How the mole and mongoose got their names: Sanskrit akhu- and nakula-.

• Author: Katz, Joshua T.

THE ETYMOLOGY OF THE SANSKRIT WORD FOR (supposedly) 'mole', akhu- (RV +), is unclear. There are two proposals in the scholarly literature: according to the more popular view, it is a derivative of the Sanskrit root ([LANGUAGE NOT REPRODUCIBLE IN ASCII]) 'dig', with the preverb a; alternatively, it is to be connected with words for 'mole' in Anatolian and Greek. Both are problematic, the former on morphological, the latter on phonological grounds. On the one hand, it is by no means obvious by what derivational process a ustem, like akhu-, could be connected with a root like ([LANGUAGE NOT REPRODUCIBLE IN ASCII]), despite the plausible semantics. Although Manfred Mayrhofer distances himself from this fairly standard etymology in his dictionary (Mayrhofer 1989: 446), he seems at the same time not to rule it out; (1) however, his comment that the "Verbindung mit [KHAN.sup.I] wohl eine Parallelwurzel * kha voraussetzt [~ kha- ?]" (cf. ([LANGUAGE NOT REPRODUCIBLE IN ASCII])]- 'gain, obtain' [LANGUAGE NOT REPRODUCIBLE IN ASCII]; Mayrhofer 1996: 696f.) skirts the issue, (2) all the more so as the very existence of this parallel root, which Mayrhofer writes with an asterisk, remains in doubt. (3) On the other hand, even though Jaan Puhvel has cautiously, but in my view probably correctly, interpreted the Hittite animal asku- as meaning 'mole' (connecting it with the etymologically disputed Greek words for this creature, aphaeretic [LANGUAGE NOT REPRODUCIBLE IN ASCII]and ~ metathetic [LANGUAGE NOT REPRODUCIBLE IN ASCII] (4) forms to which I return briefly at the end of this paper) (5) and noted that the "u-stem asku- is also reminiscent of Skt. akhu- 'mole', the explanation of which via a + kha-/khan- 'dig' leaves a lot to be desired" (thus, Puhvel 1981b: 242 [similarly, 1984: 216]), it remains the case that the Sanskrit word does not have a sibilant (it is not *askhu- [vel sim.]), so the connection would seem to be far from impeccable from the point of view of phonology. (6)

Since both loan words and animal names are notoriously subject to linguistic "deformation" (consider just the case of the Wanderwort for 'mole' in Greek), Puhvel's tentative connection between Hitt. asku- and Skt. akhu- should not be dismissed out of hand. But it can be proved, or at least considered the preferable explanation of the Sanskrit word, only if there is independent support for the phonology. For example, is there any other plausible equation between a Hittite word in (-)Vsku-and a Sanskrit word in (-)VKu-? It is my contention that this question can, rather surprisingly, be answered in the affirmative.

In Katz 1998, I discussed the only word for 'badger' that shows up in more than one branch of Indo-European, (7) adding Hitt. tasku- to the Germano-Celtic quasi-equation between Mod. Germ. Dachs and other Germanic words for this animal (< PGmc. *pahsu- < pre-PGmc. *taksu-) and such Celtic forms as the Gaulish onomastic element Tasco(-), Tasgo(-) and the Irish personal name Tad(h)g (< (pre-)PCelt. *tasko-/*tazgo-), as in the hero Tadhg mac Cein, whom Mac an Bhaird 1980 has shown to be specifically associated with badgers. (8) Now, the Hittite word is attested only with the meaning of a subcaudal body part (something like 'anus' or 'scrotum'), but I endeavored to explain how this meaning developed out of 'badger', (9) suggesting also that the original sense is preserved in a number of Anatolian personal names (most notably Hitt. mTaskuili- and HLuv. Ta-sa-ku-wa-li/Ta-sa-ku-li /Tasku(wa)lis/ [~ Hellenized [LANGUAGE NOT REPRODUCIBLE IN ASCII] (10)) and probably also the place name [LANGUAGE NOT REPRODUCIBLE IN AS CII] (best-known from Bithynia, but apparently used for a number of locations in western, especially northwestern, Asia Minor), on which see now Neumann 1999a: 17f. (11) The sheer number of discrepancies among the forms strongly supports the idea that we have to do with a Wanderwort: Celtic seems to require both *tasko-, with voiceless *-sk-, and *tazgo-, with a voiced cluster; (12) the Celtic and Hittite forms both show -SK- rather than the -KS- of Germanic; and the Germanic and Hittite forms are both u-stems, whereas the Celtic is a garden-variety 0-stem. A rough approximation of what this word for 'badger' would have looked like in prehistoric times--at least in Near Eastern languages of the late third or second millennium B.C.--is thus *tasku-, a u-stem with inherent a-vocalism.

There is evidence that in ancient Anatolia, badgers and moles were linked, and in my discussion of Hitt. tasku-, I commented briefly also on asku- (Katz 1998: 70 n. 19 and esp. 73 n. 31), making two intertwined points. First, one ancillary argument in favor of understanding tasku- as meaning, or having once meant, 'badger' is the existence of a morpho-phonologically strikingly similar word for another burrowing animal with a pointed head or snout. (13) Put another way, the similar appearance of the creatures in question, coupled with the likeness in linguistic structure, means that one would not be surprised to learn that tasku- (< *tasku-) and asku- (< *asku-), both animal- Wanderworter, acted as a pair and, perhaps, that they were borrowed together into (or, for that matter, even from) Hittite.

