik na makedonskata narodna poezija. Vol. 2., Skopje: Institut za makedonski jazik "Krste Misirkov".

Dumezil, G. 1963. Caucasique du Nord-Ouest et purlers scythiques. Annali. Sezione linguistica. Instituto orientale di Napoli. V. Roma.

Du Nay, Andre. 1977. The Early History of the Roumanian Language. Edward Sapir Monograph Series in Language, Culture and Cognition 3. Lake Bluff: Jupiter Press.

Edel'man, D. I. 1997. Burusaski jazyk. In: Volodin, A.P. et al. Paleoaziatskie jazyki. Series Jazyki mira : 204-220. Rosiijskaja akademija nauk. Moscow: Indrik.

Edel'man, D. I., Klimov, G. A. 1970. Jazyk burusaski. Moscow: Nauka.

Ernout, A., Meillet, A. 1959. Dictionnaire etymologique de la langue latine. Paris: Librairie Klincksieck.

Fortson IV, Benjamin W. 2004. Indo-European Language and Culture. Oxford: Blackwell Publishing.

Gamkrelidze, T. V., Ivanov, V. V. 1984. Indoevropejskij jazyk i indoevropejci. Tbilisi: Tbilisi State University.

Georgiev, V. 1981. Introduction to the History of the Indo-European Languages. Sofia: Bulgarian Academy of Sciences.

Gluhak, Alemko. 1993. Hrvatski etimoloski rjecnik. Zagreb: A. Cesarec.

Gottlieb, Eugene. 1931. A Systematic Tabulation of Indo-European Animal Names: With Special Reference to Their Etymology and Semasiology. Language. Vol VII, no. 3: 5-48.

Haas, Otto. 1966. Die Phrygischen Sprachdenkmdler. Sofia: Bulgarian Academy of Sciences.

Hay ward, G. W. 1871. Hunza and Nager, and Yassin. Vocabularies. Journal of the Royal Asiatic Society. Vol. XLI, p. 18ff.

Illic-Svityc, V.M. 1967. Materialy k sravnitel'nomu slovarju nostraticeskix jazykov (indoevropejskij, altajskij, ural'skij, dravidijskij, kartvel'skij, semito-xamitskij). Etimologija. [1965]. Moskva.

--. 1976. Opyt sravnenija nostraticeskix jazykov. Vol. 2. Moscow. Illyes, Elemer. 1988. Ethnic Continuity in the Carpatho-Danubian Area. Eastern European Monographs. CCX1IX. New York: Columbia University Press.

Katicic, Radoslav. 1976. Ancient Languages of the Balkans. The Hague: Mouton.

Leitner, G. W. 1889. The Hunza and Nagyr Handbook. Pt. 1. Calcutta. (Grammar and dictionary of the Nagar dialect).

Lorimer. D. L. R. 1935. The Burushaski Language. Vol. 1. Introduction and Grammar with a preface by G. Morgenstierne. Oslo: Instituttet for Sammenlignende Kulturforskning. H. Aschehoug & Co.

-1 1938. The Burushaski Language. Vol. 3. Vocabularies and Index. Oslo: Instituttet for Sammenlignende Kulturforskning. H. Aschehoug & Co.

--. 1962. Werchikwar-English Vocabulary. Oslo: Norwegian Universities Press.

Mallory, J. P. and D. Q. Adams, eds. 1997. Encyclopedia of IndoEuropean Culture. London--Chicago: Fitzroy Dearborn Publishers.

Mallory, J. P. and D. Q. Adams. 2006. The Oxford Introduction to Proto-Indo-European and the Proto-Indo-European World. Oxford: University Press.

Morgenstierne, Georg. 1935. Preface to Lorimer (1935 I: vii-xxx).

--.1945. Notes on Burushaski Phonology. Norsk Tidsskrift for Sprogvidenskap 13: 61-95. Neroznak, Vladimir. 1978. Paleobalkanskie jazyki. Moscow: Nauka. Orel, Vladimir. 1997. The Language of Phrygians. New York: Caravan Books, Delmar.

