"impersonal" passives with agents

[Annie Montaut]

on impersonal passives
frequent in Hindi as well as other Indo-aryan languages. Morphologically
new
forms built with the past participle of the main verb + auxiliary « go »
jânâ
(up to the medieval stage, ie 14-18th c the heritated form was
synthetic, using
the Sansdrit affix –ya-)
transitive passive : âtânkvâdiyon se/dvârâ do âdmî mâre gae
  terrorists by two men killed go-past-mpl
object is not necessarily promoted, since it often maintains its
accusative
marking (default agreement on the Verb, in the masculine sg)
donon ko mârâ gayâ, two acc killed go-past-sg, « both were killed »
intransitive passive kuch Tahlâ jâe ? some walk go-subjunctive-msg, «
should we
go for a walk ? »
aise nahin hansâ jâtâ, like-that neg laugh go-pres, « (one) should not
laugh
like that)
mujhse nahin calâ gayâ / bolâ gâyâ, I-instr neg walk go-past-msg /speak
go-past-msg
, « I was totally unable to walk, to speak »
the last 2 ex with the « incapacitive » meaning (negative or
paranegative
context always, instrumental agent always) seem to be attested very
early in New
IA, even with the synthetic form in –ya-, probably before the unmarked
(no modal
meaning) passive (Gaeffke, Untersuchungen zur Hindi Syntax, 1997 ;
Montaut La
construction passive en hindi moderne, Bulletin de la Société de
Linguistique de
Paris 1990, passif et moyen en hindi, Typologie 2, Linx Paris
10-Nanterre, in
press)
as for the maya gamyate examples in Sanskrit, Speijer (Sanskrit syntax,
Leiden,
Brill) gives some attestations from classical Skrt (Dashakumaracarita)
and it is
commented by Panini who gives a few exceptions of intransitive verbs
which do
not passivate.
What is usual in classical Skrt is the passive participle of both trans
and
intro to indicate a past process (maya gata, I went, like maya tat
kritam,I did
that), a pattern which developped into the ergative structure in  modern
IA but
only for transitive verbs.Renou (Grammaire sanscrite, Maisonneuve, 1961)
is more
detailed, although only 1 page (497-8, section 367). He says that right
from the
origin passive is purely formal, since verbs like âsyate (sit passive)
passivize, that grammars give the classical ex of intransitive verbs
with " the
noun of direction or time" as subject (gamyate grâmah, go-passive-3s
village-nom-ms, mâsa âsyate month-nom-s stay-passive-3s) sometimes the
noun
indicating time in the accusatif (mâsam). "Predominance of the passive
construction results into an impersonal which propagates and tends to
associate
with an instrumental oblique: as soon as Vedic for the present in -ya,
from the
beginning of literary period for aorist in -i, from Ramayana for the
obligative
participle in -tavya and from Kalidasa for the past participle in -ta,
the
latter generalised everywhere". He adds: in the Grammarians, all
passives can
occur as impersonnals, and reversely the 'authentic' impersonnals can
occur in
personnal constructions.



>