Re: THEORY: unergative

Philippe Caquant wrote:

> I already noticed that some languages are half-accusative,
> half-ergative, like Georgian, where the same sentence
> will be expressed a different way, depending if it is
> in present or in past tense. This I find absolutely
> confounding. I have to find out why it is so !
The answer to your question is explained in Dixon’s book “Ergativity”
(Cambridge University Press, 1994) in which he states that many languages
distinguish absolutive-ergative patterning from nominative-accusative
patterning based on either tense, mood, or aspect and that it will almost
always be past tense, perfective aspect, or a factual/indicative (as
opposed to subjunctive) type of mood that will be show absolutive-ergative
marking. The reason is as follows:  since non-past tenses, imperfective
aspect, or non-factual moods do not indicate events that have actually
happened, they are best thought of as indicating a propensity of the
potential AGENT, not the PATIENT, e.g., ‘That man might hit someone,’
NOT ‘That man might get hit by someone,’ this being a distinction between
transitive subject versus intransitive subject (i.e., ergative pattern), as
opposed to a distinction involving a transitive object (i.e., accusative
pattern).  To quote Dixon directly:

“…ergative marking is most likely to be found in clauses that describe some
definite result, in past tense or perfective aspect.  An ergative system is
less likely to be employed when the clause refers to something that has not
yet happened (in future tense), or is not complete (imperfective aspect) …
or where there is emphasis on the agent’s role (imperative or hortative
moods).”

Dixon goes on to point out that besides tense/aspect/mood splits in
ergative versus accusative patterning, there are other kinds of ergative
vs. accusative splits seen in natural languages, e.g., main versus
subordinate clauses, positive versus negative sentences, and third-person
pronouns versus first and second person.

I believe that cognitive linguists have also explained such ergative versus
accusative “splits” on grounds similar to Dixon.  Essentially, the idea is
that where a language imposes an unconscious cognitive need to distinguish
the definiteness or reality of an agentive result on a patient from a mere
possibility or propensity of a result, then such splits will be seen.
Applying this logic to the pronoun split, because first and second person
pronouns involve the speaker and addressee directly, each (presumably)
knows whether a definite result has occurred.  However, third person
sentences may be spoken about parties not present (and who therefore cannot
verify the definiteness of a result); thus third-person sentences become
more likely to bear ergative patterning as opposed to accusative patterning
(as in Georgian).  Anti-passive constructions in turn arise as alternative
syntactic patterns to “de-emphasize” or "question" the definiteness of a
result.  All of which sounds to me like fairly good evidence in support of
the Sapir-Whorf hypothesis (the "weak" version, not the strong, that is).

--John Quijada

Reply