--. 1998. Albanian Etymological Dictionary. Leiden: Brill. Peev, Kosta. 1988. Kukuskiot govor. Skopje: Studentski zbor. Poghirc, Cicerone. 1967. Considerations sur les elements autochtones de la langue Roumaine. Revue Roumaine de Linguistique. XII: 1936.

--. 1969. Istoria Limbii Romane. Vol. 2. Bucharest. Pokorny, Julius. 1959. Indogermanisches etymologisches Worterbuch. Bern-Munich.

Ranjung Yeshe Tibetan-English Dharma Dictionary (online). (www.nitharta.org). Nitharta International: Ranjung Yeshe Publications.

Rasmussen, J. E. 1999. Selected Papers on Indo-European Linguistics. Miscellaneous problems in Indo-European languages VII. Copenhagen: Museum Tusculanum Press (esp. 644-656).

Recnik na makedonskiot literaturen jazik. Vols. I-III. 1961-1966. Skopje: Institut za makedonski jazik.

Russu, Ion I. 1967. Limba traco-dacilor. 2nd edition. Bucharest.

--1970. Elemente autohtone in limba romana. Substratul comun romano- albanez. Bucharest.

Skok, Petar. 1974. Etimologijski rjecnik hrvatskoga Hi srpskoga jezika. Vols. I-IV. Zagreb: Jugoslavenska akademija znanosti i umjetnosti.

Starostin, Sergej A. 1988. Indoevropejsko-severnokavkazskie izoglossy. in Drevnij Vostok: Etnokul'turnye svjazi. Ed. Bongard-Levin, G.M. and V. G. Ardzinba. 112-163. Moscow: Nauka.

Steingass, Francis Joseph. 1999 [1892]. A Comprehensive PersianEnglish Dictionary Lahore: Sang-e-Meel Publications. [Orig. published: London: Routledge & Kegan Paul.]

Szemerenyi, O. J. L. 1996. Introduction to Indo-European Linguistics. London: Clarendon Press.

Tiffou, E. 2000. Current Research in Burushaski: A Survey. History of Language [special issue on Burushaski]. Vol. 6, No. L: 15-20.

Tiffou, E. and Y. C. Morin. 1989. Dictionnaire complementaire du Bourouchaski du Yasin. AMI 17. SELAF 304. Paris: Peeters/Selaf.

Tiffou, E, Pesot, J. 1989. Contes du Yasin. (Introduction au bourouchaski du Yasin avec grammaire et dictionnaire analytique). Paris: Peeters/Selaf.

Tomaschek, Wilhelm. 1980. Die Alten Thraker (Eine ethnologische Untersuchung). Vienna: Austrian Academy of Sciences. (Orig. published 1893-1894.)

Trubacev, Oleg. 1974-. Etimologiceskij slovar' slavjanskix jazykov. Moskva: Nauka.

Tucker, T. G. 1931. A Concise Etymological Dictionary of Latin. Halle (Saale): Max Niemeyer Verlag.

Turner, Ralph L. 1966. A Comparative Dictionary of the Indo-Aryan Languages. London: Oxford University Press.

Van Driem, George. 2001. Languages of the Himalayas. Vols. 1-2. Leiden: Brill.

Varma, S. 1941. Studies in Burushaski Dialectology. Journal of the Royal Asiatic Society of Bengal. (Letters) 7: 133-73.

[Vasmer, Max] Fasmer, Maks. 1986-1987. Etimologiceskij slovar' russkogo jazyka. Vols. I-IV. Moskva: Progress.

Walde, Alois and Julius Pokorny. 1927-32. Vergleichendes Worterbuch der indo-germanischen Sprachen. 3 vols. Berlin: Mouton de Gruyter.

Watkins, Calvert. 2000. The American Heritage Dictionary of IndoEuropean Roots (2nd edition). Boston: Houghton Mifflin.

Willson, Stephen, R. 1999. Basic Burushaski Vocabulary. Studies in Languages of Northern Pakistan, vol. 6. Islamabad: National Institute of Pakistan Studies/Summer Institute of Linguistics.

Zarubin, I. 1927. Versikskoe narecie kandzutskogo jazyka. Kollegija vostokovedov. Zapiski 2: 275-364.

