Personal Pronouns / Ergativity

[Wolfgang Schulze]

Vidhyanath Rao schrieb:

> Wolfgang Schulze <W.Schulze at lrz.uni-muenchen.de> wrote:

>> If we look at the question of ERG and passive, we can sometimes observe
>> a tendency to establish a pseudo-ergative strategy based on the passive
>> (some modern Indo-Iranian languages, partly Cl. Armenian etc.).

> The (middle and) ModIA ergative construction in the past should not be
> traced to a passive. It is based on the descendent of the -to adjective,
> which in Indic was always >resultative<. mr.ta means dead, not killed. More
> tellingly, a:ru:d.ha, `mounted' had an active meaning. It is called ``past
> passive pariticiple'' because 19th c. grammar writers did not have better
> terminology.

I used the term 'passive' in a functional sense which means that
'passive' refers to ANY means to background an agent and/or foreground a
patient. Naturally, this function is not restricted to the technique of
'classical' passivization. The pragmatics and semantics of such a
strategy are strongly language-dependent though certain universal
tendencies (not universals!) can be observed. One typical effect is that
of resultiveness which is especially strong with the perfective aspect
and non-SAP agents (see Hopper/Thompson 1980 for details). The main
point with (anti)passives is that their 'demoted' NP is located in the
periphery of a clause: Normally (I don't say ALWAYS!), it can be
deleted. In ERG structures it is important to note that ERG plays the
role of a subject or object (depending on the other techniques relevant
to accusativity/ergativity present). Demoted NPs in (anti)passives never
play one of these roles. AGAIN: Even if there are secondary
ergative-like structures in some IE languages (grammaticalized via some
kind of 'passive' / resultative or what so ever), this does not tell us
ANYTHING about ERG in PIE. One the contrary: PASS-based ergatives can
only emerge from accusative strategies. So, if you find parallels for
the (new) Indo-Iranian structures in PIE, this only tells us that PIE in
fact WAS of the ACC type!

>  There are good pragmatic reasons why resultative adjectives seem to have
> ergative-like agreement. If such constructions are used for general
> adjectives, it may seem like there are traces of ergativity. I wonder if
> such an explanation would work for PIE.

Ergative-like agreement with resultatives does not stem from an
'ergative' strategy but from the fact that resultatives often behave
like predicative adjectives. Their agreement structure is triggered by
the appropriate NP (something like 'man-sexus dead-sexus by NP'). A
'true' ERG agreement pattern would not by confined to such resultative
structures, but would be present in the whole paradigm, cf. (as an
example)

Chechen (East Caucasian)

	stag w-öl-u
	man(CLI):ABS I-laugh:PRES-PRES
	'The man is laughing'

	stag-a baga y-illi-na
	man-ERG mouth(CLIV):ABS IV-open:PAST-INFER
	'The man has opened (his) mouth'

Here, the noun class marker /w-/ (class I [male humans]) is triggered by
the intransitive absolutive, just as /-/ (class IV (some inanimates)) is
triggered by the transitivie absolutive. If we check AGR in PIE we
should stick to the reconstructed AGR system itself, but not for
possible secondary AGR systems derived from pseudo-adjectival
(participle-like) forms. These do not tell us anything with respect to
the basic question.

Finally, let me briefly refer to Pat's and Roz' discussion on animacy,
ergativity etc. First we should note that the semantics of animacy is
heavily depending on how people exeprience and categorize their world(s)
(at least in a diachronic sense). To caracterize "animacy [as] the
quality of a "conscious, volitional entity"" (Pat) surely is an option,
but not a must. Especially, 'volitionality' is a very doubtful feature.
In some so-called active languages 'volition' is NOT a distinctive
feature, rather an inferential one. 'Animacy' is not a categorial entity
but a name for the behavior of the lexcial representation of a cognitive
concept with respect to a scale which often is labeled animacy
hierarchy. Second, the problem of 'animacy' and 'agentivity' has - in
itself - nothing to do with ergativity/accusativity. It represents
nothing but a 'natural' tendency to prefer animate NPs with
high-transitive contructions. This is why speech act participants often
are 'neutral' with respect to the morphological part of the AEC
(Silverstein hierarchy). High/Low animacy is crucial with non-SAP
agents: The more 'animate' a nSAP agent is the less a specific marking
of its role becomes necessary (cf. names in the role of nSAP agents in
many ERG languages that are unmarked). Hence we can conclude that ONE
possible semantic aspect of ergative marking is the LOW degree of
inherent agentivity which necessitates its supplementory marking (there
are other strategies to use ERG morphology in order to overemphasize the
left side of the agentivity hierarchy, but that does not matter here).
But remember that again we deal with high-transitive structures only. An
ERG typology clearly demands that intransitive (or low transitive)
contructions are exempted from this strategy. Hence, we cannot not talk
about 'agents' and 'animacy' in general (when discussing possible ERG
features), but only about those that occur with (high) transitive
structures. To get back to PIE: There are no clear signs that animate
'agents' were ever differentiated according to the degree of
transitivity exerted by them. This again is a hint at an even *semantic*
accusativity of PIE.

Wolfgang

[Please note new phone number (office) :+89-2180 5343]
___________________________________
| Prof. Dr. Wolfgang Schulze
| Institut fuer Allgemeine und Indogermanische Sprachwissenschaft
| Ludwig-Maximilians-Universitaet Muenchen
| Geschwister-Scholl-Platz 1
| D-80539 Muenchen
| Tel:	+89-21802486 (secr.)
|      	+89-21805343 (office) NEW ! NEW !
| Fax:	+89-21805345
| Email: W.Schulze at mail.lrz-muenchen.de
| http://www.lrz-muenchen.de/~wschulze/
_____________________________________________________