¹ According to K. Hoffmann, the unaugmented forms distinguish aspect, but not tense or mood; their negative content is summed up in the term ‘memorativ’ (Der Injunktiv im Veda, Heidelberg, 1967, cited below as H.)Google Scholar.

For J. Gonda, following Renou, the content is still more negative: ‘a hard-and-fast line between aor. and pres. stem seems to have been wanting in the case of the injunctive… indifferent with respect to time and modality’ (The aspectual function of the Rgvedic present and aorist (Disputationes Rheno-Trajectinae, , VII), ‘s-Gravenhage, 1962, 266Google Scholar; this study will be referred to as G.).

In Aorist v ‘Rigvede’ (Moscow, 1960Google Scholar, in Russian), T. Ya. Yelizarenkova paraphrased Renou: ‘the injunctive… an unmarked and neutral mood… possesses a wide range of modal meanings’; ‘tenseless’, it is ‘more accurately for the most part indifferent to distinctions of tense’; aspectually ‘it can be treated as expressing both perfective or imperfective action and state’(pp. 56 and 120 ff.). The reference to state does not stem from Renou.

There will be reference also, as K., to Kurylowicz, J., The inflectional categories of Indo-European, Heidelberg, 1964Google Scholar.

² Imper. 3rd sg. vastām, 3rd dual and p1. ἔοτων suggests that the 3rd dual was used for p1. and sg. (where -tōm → → -tōt originated: cf. 2nd sg. -som « dual -tom in λ⋯ςσν ?). It seems likely that the type vittāt/vastām antedates dāt (→ dātu, dati)/dātām and δῷ δ⋯σθω, with *dōt based on 2–3 dual, 2 pi. Imper. (στητν/dātm, στxetaτε/dāta).

³ For redupl. Aor.vīvāpat, cf. RV 10.142.4 vápteva śmáśru vapasi prá bhṹtma (ague) and (with H. Humbach, ad loc.) 6.6.4. That vōiždaṯ 'stimulated, made erect’ is a thematic Aor. equivalent to RV váyo dhāt (vayodhā ‘stimulator’) seems certain. The idea of stimulating the daēvic weapon against the rtāvan occurs in RV 8.12.7 ‘Indra's ketávah and vájra grew within his hands, as he raised up Heaven and Earth, usurping the function of Sūrya (sūtryo ná… ávardhayat)’; the Pahl. gloss accurately renders vadarā vōizdat by pad paydāgīh ōŝiŝn dād, where ōŜiŜn may convey the basic implications of ‘perdition’ better than vadarā, and ‘manifestly made’retains much of the force of váyo dhá-. Similarly YAv.vōizdat ‘raised up’ presupposes the Indo-Ir. connotations of the weapon, and not merely the GAv. passage. The latter yields *vayah ‘evil manifestation, perdition’in Y 53.6–7 vayū.bzrvdubyō dregvō.dabyó and vayói anhaitī vacō. Here vayōbrt (RV váyaḥ bharāmahe) has acquired the sense of vajrabŕht, viz. ‘bringing destruction’, in view of the other passage (Pahl., rightly: ōŜiŝn ast pad gōwiŝn) which reverses the original sense. For the palatal nom. sg. vayōi, compare the labializedvayū-; cf. also RV vácase váyo dhāh. Dr. D. N. MacKenzie kindly drew my attention to the gloss ōŜiŜn for both vayū- and vayōi: it had been misread in AIW.

⁴ In Y 32, sru- has the connotation of heroic tales (śrávo nrrāam). The basic Indo-Ir. śruti might appear to the Iranians a derogatory smrti which perverted God's revelation. Yama, child of the Daēvas, misuses his power (vadara — RV 8.12.7 vájra) to intrude upon his parents Earth and Heaven (8.12.8b, the devouring of Gau; 8.12.7c and 9a, the dispossession of Hvar: assumption of the parental relationship seems to point the moral best, multiple affiliation being normal; cf. the offence of seeing Purūravas). For the devastation and the offence against the single Rtāvan, cf.ny òsati in 8.12.9 and 1.130.8.

⁵ I use ‘eventual’ to denote a modal category embracing all inj., subj., fut., and opt. values; and the terms ‘contingent’ and ‘prospective’ will distinguish subordinate-clause functions (excluding the non-modal relative) and main-clause functions within this category.