Zoller, Claus Peter. 2005. A Grammar and Dictionary of Indus Kohistani. Trends in Linguistics. Documentation. 21-1. Berlin: Mouton de Gruyter.

(1) There have been many unsuccessful attempts at correlating it with Basque, Nubian, the Dravidian languages, Munda, various Caucasic languages, Yenisseian (Van Driem 2001), Sino-Tibetan, Sumerian (see Bashir 2000: 1-3). For a recent rejection of the attempts to relate Burushaski to the Dene-Caucasian hypothesis, see Alonso de la Fuente (2006: 562-6).

(2) The grammatical correspondences are outlined in Casule (2003b: 69-80). We analyse the correlation of the Burushaski numeral system with Indo-European in Casule (2009).

(3) For a detailed appraisal of this evidence, see Alonso de la Fuente (2006).

(4) In Casule (2003b), we provide an in-depth analysis of the Burushaski laryngeal and their direct correspondence with the Indo-European laryngeals. We follow the approach by Adams and the editorial board of Mallory-Adams (1997: 462) where four PIE laryngeals are assumed: [h.sub.1], [h.sub.2], [h.sub.3] and [h.sub.4]. In this set, [h.sub.2] and [h.sub.4] colour an adjacent *e to *a and [h.sub.3] colours an adjacent *e to *o, and are considered to have been pharyngeal and/or laryngeal continuants (fricatives). The first laryngeal h, does not cause colouring and has been assumed to be a glottal stop. Adams uses the symbol ha when, because there are no Hittite and Albanian forms, it is impossible to determine whether the laryngeal is [h.sub.2] or [h.sub.4]. A generic [h.sub.x] is used when there is evidence for a laryngeal, but its exact nature cannot be determined. Most Indo-Europeanists, if not all, accept the existence of at least one laryngeal confirmed by its attestation in Hittite and other evidence, but the three-laryngeal theory also enjoys wide acceptance, whereas the fourth laryngeal has been more often disputed than not.

(5) I am grateful to E. Bashir and CP. Zoller for their assistance with the Indo-Aryan material, and to J.A. Alonso de la Fuente and E. Tiffou for their comments, but the responsibility for all shortcomings is mine.

(6) Morgenstierne (1935: XXI1-XX1I1) indicates that "Burushaski retains here the $ which would have been lost in a modern Shina word, and also a very ancient meaning of the word."

(7) Some reconstruct the PIE form as *h3eiii- (e.g. Danka 1986: 314) > *hoi|i- -both interpretations are possible for Burushaski, where subsequently ou > u.

(8) Starostin (1988: 114), in a tentative list of lexical correspondences between Proto-North Caucasian and PIE, correlates both PIE stems [*haeigs- and *haegos] with Proto-Adyghian-Kabardian *aca 'he-goat', Adygh aia, Kab aza 'same', also with East Caucasian and notes Dumezil's (1963: 13) earlier correlation of the Adyghian word with IE. Starostin (1988: 154) is inclined to see here borrowings from Proto-North Caucasian into IE. Burushaski seems to fit well in this analysis--having the stem of PAK. *aca + the IE suffix (?) -s, although the Burushaski word could have developed independently and directly from PIE.

(9) A hyphen preceding a word indicates that it is used only with pronominal prefixes.

(10) The basic Burushaski word for 'lamb' is mamusi, derived by Berger < mamd 'milk' + se- 'to eat' (L 253) (B 277), clearly an independent innovation.

(11) Berger (ibid) indicates that the forms without -k- should be seen as secondary.

(12) For the change rs > s see e.g. the derivation of gas 'price' (B 150) or bas 'bridge' (B 43).

(13) There is a curious parallel between the Burushaski reduplicated form du do provided by NH and lllyr deda 'foster mother' which Pokorny (IEW 235) indicates may be from IE *dhe(i)- 'to suckle' rather than from a child word for 'grandparents'. There is a further intriguing correspondence between Bur du 'kid' and Med (dial.) dujak 'kid' (-jak is a suffix) (Peev 1988: 41) (the informant explains that it is called dujak from duj- 'to suckle, breastfeed') which is likely to be a coincidence, yet note also Med dude 'endearing term (usually) for a small child' (Dimitrovski 187) which could be a diminutive/hypocoristic in -e from a basic form *dudo (as e.g., tate (dem. hyp.): tato (basic) 'father').

