transitivity, locative prefixes, etc.

[Marianne Mithun]

Dear All,

I'm inclined to agree with Pam and Bob and others of you. I've always
thought of these prefixes (a-, I-, etc.) as prototypical applicatives,
that is, they derive a new verb stem whose added core argument is a
location, an instrument, etc.

As many of you know, applicatives can have two kinds of grammatical
effects when applied to verb stems that are already transitive. In some
languages, the added argument (location, etc.) replaces the original
second argument (usually a semantic patient), which then is not mentioned
at all or is grammatically oblique. In other languages, applicatives add
an argument, deriving a ditransitive stem from a previously transitive
one. And yet other languages have some of each.

Since the applicative morphology is derivational, and creates new lexical
items, the new stems can take on a life of their own and develop in
idiosyncratic ways, sometimes with a shift in transitivity. So what might
have started life as a derived transitive could be reinterpreted as an
intransitive, etc.

Actually, Athabaskan languages show very much the same kind of situation,
as many of you know. Most have a large inventory of separate
'postpositions', which carry pronominal prefixes just like verbs do:

	me-for   you-cook

etc.
Over time, some combinations of postposition and following inflected verb
have come to be recognized as lexicalized expressions, and the word
boundary has broken down, so that in those cases, the former postposition,
preceded by its pronominal prefix, has become the initial element of the
verb word. In some cases both forms persist in the language: independent
postpositions (with pronominal prefixes) and cognate applicative verbal
prefixes (with pronominal prefixes), typically with the verbal prefixes
slightly more eroded in form, as would be expected. Some of the group that
were undoubtedly originally postpositions now survive only as applicative
prefixes on verbs; their postpositional sources have dropped out of the
language.

In some cases, speakers still have choices between a postposition followed
by a verb, and an applicative verb. As might be expected, the independent
postposition in these cases, puts separate attention on the postposition,
while the applicative verb is interpreted more as a single lexical unit,
often with more idiomatic meaning. Of course applicative constructions
make that additional argument a core argument, so it is more likely to
happen to animate arguments, especially first and second persons.

I tend to feel the same way as Pam: a major formal sign of the difference
between core and oblique status is whether pronominal reference is in
the verb or not. Of course with Siouan third persons, there is less
visible difference, with zero third persons.


Marianne


On Sat, 28 Sep 2002, R. Rankin wrote:

>
> All,
>
> Many thanks indeed for all the really fascinating
> input.  I sort of figured that this might be like
> squirting the garden hose at the hornet's nest.
> Obviously a lot depends on what model of grammar one
> tends to "think in".  It also depends on whether we
> take a semantic or a purely morpho-syntactic point of
> view.
>
> > To me, the terms "direct" and "indirect" for objects
> don't seem
> > particularly appropriate for this type of language,
> because there is a
> > mis-match. Often, in a language like this, the
> indirect (recipient)
> > object of a verb like 'give' is the one that will
> agree ( if it's one of
> > Shannon's "local persons").
>
> I've always sort of liked Matt Dryer's notion of
> "primary object" and "secondary object".
>
> > Thus I was not fully informative in saying that I
> thought the locative
> > argument in Bob's original sentence was (probably) a
> DO. (I don't know
> > Osage, so this is really all speculation. But that's
> my gut feeling.)
>
> Well, Osage works generally very much like Dakota.  I
> think we can say that of nearly all Mississippi Valley
> Siouan languages (I haven't looked at Winnebago).
> Pam's "gut feeling" was mine too, and that's what I
> told Carolyn.  Like Randy, I noted that very many
> locative prefixes are part of their verbs and range
> from totally transparent to totally opaque in
> analyzability (I haven't had my coffee this a.m. so I
> don't know if that's spelled right.).
>
> > So these locatives might be "semantic obliques", but
> they aren't
> > "formal" obliques, at least for me.
>
> I have a feeling that talking about "semantic obliques"
> (or other semantic analyses) is just a fancy way of
> saying "we're translating this sentence into English
> and then analyzing the English." . . . something that
> used to get Mary Haas's hackles up.  I guess, like Pam,
> I'm trying to think of things in morphosyntactic terms
> taking into consideration "the genius of the Siouan
> languages", i.e., the fact that they have real
> postpositions that allow for oblique arguments,
> ordinary PP's, BUT they also have these three locative
> prefixes that are "different" from postpositions.  Why?
> What does this difference imply grammatically?  Etc.
> That's what you guys are helping so much with here.
> Thanks again for your comments, feelings and analyses.
> All are very valuable to us.  I'd be happy to hear more
> along the same lines.
>
> Best,
>
> Bob
>
>