RV offers a single eventual mode, in which the use of ‘Inj.’, ‘Subj.’, ‘Opt.’, and ‘Tut.’ morphemes is regulated by inherited patterns of conjugation (e.g.paśyad aksayvān ‘who has eyes will see’, bhasad yamasānáh ‘curbed, he will snap’, prbánn abhi syāt ‘who gives will prevail’; vocā, vocam, voceyam; mā hāsmahi, nét tápāti, mā bhujema in final clauses. An obsolescent contrast subsists between contingent and prospective formations: compare consecutivemádema, ásāmā, jesāma with prosp. madāma, syāma, jesma (beside more modern prosp. madema, jayema, Av. ánhāmā). Were it not for prejudice based on Gk. and later Vd., more consideration would surely have been given to the explanation of e.g. bhavāma/bhavema and bhavāsi/bhaveh in terms of formal renewal, with semantic differentiation that is substantially post-RV.

The old explanation of the Gk. and Vd. long-vowel Subj. as an Aor. in *ē (and *ā, *ō, with adaptation to the indicative ?), and the current explanation of the Opt. as an Aor. in*ē may be true in the same sense that the pret. has made a contribution to the -t Subj. once constituted (gāt, gamut, etc.). The Opt. has features which recall the common 2–3 sg. (cf. peyāh) and defective 1st person of the Imper. If, as suggested above, 3rd dual Imper. active *-tōm underlies both the active and middle radical Imper. and Subj., a thematic 3rd dual Imper. middle *-eitām (Sk. bhavetām, cf. K., 155) could yield the basic active paradigm with 2–3 dual θεῖτην/(dhethe), 2 pl. θεῖτε/dhetana, 3 pi. θεῖτην/dheyuh→ 1 sg. *dheiō and 2–3 sg. *dheiēs. The change *dheiō » θε⋯ην/dheyām seems to be attested bydhām « *dhāni and the complete absence of 1 sg. Subj.-(a)yā in RV. As in the subsequent ατη-ι/-t/sthāt(i) Subj., which I assume underlies the long-vowel Subj., there is no ancient form of 1st dual and pi., judging by rival θε⋯ημεν/aśyāma θεῖμεν/sol;asema and divergent θε⋯μεθα ∼ (n)aśīmahi.

Fundamental may be the contribution by the Perf. of a short-vowel Subj. in -a(l)/-epsiv;(iota;) and -ate/-εται (stavate < stave). Such derivation would indicate that the contingent function is primary, since condition and purpose will be conveyed by the contrast of anterior subord. clause and non-anterior main clause, failing a Subj. (cf., renovated with Aor. morphemes, 7.20.6yá āvivāsāt ‘who has gained > who will have gained, will gain’and Y 32.4 yāat framīmaθā ‘since when you have instituted > so that you would institute, were about to institute’; renovated afresh from the Perf. is 6.14.1 yó jujósa… bhasat… vurīta). The secondary prospective function, embracing opt., pot., prescr., and fut. values (all may be attributed to the prospectives in 7.20.6), is a linguistic luxury which may be replaced by the indicative (dadhanti, naksati, vivāsati, 6.3.1 pāsi, 1.41.1yám ráksanti… nū cit sá dabhyate jánalj,): hence the tendency to impute pres. value to resat, ksayat, etc., although they remain restricted to prosp. contexts.

The development of the Indo-Ir.-sya- eventual from the sigmatic Opt. is reflected in the reinterpretation of Aor. Opt. karisyāh (Say. ‘krtavān asi’) as a contingent in -ā-: hence prosp. karisyati and (if corruptly 1st sg.)karisyā, after the model of contingent dāśāt∼prosp. dāśat(i). For the Baltic Fut., see now K., 115.

⁶ Thus bhavāti, Av.b^uvat(i) versus obsolescent bhuvat, Av.bavāt; abhavat, MIA ahu(vā), Av. (a)bavat versus obsolescent bhuvat, Av.b^uvat. Despite H., p. 236, n., the distinction is well maintained between pret. (pouiruyō) bavat in Y28.ll (with universal MS support) and Subj. buvat(i) in Y 30.9–10 and 33.10 with strong MS support (incl. a case of bavanti sec. m. for bvanti).