(14) Compare with the common genitive syntagms like e.g. car 'sentry, watchman' : care sikiari 'watch tower' (B 69).

(15) For example: IE *uer- 'high raised spot', zero-grade form *ur- (Wat 99) > Bur -riri and -dris 'crest, ridge, peak; prong; fingernail', and Bur buuri 'crest of hill, peak' (B 66).

IE *uel-7 'turn, wind; round' (IEW 1441--1444): (from the zero-grade) Bur du-iil- 'wind, become wound up' (B 454), also: Bur bal-, -wal- 'fall; find oself in, land in, be beaten; settle down' (B 32), (cp. with OS1 valjati se 'roll, fall clumsily'). And further: Bur bal 'wall' (B 31), (cp. with Lat vallum 'palisade, wall') (IEW 1140) and Bur du-wal- 'fly, fly away', d'~wal- 'winnow' (B 463) balds 'bird' (B 33), cp. with OS1 vlajati se 'to be cast up'.

IE *uel-g 'to tear, pull' (e.g. in PS1 *ob-velkti 'put on clothes') (M-A 567) > Bur bei- : Ys wel- 'put sth on, don, wear' (B 47). (Refer to |27] Iduyo 'tuft of wool, etc.'.)

IE *uer-g 'perceive, watch out for' (Lith vert 'see, look into', OHG gi-war'careful', OEng warian 'beware', Gk horan 'see' (IEW 1164) > Bur baren-, bare- 'look, look at; look for, search for; look after, look about' (B 40), also d' waran- 'require, be in need of (B 465).

IE *uod-or- 'water' or *ud-5r- (suffixed zero-grade form) (Wat 95) (or *ud-ro ?): Bur buddo 'rinsing water', Hz also 'water which becomes warm in the sun'(B 61).

IE *uos 'you plural' (Wat 102) or *uohl- 'you two' > the Yasin doublet for the pronoun for 2 p.pl. wa : ma (T-P 105, 151), Hz Ng ma (through an intermediate *ba ? ) (no word in Bur begins with wo-, i.e., all Bur words in w(in the anlaut) have wa- in Hz Ng (B 462-466) and wa- and we- in Ys). In total we have identified 25 such correspondences.

(16) While Lorimer's phonological analysis was not entirely adequate, we must take into account his perception of the vowel sounds and phonetics. Being able to draw upon the rich system of vowels in English would have helped him describe more precisely the vowels of Burushaski.

(17) Note in this example once more the alternation we : wa.

(18) For various examples of haplology in Burushaski, see Berger 2008: 6.13.

(19) Note e.g. Bur bur/6 no '(of people) corpulent' (B 64), which we have derived from IE *bhris-, *bhers- 'fast' (sem. in IE > 'sprightly, virilis; defiant, forward; big, a lot' (IEW 143).

(20) Note also Ind Koh kAra 'a ram whose horns are turned inwards' (Zoller 110) which seems to be a loanword from Shina or Burushaski. Zoller (16) estimates that about 10% of the Indus Kohistani vocabulary consists of borrowings from Burushaski.

(21) The basic Burushaski word for 'horn' is 'tur Hz Ng Ys in Hz Ng also '-ltdr (L 252, 360) (B 270), which in turn we (Casule 1998a: 47) have very tentatively related to IE *tauro- 'bull' (Wat 89) < 'the strong, sturdy, big, [solid] (animal)' (Gottlieb 22), or rather: 'homed' (?).

