On the Etymology of Vedic

• Author: Rothstein-Dowden, Zachary

1. The clitic (1) particle áha is attested throughout the Vedic period, from the Rgveda onward. (2) Like many other particles, it is nebulously defined in the lexica. Grassmann classifies áha as an emphatic particle with scope over the preceding word and analyzes it as consisting of the pronominal stem á-, found in the oblique forms of idám and locational adverbs like á-tra and á-tas, and the clitic particle ha < *ghe, the apophonic variant of gha < *gho. (3)

   The same etymological analysis is adopted by Delbrück, (4) who provides a detailed discussion of the use of the particle in prose syntax. (5) He likewise sees in áha an emphatic particle and traces its other uses back to this basic function. One particularly common use of the particle in Vedic prose is to introduce a subordinate clause that is contrasted with a following main clause:

   SB I 4,1,4 gayatrlm evaiiád arvacim ca páracim ca yunakti [[páracy áha devébhyo yajñám váhaty] arvaci manusyan avali]. (6) He thereby joins a gayatri verse directed hitherward to one directed away from here: [[the one which tends from hence áha carries the sacrifice to the gods,] and the one which tends hither-ward pleases the men.] (7) Delbrück conjectures that this use of áha developed out of its use as an emphatic particle. (8) I will return to this question in my analysis below.

   While this older etymology is not wholly compelling, it has yet to be superseded. Mayrhofen (9) without rejecting the earlier analysis, tentatively suggests that the word could be of exclamatory origin like German aha! It is unlikely, however, that a sound expressing surprise or approbation would have been transformed into a second-position clitic or would have acquired the requisite semantics. I know of no parallels for such a development.

   Oberlies (10) has suggested that áha could in fact be the lsg. form corresponding to pf. 3sg. aha 'says'. His reasoning, unfortunately, is not provided. Though positing verbal origin of the particle does offer significant advantages for explaining its syntactic properties as I will argue below, Oberlies's proposition is untenable on formal grounds. There is no reason to think that the initial long vowel of the expected lsg. *ãha < *HaHádhHa would have become short. The expected first-person form may even be attested in OAv. ada (on which see below). A morphological remodeling of this within Vedic by an analogy of the type 3sg. cakara: 1 sg. cakara :: 3sg. aha: .X, where x was resolved as 1 sg. áha would be poorly motivated; roots of the structure (H)aC regularly form a synchronically non-ablauting stem aC throughout the perfect in Vedic, a result of the fact that the strong stem HatiáC- and the weak stem HaHC- fell together. Shortening of the initial vowel of the particle in allegro speech would be an equally ad hoc explanation.

2. If all that we had to go on were limited attestations, semantics that are difficult to determine, unremarkable syntax, and no clear Indo-European cognates, the origin of the particle áha would be impossible to determine. The situation is, however, not quite so desperate. A first clue to the particle's origin comes from its odd accentual properties: Panini prescribes that a verb be accented in a sentence beginning with tu, pasya, pasyata, aha, and aho when they express respect." This accentuation rule appears to obtain in practice in Vedic prose, though unambiguous examples are few. Delbrück (12) identifies two instances, both from the Satapathabrahmana:

   SB I 2,3,2 Í'ity áha lád indró 'mucvata, devó hi sáh. Indra, assuredly, was free from that (sin), for he is a god. SB I 3,1,11 tád dháike sruksammarjanany agnav abhya dadhati vedasvahabhuvant srúca ebhih sám amarjisttr idám vái kimcid yajñásya néd idám bahirdha yajñad bhávad iti Here now some throw the grass-ends used for cleaning the spoons into the (Ahavaniya) fire. "To the veda (grass-bunch) they assuredly belonged. They cleaned the spoons with them: (13) hence it is something that belongs to the sacrifice, and (we throw it into the fire) in order that it should not become excluded from the sacrifice," thus (they argue). Particularly striking is the contrast in SB VI 6,4,12, where the same clause is given first without áha and with unaccented verb and then repeated with áha and with accented verb:

