44. It was rather abruptly that the earliest past forms Asa Asus and Aha Ahus were analysed as a-as-a a-as-us and a-ah-a a-ah-us, and then two unforeseen verbs as (to be) and ah (to say) emerged. Then the bisyllabic verbs, e.g., cara began to be reduplicated through the first syllable : ca-cara. This was the beginning of the perfect formation in Vedic.

44a. The agglutinative verbal form A-sa (it was, there was) itself later lost the final a due to accent, and As (RV 10,129,3b) was there at a very late stage. When the persons began to be distinguished, As- was extended by -am for the 1p. but by -s- and -t- with the interposed -I- for the 2p and the 3p. respectively. Thus three separate forms emerged:

3. As-I-t

2. As-I-s

1. As-am

A similar exercise was carried out for the plural:

3. As-an

2. As-ta

1. As-ma

We know how these forms were explained by Panini: he took A- as the past augment, as the strong base in sg., and s as the weak base elsewhere; and then he added the endings:

3. A-as-I-t	A-s-an
2. A-as-I-s	A-s-ta
1. A-as-am	A-s-ma

45. The past participle with -Tha at the agglutinative stage of Indus is attested. But there was another such affix, namely -ma, added specifically to the verb Sa (to be), making Sa-ma (be-past), which appears as the contracted past postposition sma after the present forms in Sanskrit. Thus gacchati sma = he goes-past (he went, used to go). It is presumed that this past participle expressed a habitual action. It could not, however, be popular in the language.

a. Corresponding to the past affix -ma at the agglutinative stage and contracted with the verb ºa (to be) as s-ma at the inflexional Vedic stage, coming down to Sanskrit, there were two more affixes at the agglutinative stage, remaining unnoticed till now.

(i) The affix ha made perfect tense, as even now in some rustic dialects of Hindi, but it was fossilised with the verb sa (to continue) as the particle saha (with), lit. ‘has continued’, in Sanskrit.

One more Indus perfect form appearing as suffix in Sanskrit is –maha: ma-ha (time-perfect = hs grown old) appears after pitA and mAtA, pitA-maha (grandfather) mAtA-maha (grandmother), literally ‘father (and) mother grown old’.

(ii) The affix ra made pluperfect tense, as even now in some rustic dialects in Hindi, but it was fossilised with the verb ca (to take place) as the secondary suffix -cara (had taken place). It was noted by Panini (5,3,53) as having ‘bhUta-pUrva’ sense.

Besides –cara appearing as the secondary suffix, other pluperfect forms, e.g. mAra cupid), dAra (wife), jAra (parampur), etc. have appeared as nouns.

b. Needless to say, it has been repeated so often, the future affix gha of the early agglutinative stage, appearing more often as ghA at the Vedic stage, remained dormant for a long time till it appeared in Hindi as the future gA affix in the same condition (jA-e-gA ‘he will go’ from jA-e ‘he should go’ from jA ‘to go’).

The Middle Voice

46. At least from the statement of Panini (P. 3,4,79-81) it appears that the middle endings came later through the e-bending in the original affixes. It is also clear from the fact that the older verbs do not have the middle forms. The middle forms should have come after the emergence of the present forms. Therefore, there are 3. -te 2. -se 1. -e corresponding to 3. -ti 2. -si 1. -mi. Just like the 1p.sg. -e (for the expected *me), there is some anomaly in pl., except in the 3p. (-nte). The 2p. pl. is the unexpected -dhve. The 1p. pl. -mahe derives from -masi when a corresponding -mahi had developed in the past tense.

a. The middle endings for the past tense show
innovation only in the 2p., -thAs in sg. and -dhvam in pl. The 3p. -ta in sg. and -nta in pl. represent the original past endings for the 3p. The 1p. -i sg. and -mahi pl. are the expected endings.

b. If we try to imagine the corresponding Indus affixes at the agglutinative stage, we find that the middle
endings were not contemplated in the past tense. They were the same in active and middle. Then -thAs and dhvam of the 2p are almost the Vedic development. The -i and -mahi of the first person are the result of the weakening of -e and -mahe due to accent on the first syllable (á). Therefore, they may have originated even at the inflexional Indus stage.