(22) Other historical examples: Bur bada 'sole of foot; step, pace' (B 29) : OInd pada 'footstep, track, pace', in Ys baya 'same' < (?) Pkt paya- 'footstep, foot' < IE *ped-, *pod- (nom. root) 'foot' and from IE *ped-, *pod- (verbal root) 'fall, stumble': Ys badari -wal- 'fall down on one's back, fall over, faint' (BYs 131) (Wat 62). Even in such examples we find the alternation p:b as Burushaski also has pada ay -t- 'kick a stone with the foot' (which Berger links with a ? with bada above) (B 310) (See also |17J.) And further examples of a historical change:

IE *peh,(i)- 'misfortune', *pih,- 'revile' (e.g. OInd paman- 'skin disease', plyati 'insults' (M-A 313). In Burushaski we have Ys bihai, Hz Ng biai 'illness, disease' (B 50), where the retention of the laryngeal clearly points to an original Burushaski word.

IE *perR- 'to dig out' (Wat 66) > Bur biraq- 'to dig, to dig anywhere' (B 54).

IE *pel(i)s- 'rock; cliff (Wat 64) > Bur baloos 'a kind of stone' (Will 20).

(23) Starostin (1988:113) gives as an isogloss with the IE words, Proto-East Caucasian *wVtVrV 'child (up to one year old)': Tsakurian vudra 'kid up to one year old', Tsez beduro 'cub', Batzbean bader, Chechen ber 'child' and possibly within a Nostratic etymology (Illifi-Svityc 1967: 337).

(24) Bur busooso, Ys bos6 'calf, can be traced to IE *nrs- (zero-grade of *uers'dampen, wet; beget') esp. to Lith vefsis 'calf, Lett versis 'ox' and Lat verres 'boar', Av varasna- 'male of animal', OInd vrsa-, vtsan- 'male of animal' (M-A2 204) (Gottlieb 17-18).

(25) From the same IE stem, i.e. from *pau-kos 'little, few; small' (M-A 200), Bur has phuk 'a small speck of any substance', phriko adj. 'small, tiny' (B 334).

(26) There is another possible Paleobalkanic parallel with the Thracian word ebros 'buck' < IE *[h.sub.1]eperos 'boar' (M-A2 142) (Gottlieb 10-11), e.g. OS1 veprl, Lett vepris 'boar' (with an unexplained V-, derived by some etymologists (G 666) < *wep- 'throw, sprinkle (semen)' (<*[h.sub.1]wep-?), Lat aper, OEng eofor, OHG ebur, epur, all 'boar'. A correlation is possible with Bur beepay, also bepay 'yak'(Tib gyeg), Sh bepo (B 48) (retroflex y possibly < *r-io). It is very important that Bur also has bipher n. 'breed, engendering' (B 53-4) which fits directly with the semantics of 'begetting', and provides evidence for a final-r-in Bur beepay.

(27) Yet interestingly, Lorimer (L 438) notes Traqhandatiri (B 438) "a dynasty whose headquarters were at Gilgit. The name is derived from Trakhan (...)" who was its ruler. It is very tempting to seek a historical correlation with the Thracian ethnonym, i.e. with Thrikes or Thrakion ethnos (KatiCic 131-132) although it is more likely that this is a more recent regional development.

(28) Consider in regard to Thracian consonantism the discussion and examples in Katicic (1976: 143): "The only safe conclusion is that Greek and Latin writers met with difficulties when rendering Thracian stops in their familiar alphabet. This has been explained (...) as a full-fledged consonant shift, (...) as the orthographic representation of stops somewhat different from the Greek one and similar to the articulation of stops in Albanian".

(29) "A most characteristic Thracian feature is the alternation of b and m." (Katicic 1976: 144).

(30) There are also a few words of non-Slavic origin in the highly Balkanised Macedonian language, at the centre of the Balkan Sprachbund, possibly substratal, that find a direct parallel with Burushaski, e.g. Med bara 'look for, search for; seek; demand; look after' (RMJ I: 22) : Bur baren-, bare- 'look, look at; look for, search for; look after, look about' (B 40), possibly < IE *wer-4 'perceive, watch out for' (Wat 99) or more tentatively Med vrne 'to rain', a unique Macedonian verb most probably of substratal provenience [for the etymology see Casule 1998b] : Bur war man- 'rain heavily (B 464) < IE *we-r- 'water, liquid' or IE *wers- 'to rain, drip' (Wat 100) etc. (For other examples, see Casule 2001: 125-8.)

