Adhrigu and drigu: on the semantics of an old Indo-Iranian word.

• Author: Thompson, George

IN HIS LONG and characteristically thorough article, "Wortkundiche Beitrage zur Arischen (Indo-Iranischen) Kulturgeschichte und Welt-anschauung" (1958), W. Wust surveyed the many attempts that have been made to analyze and interpret the obscure Vedic word adhrigu. The great variety and mutual incompatability of these attempts demonstrates, at the very least, that the word has been perceived to be a problem ever since Yaska's analysis of it in the Nirukta (where it is rather cavalierly proposed that adhrigu = adhi, prep., + -gu, 'bull, cow'). In this brief article, offered in homage to the American Oriental Society's master Indo-Iranist, yet another attempt will be made to come to terms with this still opaque Vedic word. It is humbly offered as a drigu's gift to one whose judgments on all things Indo-Iranian have been unfailingly illuminating and instructive.

Leaving aside other highly unlikely analyses (such as Rajavade's [1932]: adhrigu < *adrigu- 'one that possesses cows shut in mountain strongholds'), prior to Wust's article three analyses have been generally preferred:

(1) adhrigu = a-dhri-gu (i.e., alpha-privative + verbal root dhr- + verbal root ga- 'to go'), thus = 'unhaltsam gehend' (Grassmann). (1)

(2) adhrigu = a-dhri-gu (i.e., alpha-privative + verbal root dhr + nominal stem gu- 'bull, cow'), thus = 'dont la vache ne retient pas (son lait), dont la vache est genereuse, qui a pour vache adhri' (Bergaigne). (2)

(3) adhrigu = a-dhrigu (i.e., alpha-privative + the Vedic equivalent of Avestan drigu 'poor'), thus = 'not poor, rich, liberal' (Bloomfield). (3)

A measure of the uncertainty confronting us with regard to this word can be found in the peregrinations on it in Mayrhofer's etymological dictionaries, covering a span of over thirty years. (4) In KEWA I.31 (s.v. adhrih) (5) Mayrhofer embraced the view that adhrigu is to be analyzed as in number 1 above (explicitly rejecting the proposed association with Avestan drigu in number 3, while completely ignoring number 2). In his later addendum (KEWA 1.548), Mayrhofer once again rejected the derivation of Vedic adhrigu from Avestan drigu (i.e., number 3), this time treating in some detail the views of K. Barr 1953 (on which, see below). In this treatment, Mayrhofer acknowledged an etymological connection between the two words, but he explicitly rejected the suggestion that the Avestan word was the primary member of the pair. The basis for this conviction was his view that there was no good reason to interpret the initial a- in adhrigu as an alpha-privative (this of course contradicts his earlier position at KEWA 1. 31). Instead, Mayrhofer insisted that there was "klar Zeugnis" for an element *adhri- not only in ddhrigu but also in the rare form adhrija (sic). Furthermore, he considered the Avestan form drigu to be a reanalysis of an inherited Iranian word *adri-gu, similar to the sort of back-formations found in Sanskrit, such as sita 'white', from asita 'black', and sura 'god', from asura 'lord', later 'demon'. Mayrhofer later returned to the problematic form adhrih in the addendum to KEWA (III.627), this time deriving from it the form adhrij (instead of adhrija) as well as adhrigu. Here Mayrhofer cited, very briefly and for the first time, the supporting evidence of Wust, as well as the important article of Bailey (1973). Finally, in EWA 1.67, in his final treatment of the problem, Mayrhofer re-asserted his view that the term ddhrigu contained a final member -gu 'Rind', as well as a first member adhri-, for the interpretation of which Mayrhofer expressed a clear preference for Wust's analysis.

The point of this extensive summary of Mayrhofer's views is that, while hesitating between different interpretations of the word, Mayrhofer has consistently assumed the primacy of an otherwise unattested form *adhri-, from which he has derived the attested forms adhrij and adhrigu. In doing so, Mayrhofer has relegated the Avestan form drigu to the status of an incorrect back-formation. This invites the question: is there a principled reason for Mayrhofer's preference?

Wust's contribution to the discussion was to come up with a novel etymology for this element adhri- in adhrigu. Rejecting both the proposed association of -dhri- with the verbal root dhr-, as well as the alternative derivation from Avestan drigu (and rejecting the presence of an alpha-privative in either case), Wust related ddhri- to the Greek forms [LANGUAGE NOT REPRODUCIBLE IN ASCII] 'eunuch', (6) and glossed it as "verschnittene (und somit im Effekt: domestiziert feist)" (p. 25); i.e., he interpreted adhrigu as a bahuvrihi compound with a first member adhri- 'castrated', and a final member -gu 'bull': adhrigu in his view thus = 'having castrated bulls, therefore wealthy, powerful'.

Admittedly, Wust's rather general remarks about the vital role of castrated and domestic animals in the Vedic economy are valid and even at times valuable. (7) But in the end this etymology is hardly convincing, for the simple reason that there is no evidence for it in the texts. Wust has closely examined each of the fifteen attestations of the term adhrigu in the RV, but he has not been able to point to any passage that even hints at an association of the word adhrigu with castrated animals. (8) In fact, as I will attempt to show, all of the attestations of the word in the RV are frustratingly opaque. Nevertheless, because the term is applied to several Vedic gods, notably Indra, Agni, the Maruts (Rudras), the Asvins, and Soma (though notably not to any one of the Adityas, Mitra or Varuna, etc.), as well as to a handful of otherwise obscure human figures, a generally positive sense like 'rich, liberal, powerful' etc., can be reasonably assumed. Beyond this rather generally positive sense, however, precious l ittle can be inferred from these collocations.

Since the publication of Wust's article, another important visitation to this problem was made by J. Narten in her 1986 edition, translation, and commentary on the Old Avestan text, the Yasna Haptanhaiti. In her discussion of the term drigudaiiah, 'nourishing the drigu', which is applied to the divine waters at YH 38.5, Narten briefly surveyed the attestations of the Avestan word drigu. (9) She presented a strong case for the view, held by most Iranists, that the Avestan word drigu and the Vedic word adhrigu are indeed related, but that the primary member of the pair is in fact the Iranian form drigu, 'poor, needy, dependent'. She explicitly rejected Wust's detailed analysis, including his suggestion that the Avestan word drigu should be interpreted as a Kontrarbildung to the supposedly inherited Indo-Iranian term *adhri-gu. Furthermore, like Barr (1953) and Bailey (1973) before her, Narten reversed the direction of the presumed re-analysis: instead of the derivation of the Avestan form from a falsely re-anal yzed Vedic one (drigu < Iranian *a-drigu < Vedic adhri-gu), as preferred by Wust and Mayrhofer, these Iranists all proposed a derivation of Vedic adhrigu as the negation of an unattested Vedic form *dhrigu, 'poor, needy' (cognate with Avestan drigu): therefore, in their view, Vedic adhrigu literally = 'not poor, therefore rich, powerful, etc.'. Again, one is tempted to ask: is there a principled reason for their preference?

The major problem with this derivation is the existence of strong forms like nom. p1. adhrigavah (at RV 1.64.3 and 8.22.11) instead of -avah, (10) as well as the collocation of the word adhrigu with dasagva at RV 8.12.2 (where both are proper names whose second elements appear to be alternative stems from the noun gdu 'bull, cow'). Such passages indicate that in fact there was a tendency already in the RV to analyze the Vedic word adhrigu as adhri-gu, with the final element understood as a reduced form of gau 'bull, cow'. The suggestion made by Narten and other Iranists that such forms are themselves simply...