Causatives in Sanskrit

[Vidhyanath Rao]

The recent thread about passive reminded me about the questions I have
had for a while, about causative and double accusative in Sanskrit. I
hope that members of this list can shed some light on this.

Speyer (sect. 49, p36 of Sanskrit Syntax) says (essentially) that the
case, in the causative construction, of the agent of the simple
(transitive) verb is determined `naturally'. Thus `.r.siin karam
adaapayat' (M.Bharata 1.70.26b) means `(he) made the rsis give tribute';
but `lekha.m "saka.tadaasenaalekhayat' would mean `(he) got the letter
written by S.'

In terms of Panini's rules, this would seem to result if Panini had
simply left out 1.4.52, 53. But, the presence of these rules, especially
the `anyatarasyaam' of 1.4.53, implies that this is not allowed for
other verbs. In that case, how would we make the distinction made above?
The problem is that there are no automatic demotion rules in Sanskrit.
`.r.sibhi.h karam adaapayat' would imply that the rsis were the means or
the cause etc. Now one could presumably say `.r.siin karadaanam
akaarayat'. But it seems odd that this was acceptable whereas `.r.siin
karam adaapayat' was not.

Another observation of Speyer adds to the problem. He says that with the
ta-participle and the gerundive of the causative, the agent of the
simple verb is put in the nominative, even when the active causative would
not admit double accusative.

In languages with morphological causative, the case of the subject of
the simple verb varies. Comrie (in Ling. Universals and Lang. Typology,
the chapter on causatives) says that putting the agent of the simple
verb in accusative is more or less limited languages with double
accusatives. [I presume that he is leaving out languages such as English
in which causative is expressed with a separate verb.] The interesting
thing here is that double accusatives become less common in Sanskrit as
time goes on (Speyer, p.35  fn 1). The apparent conflict on the number
of verbs that allow double accusative in the Bhaa.sya on 1.4.51 may also
be from the same cause. This brings me to the next question: How do
causatives work in Prakrits? How common are double accusatives there?
[This may be silly question because the -aya- would become -e-, which
often forms >simple< presents in Prakrits. But I am not sure that
morphological causatives are totally lost in Prakrits.] What about
Modern Indian languages, IA or not?

-------

Regarding the sentence `daana.m dadau raamam' in the thread:
daana.m dadau is of course possible, like donum dedit. But
wouldn't we translate `daana.m dadau gaam' as `He gave a cow/bull
(as a) gift'?

-------

Regards
-Nath