(31) In the wider semantic field of 'domestic animals' note the direct correspondence between the Phrygian gloss ma 'Phrygian call to cattle' (Neroznak 1978: 150), interpreted as an interjection (prominent and specific enough to be recorded by the ancient lexicographers) and the Burushaski interjection mi ha 'come! - a call to a horse' (B 275). Interjections of this type are generally culture-specific, even if they have an expressive component (Bashirp.c).

Note also in reference to 'horse', the discussion (Casule 2004: 78-79) of the link between the Bur Ys word kabrit 'white horse' (T-M 33), and Bur kabulek 'roof-posts' and Phrygian kubela, Thracian *kabula 'horse', OS1 kobyla 'mare' > kobyluka 'pole for carrying loads', kobylica 'scaffolding, supporting beams etc.', Lat caballus 'horse, gelding, work horse', words which Buck (168) considers of Anatolian or Balkan ethnic origin.

Ilija Casule

Macquarie University
Table 2. Summary of Burushaski-Indo-European phonological
correspondences.

IE a                            >    Bur a
IE e                            >    Bur Ys e : Hz Ng i
IE e (unstr.)                   >    Bur a
IE e                            >    Bur ee, i
IE o                            >    Bur o
IE o (unstr.)                   >    Bur a
IE o                            >    Bur oo, oo
IE i                            >    Bur i
IE u                            >    Bur u

IE ai, ei, oi; eu               >    Bur a
IE au, ou                       >    Bur a

PIE [h.sub.1-]                  >    Bur [h-.sup.4]
PIE [h.sub.1]e-                 >    Bur he-
PIE [h.sub.1]uer-               >    Bur har-:-war-: her-
PIE [h.sub.2-]                  >    Bur h-
PIE [h.sub.2]e- > [h.sub.2]a-   >    Bur ha-
PIE [h.sub.2]ou-                >    Bur hu-
PIE [h.sub.2]ue-                >    Bur -we- : -wa-
PIE [h.sub.a]-                  >    Bur h-
PIE [h.sub.a]e- > [h.sub.a]a-   >    Bur ha-
PIE [h.sub.4]-                  >    Bur h-
PIE [h.sub.4]e- > [h.sub.4]a-   >    Bur ha-
PIE [h.sub.3]-                  >    Bur h-
PIE [h.sub.3]e- > [h.sub.3]o-   >    Bur ho-
PIE [h.sub.x] -                 >    Bur h-
PIE [h.sub.1]/[.sub.2]i-        :    Bur i-
PIE -[eh.sub.1]-                >    Ys -e- : Hz Ng -ee-: -ii-
PIE [h.sub.2][eh.sub.x]-        >    Bur -aa-
PIE [h.sub.3]e[h.sub.x]-        >    Bur -oo-
PIE -er/[lh.sub.1/2] -          >    Hz Ng -ii-r/1: -aa-r/1
                                     [Ys:-aha-r/1]
                                       IE 1, m, n, r > Bur 1, m, n, r
                                       IE u          >    Bur -w/-u
                                       IE u-         >    Bur b-
                                       IE y          >    Bur y/i
                                       IE m          >    Bur -um, am
                                       IE n          >    Bur -un,-an
                                       IE r          >    Bur -ur, -ar
                                       IE l          >    Bur -ul, -al
IE p-                           >    Bur ph-, also b-
IE -p-, -p                      >    Bur -p-, -p
IE bh                           >    Bur b
IE b                            >    Bur b, also m
IE t                            >    Bur t: th : t , also d-
IE dh-                          >    Bur d-
IE VdhV, -dh                    >    Bur -t-,-t-,-t
IE d                            >    Bur d
IE k                            >    Bur k : kh, also g-: g-
Ie [k.sup.w]                    >    Bur k
lE k                            >    Bur k, kh, also g-: g-
IE g, gh                        >    Bur g, g
Ie [g.sup.w]                    >    Bur g, g
IE g , gh-                      >    Bur g, g
IE s                            >    Bur s ors : c , ch
IE ks                           >    Burs
Gale Copyright: Copyright 2009 Gale, Cengage Learning. All rights reserved.