Second and arguably more interesting, there is the curious fact, hitherto unexplained, that the name of (apparently) one and the same Galatian Christian sect--that is to say, a group of Celts in Anatolia--is attested as both [LANGUAGE NOT REPRODUCIBLE IN ASCII] and Ascodrogitae (among other variants). (14) Our best source for these people is an account, in Greek, by the fourth-century A.D. bishop Epiphanius of Salamis, who writes in his Panarion of 374/75 that the [LANGUAGE NOT REPRODUCIBLE IN ASCII] are literally 'Peg-noses' (Gk. [LANGUAGE NOT REPRODUCIBLE IN ASCII]), so-called because [LANGUAGE NOT REPRODUCIBLE IN ASCII] means 'peg' (Gk. [LANGUAGE NOT REPRODUCIBLE IN ASCII]) and [LANGUAGE NOT REPRODUCIBLE IN ASCII] means 'nose' (Gk. [LANGUAGE NOT REPRODUCIBLE IN ASCII] 'snout, muzzle'), (15) and that when they pray, they put their index finger on their nose: [LANGUAGE NOT REPRODUCIBLE IN ASCII] (48.14.4 = p. [II.]239 Holl--Dummer). (16) Modern scholars have followed Epiphanius in interpreting [LANGUAGE NOT REPRODUCIBLE IN ASCII] as 'Peg-noses', but the existence of various by-forms without an initial dental stop (e.g., Ascodrobi and Ascodrugitae; see n. 14) makes it in fact more probable that these people are instead 'Badger-noses'. Rather than assume a sporadic and, as far as we know, phonologically completely unmotivated loss of the initial consonant of the first element of the compound (whether it means 'badger' or 'peg'), it is preferable to think that the 'Badger-noses' might sometimes be referred to instead as the 'Molenoses', (17) with the simple substitution of the word for one snouty animal indigenous to Anatolia with a nearly identical word for another such creature. (18)

With due caution, I suggest that two other languages attest to the same sort of confusion: Basque and Modern Greek. (19) In Katz 1998: 70 n. 19 and esp. 71, with nn. 24f., I mentioned the Basque word for 'badger', azkoin (and many variants), which clearly reflects a pre-Basque preform *azkone, and quoted the standard etymology, namely that it goes back to something like *(t)axonem (cf., e.g., Span. tejon), that is, to the oblique form of the Late Latin word for 'badger', taxo (itself a borrowing from Celtic or Germanic). (20) There are two difficulties: a comparatively minor matter is that * azkone shows * -zk-rather than *-ks- (as in taxo); while this could perhaps be nothing more than a low-level, but seemingly unparalleled, alteration of the cluster /-ks-/ (which Basque does not tolerate at any period), it is also possible that a Celtic language (cf. Gaul. Tasco(-)) is somehow involved (compare Mac an Bhaird 1980: 153f.). Rather more dramatic is the apparent loss of the initial *t-. In my paper on badgers, I wrote that a parallel for this "would seem to be the form azkonaga, defined in [G.] Aulestia['s Basque-English Dictionary] as 'place of yew trees [cf. Lat taxus 'yew']. Used as a last name'" (Katz 1998: 71 n. 25). Larry "Trask tells me, however, that Aulestia's work is unreliable as far as etymology is concerned and that it is generally agreed that Azconaga (which in fact functions only as a name and has no synchronic meaning) is derived from the word for 'badger' (thus, e.g., Michelena 1973: 64). What this means, of course, is that the only parallel for the loss of *#t is not a parallel after all. But rather than despair, the absence of a potentially straightforward phonetic explanation gives us license to consider a completely different sort of solution to the problem: I propose that the Basque word for 'badger' begins with a vowel because it is intimately connected with forms elsewhere that mean 'mole' (e.g., Hitt. asku-)--forms that look like the respective words for 'badger', but crucially lack the init ial consonant. (21) Now, the ending *-one of *(t)azkone can hardly reflect anything but a Latin/Romance n-stem, like taxo (there is no evidence alongside talpa for a corresponding Latinate word *axo 'mole'); nevertheless, as noted above, the cluster *-zk- is not easy to reconcile with Lat. -x- (i.e., /-ks-/) in either the order of the consonants or the quality of their voicing. I therefore conclude that Basque azkoin 'badger' is most likely a cross between the Iberian Romance word for this animal, /taksone(m)/, and a manifestation, presumably unmediated by Latin, of the same vowel-initial Wanderwort for 'mole' that I have been discussing, namely something like /askV-/--in particular, /azgV-/. (22)

The existence of *azgu-, a u-stem with a voiced cluster, may find surprising confirmation in Modern Greek, where the word for 'badger' is [LANGUAGE NOT REPRODUCIBLE IN ASCII]. The only etymology known to me does not inspire confidence: according to Andriotis 1983...