Published online by Cambridge University Press: 22 April 2026
Within the Vedic verbal system there is a series of second singular forms made by suffixing -si to the full-grade root, which differ from the present type ási ‘thou art’ in that they function primarily as imperatives; e.g. yákṣi ‘sacrifice’. This archaic formation is found frequently in the Rigveda, where approximately 150 forms occur from 23 roots (§2); in other Vedic texts -si imperatives occur only for roots which have such forms in the Rigveda and predominantly in Rigvedic mantras; e.g. 
maghónām ‘bring safely across the gift-giving of the patrons’ (RV 8.103.7, 9.1.3); 
vakṣi yákṣi ca ‘bring the gods and sacrifice’ (.5.21.6, 6.16.2, 8.102.16); see Bloomfield, Concordance sv. párṣi, 
.
1 This study was written in the summer of 1963 and communicated to the American Oriental Society on 8 March 1964. Other commitments prevented me from submitting a version for publication before November 1964. Meanwhile I have received a copy of an excellent work by Johanna Narten, Die sigmatischen Aoriste im Veda (Wiesbaden, 1964), in which she reaches the same general conclusion about the -si imperatives as I, namely that they function as part of the sigmatic aorist system, correlated with 3sg. subjunctives in -sat; cf. also my discussion of the correlations yakṣi/nakṣat, vákṣi/vakṣat, and satai/sátsat in Lg. 39.14–16 (1963). Since I approach the question from a different angle and consider some features not treated by Narten, I have thought it worthwhile to publish my findings. In the present study I indicate the main points of agreement or difference between Narten and me in notes, where the bare reference Narten indicates her work, mentioned above. In this work, Narten has also presented a new explanation of the Sanskrit seventh aorist, type ádhukṣat. In view of her treatment, I have withheld publication of my own study, ‘The origin of the Sanskrit seventh aorist’, which elsewhere (footnote 5 of my paper ‘The formulation of Pāṇini 7.3.73‘, JOIB 14 [Sept. 1964]) I indicated as forthcoming. Let me stress that Narten and I differ only in details; I am most happy to find that on the -si imperatives in general, as on sátsat (Narten 263 n. 827a), we agree almost completely.
I wish to express my thanks to Bernard Bloch for his advice on the form of this paper.
2 Benfey also lists ghoṣi as a -si imperative of ghuṣ ‘resound’. The form occurs in RV 4.4.8 and 6.5.6; 4.4.8ab: árcāmi te sumatíṃ 
sáṃte 
jaratāmiyáṃ 
; 6.5.6cd: yácchasyáse dyúbhiraktó vácobhistájjuṣasva jaritúrghóṣi mánma. In the. second passage ghóṣi is almost certainly adjectival (Gr. Wb. ‘lauttönend’): ‘When you are praised, take pleasure in the resounding hymn of the chanter'. Though ghóṣi in 4.4.8 is usually considered a 3sg. aor. (e.g. Gr. Wb. sv. ghuṣ), Oldenberg (Noten ad loc.) considers interpreting it parallel to 6.5.6, and Sāy. appears to have given a similar interpretation, though construing ghóṣi with 
. Geldner interpreta ghóṣi in both passages as imperative; ‘horche her’ (4.4.8), ‘höre auf sein Gedicht’. But this seems inappropriate, though ā ghuṣ means ‘hearken’ (cf. Wackernagel IF 45.310 [1927 = Kl. Schr. 1250]).
3 Thus Whitney-Lanman. BR, Whitney (Index Verborum), and Whitney-Lanman assign vṛkṣi to vṛj ‘twist’; Sāy, derives it from vraśc ‘hew’, which is more probable; cf. Narten, MSS 14.42 (1959).
4 For the interpretation of -si imperatives as present forms with zeroed thematic vowel, cf. also Sāy. ad RV 1.129.5 (neṣi, parṣi), 3.30.18 (āsátsi).