47. Thus the Indus middle endings for the present and past tenses may be the following:

Present	Past
3. -te  -nte	-ta   -nta
2. -se - - -
1. -e  -mahe	-i   -mahi

Because the present endings have emerged after the emergence of the past endings, the latter are, more or less, original. That is why, they also serve for the middle forms with certain additions and alterations.

48. The past endings serve for the optative with a prefixed I. In 3p. pl., an -r- element intervenes, which is actually a phonetic transformation of -n-, as in other cases (cf. s'arvan : s'aravar-I). In 1p. sg. -i is extended by a, inviting an intermediate -y-, making -Iya. The whole set is as under:

3. -I-ta	-I-r-an
2. -I-thAs	-I-dhvam
1. -I-y-a	-I-mahi

49. Panini derives the imperative endings by -Am bending of the e of the present. Whether it goes to the inflexional Indus is very doubtful. The 2p. sg. -sva is an innovation:

3. -tAm	-ntAm
2. -sva	-dhvam

50. The subjunctive middle endings of the Vedic are just the partial strengthening of the present endings and have the distinctive a- at their beginning:

3. -a-te/tai	-a-ntai
2. -a-se/sai	-a-dhvai
1. -ai	-A-mahai

Hardly few of these Vedic endings can be taken back even to the inflexional Indus. They are the Vedic developments.

Duals

51. The dual forms should have started with declension and spread over to conjugation, initially with the subject in the third person.

Let us run over the Vedic dual endings.

a. The modal endings are not different from the past.

b. It appears that the dual endings are the arbitrary creations at the Vedic stage, though they must have germinated at the inflexional Indus stage.

c. When 3p. -tas was conceived, a -thas was conceived for the 2p. Similarly, 3p. -te was distinguished by 2p. -the in the middle. In the past tense, 3p. -tam is distinguished by 2p. -tam in active, and 3p. tAm by 2p. -thAm in middle.

i. Roughly, -t- is for the third person, and th- for the second person, the past active and middle distinguished by the absence or presence of i/Ù at the beginning.

d. In the first person, -v- is a distinctive dual sound, contrasting with the sg. and pl. -m-.

The Vedic Verbs

After the verb tan (to expand) was deduced from the vocable tánas (offspring) and even after tánas itself was further reduced to tars (to be thirsty), the monosyllabic verbs ending with consonants had begun to emerge, already at the inflexional stage of Indus.

We have also seen that, already at the agglutinative stage of Indus, a monosyllabic verb (e.g. ga, ra, etc.) had begun to be extended by the semivowels and nasals as well as, less often, by other syllables:

ga	gama	gacha
ra	rama	racha
pra	pracha

The following are based exclusively on the Vedic examples:

bha	bhava
ha	hava
ja	jaya
ka	kara

These are the monosyllabic verbs with a. It may be supposed that the monosyllabic verbs with vowels other than a were also extended by stops and sibilants:

tu	tuda
di	diva
bu	budha
bhU	bhUSa
na	nas'a

It has been said that, at the inflexional stage, when-ti-si -mi were added to distinguish persons and when -cha-bore the accent, the preceding ra was reduced to R: pracha > pRchá-ti, racha > Rchá-ti. When -nó- was added to kra, the reduced form of kara, it was further reduced to kR: kR-Nó-ti.

When the bisyllabic verbs diva, budha, etc. were further extended by ya, the preceding a was lost: budha-ya > budh-ya-ti.

When the final a of the extended verbs, itself considered as infix between the verb and the personal endings, the remaining parts contracted and reduced to hu,ji,hR, etc.

It was thus that the Vedic verbal forms were systematised as having the present, perfect and aorist bases with the present and past tenses as well as the subjunctive, imperative and optative moods. Later a future tense with the corresponding past and a precative mood developed from the aorist. The past forms without the augment were reckoned as the injunctive mood.