Extension of -i may rest on the loss of the distinction between contingent and prospective in kárate/karat, yielding eventualkarat(e) ∼ prosp. karati. A further development would be Subj. dadhat(e) naśat(e)∼Pres. dadhanti naśanti. The acquisition by the -ā- Subj. of the status of a derivative from -at(i) must account for gacchāt(i) versus bhūt, syāt, bhaveh. The category -et remains rare in RV, frequently renewing Subj. -a- (e.g. SV videt for RV vidāt) and seemingly built on occasion direct from root Aor. (AVbhideyam); its specialized conditional use is in RV uniformly hypothetical, a classical Subj. function. The distinction between contingent and prosp. seems better preserved in cases where the Subj. was slow to acquire also the status of a derivative from the Pres.: cf. dāśāt/padāti (∼dāsti, padyate) versus eventual bhavāti (<*bhuvät ?), arcat, asat(i), bhinadat, etc. Restriction to main-clause function is not peculiar to the -t Subj. but reflects a dichotomy within the eventual as a whole, with which I have sought to come to terms on the assumption of a Perf.-based -a(t) -(at)e eventual, and an Imper.-based -yāt -et -t -at eventual; -ate and -at, as recent formations, have appropriated the purely contingent function.

⁷ For anterior or historic value of śrutáh here (despite the interpretation given by v. 2), cf. the contrast between hist, śrnvisé (parāváti) and actual śrutáh (arvāváti) in 8.33.10. For syántā… ádhranah ‘wayfaring’, cf. 2.13.2 ádhvā … anusyáde’ the road is to be travelled’.

It seems that stosi is historic, and the contrast with vidāyyah (cf.tyé… tyé in 5.6.6–7) indicates that it was felt to be non-finite. It is, however, in use (as in probable origin) the past tense stusé (3rd sg. pass, in 1.122.7), and its use in the 2nd person will rest on misapprehension such as would arise from cárkrse in v. I. In 8.62.8, harnsi is felt to be 2nd sg. act., but an original 3rd sg. pass, might be involved: (grné tát … yádd hánisi vrtrám … śacīpate.

⁸ Rival interpretations of v. 6 are offered in v. 7 and vv. 12–15. In v. 7, kádarthā was taken as singular, ‘desiring what’; in v. 12, as fem. pi., ‘in the wrong direction’. In v. 7, na á grhám [prcch-] ‘attend our house for (a welcome)’ was understood, cf. Skt.āprcch- ‘salute a host or guest’; v. 7c resumes v. 6a prcchate vām, and v. 7d seems to gloss ā jagmathuh… mártyam as ‘you attacked the destructible (Susna)’. Similarly, vv. 12–15 supply verbs for each phrase:

v. 12 kiidhryág asmé [mā bhūvari]∼kádarthā, na ā grhám

v. 13tā te satyá [santit] ∼ diváh

v. 14yát ksáh [vardhata]… śúsnam ní śiśnathah∼gmáh … ā jagmathuh

v. 15utá maghónah [trāyasva]∼ca mártyam.

Hence the weird constructions in vv. 8–11.

For indicative yádd hán, cf. H., 192; tád … yád is never ‘because’, but circumstantia ‘that’, cf. 1.131.4 asyá vīryàsya… yád avātirah, 8.62.8 grné tát … yádd hámsi vrtrám, and in the present hymn, v. 11maksū tū le [satyā, v. 13]… yádd ha śúsnasya dambháyo jātám.

⁹ The refutation of a publication of 1928, long outmoded, not least by the later researches and translations of its author, has seemed a curious undertaking, justified, however, by the extent to which its findings remain the basis of discussion and controversy to-day. His guiding principle ‘l'intention rituelle rejoint ou recouvre la donnée mythologique’ remains the key to Rgvedic grammar; it may be upheld without impairing the integrity of the verbal system, observing the replacement of historic slant by actual astāvi in the ‘colophons’, and the complex interplay of theology and ritual which results from the ‘représentation actualisée ‘in 9.109.13–15 of the cosmic events described historically in 9.64.10–12 by means of unaugmented preterites.