   SB VI 6,4,12 ghrténa tviim tanvàm vardhayasva [RV X 59,5] satyah santu yájamanasya karna [RV X 116,8] iti ghrtènaha Ivám tanvàm vardhayasva yéhhya u tvam kãmehhyo yalamana adhatta tè 'sya sárve satyah santv ity et at. "With ghee make thou grow thy body, let the wishes of the Sacrificer be true!"--that is. "With ghee indeed make thou grow thy body, and for whatever wishes the Sacrificer makes up a fire, may they all come true!" The accentual properties of áha cannot be assessed in earlier samhita prose, where all instances of an accented verb in an áha-clause can have the accent for other syntactic reasons. (14)

   In early verse this accent rule is not consistently applied. We find for instance an accented primary verb in

   RV I 6,4 ad áha svadham ami púnar garhhatvám eriré I dadhana nama yajñiyam // Certainly áha. just after that they once again roused his embryonic state [= kindled the fire] according to his nature, / acquiring for themselves a name worthy of the sacrifice. (15) AV I 34,2 jihvaya agre mádhu me jihvamulè madhulakam I máméd áha kratav áso máma cittám uparasi // At the tip of my tongue honey, at the root of my tongue honeyedness; mayest though be altogether in my power, mayest though come unto my intent. (16) AV VII 39,4 ahum vadami nét [t.sub.u]vám sahhavam áha trám váda I I am speaking; not thou; in the assembly verily do thou speak. (17) But contrary to this we several times find áha followed by an unaccented verb in the RV as in

   RV X 86,2 pára [h.sub.i]lndra dhavasi vrsakaper ati vyathih I nó áha prá vindasy anyútra sómapïtaye risrasmad indra uttarah // [Indrani:] "But although, o Indra, you run away, beyond the wayward course of Vrsakapi, / you do not find áha anywhere else for soma-drinking."--Above all Indra! Though the situation is admittedly somewhat messy, the use of áha in a main clause with an accented verb is a persistent feature of the language and can be considered a "syntactic lectio difficilior." It is not well grounded in the synchronic grammar of Vedic and hence demands a historical explanation.

   One possibility is that áha started out as an emphatic particle and then developed a special association with subordinate clauses, where accentuation of the verb is the norm. The rule that the verb in an áha clause must take the accent would then have been generalized from the innovative subordinate clauses to the independent clauses in which áha operated as an emphatic particle. But this development is not trivial. The use of the verbal accent to distinguish subordinate clauses from main clauses is a central feature of Vedic syntax and prosody. It is hard to see why speakers would have mistaken the accent in a dependent clause containing áha as being conditioned by the particle at a time when áha was also used in independent clauses with accentless verb, or why this usage should have spread from the former to the latter.

   The other logical possibility, and a far more straightforward one. is that clauses with áha contain a verb the accent of which used to be well motivated under the synchronic rules of pre-Vedic grammar. Over time, the syntactic environment motivating the accent was lost but the accent itself remained as an archaism. Precisely this explanation holds for the particle hânla [[approximately equal to] Pa. handa, Pkt. hanta/ handa/handi], (18) which consistently introduces a clause with an accented verb. (19) Thieme (20) argued convincingly that hánla continues a 2pl. pres. ipv. to the verbal root han 'strike'. (21) We thus find, for example, SB I 2,5,2 hántemam prthivim vihháiainahai "Well then, let us divide this world between us" with accented verb. The accentuation of the verb in a hanta-clause can most easily be taken as a fall-out of the fact that hánta historically expressed the primary verbal action, while the second clause contained a concomitant verbal action that was logically subordinated to the former. (22) In light of this fact, a verbal origin for áha would plausibly make sense of its accentual properties.

3. With the establishment of some basis for supposing that áha could be of verbal origin, Oberlies's suggestion of connecting this particle with the verb aha takes on renewed appeal. The question then becomes, what morphological relationship could áha bear to aha? In order to answer this, it will be necessary to examine the paradigm of aha and its Iranian cognates.

   In the Samlina texis, the preterito-present aha is used only in the 3rd person, other forms and tenses being supplied variously by other...