5 Oldenberg, Noten ad loc., cites Av. cašte ‘teaches’ as support for the transitive meaning; but such comparison can be misleading: bhī means ‘fear’, but GAv. nyentē (Y 34.8) is possibly transitive (cf. Hinz, Zarathustra 180, 220 with ref.). Bartholomae, IF 2.283, took cakṣi in 7.3.6 to have infinitive value, but his interpretation involved in addition an emendation to bhānúḥ and was rightly rejected by Oldenberg.
6 Oldenberg, Noten ad loc., takes aryáḥ as an acc. pl., as does Sāy., and Geldner translates ‘werde mit den vornehmen (Nebenbuhlern), den (anderen) Leuten fertig’. Thieme, Fremdling 11 n. 2, lists this passage among those he does not consider important for the meaning of ari, possibly indicating by this that there does not appear here the contrast between arí and jána ‘l'homme du commun’ (Renou, EVP 2.110).
7 I count as one occurrence the repetition of the same form in a stanza; thus jóṣi 2.37.6ab: jóṣyagne samídhiṃ 
jóṣi bráhma jányaṃ jóṣi suṣṭutím; mátsi 9.90.5: mátsi soma váruṇaṃ mátsi mitráṃ 
… víṣṇum / mátsi śárdho 
mátsi
mátsi 
; 
2.11.4ab: 
kṣáyaṃ 
mitrám … 
śárdha … 
naḥ.
8 An apparent example of an imperative in subordination is RV 1.127.2e–g: hótāraṃ 
jūtáye víśaḥ ‘(Agni) the hotṛ of men …, whom these clans should promote …’ But the relative yám may be adjectival, standing for yám hótā,ra som that Renou (Gr. ved. §430 note) is no doubt right in considering this ‘une subordnon aéepparente'.
9 It has also been supposed that the subjunctive is used as a narrative form; cf. Neisser, BB 20.70 note, and Geldner ad 6.17.11.
10 Similarly Narten 202.
11 Narten 116 also suggests the possibility that chantsi functions as an imperative in the passage mentioned.
12 Narten arrived at the same conclusion, 262–3 with note 827a.
13 Narten 261 decides in favor of deriving śroṣan etc. from the subjunctive.
14 Narten 214, 216, considers yutsmahi more probable than yuṣmahi, though she does not rule out the latter.
15 Similarly Narten 120, 164, where the possibility is also considered that nethá is conditioned by pātha in the same verse.
16 Jeṣam has also been connected with 1sg. yeṣam (yā) etc.; cf. Renou, op.cit. §344 note. I cannot offer a clear explanation of this type, though I would hesitate to accept completely the view of K. Hoffmann (apud Narten 120 n. 321) that jeṣam and yeṣam represent full-grade precatives (*jay-īṣ-am, yā-īṣ-am).
17 The 3pl. aveṣan (2 × RV) has been considered an s-aorist pf vī, but, as pointed out by Macdonell 379 note 3, this is morphologically impossible; the forms are rather from the verb viṣ. On this verb and the possible classing of two roots viṣ, see now Narten 214–5. Narten 104 attributes the use of kṣeṣat to the need of avoiding homonymy between subj. kṣayat(i) < kṣi ‘abide’ and ind. kṣayati < kṣi ‘destroy’, though she mentions the model neṣat, jeṣat. Narten 244–6 does not directly recognize impv. veṣi; but 246 n. 752 she states that if veṣi is impv. in some instances, one may compare neṣi/neṣat etc.
18 See note 17.
19 Norton 120 also considers joṣat as possibly derived from impv. jóṣi.
20 Debrunner, Nachtrag to Wackernagel, Ai. Gr. 1.275.8, considers voḍh- the phonetic development of *-aźh-s-t. Narten 240 n. 727 accepts the voḍh- forms as evidence of a second-class present.
21 Narten 191 n. 547 accepts Debrunner's view. Narten does not consider the possible influence of *dhāsi etc. on the formation of s-aorist forms.
22 Similar objection to Debrunner's dirent connection of rāsi and rāsate, Narten 221 n. 668.
