Re: grammatical cases & semantic roles (was: ergative/accusative)

Hallo!

On Mon, 29 Jan 2007 19:58:26 +0000, R A Brown wrote:

> MorphemeAddict@WMCONNECT.COM wrote:
>
> > [Rick Morneau's book]
>
> > His treatment of argument structures is not mainstream.
>
> It certainly isn't, is it! Your mail send me to look out the stuff I
> have about Rick's 'Machine Translation Interlingua'.
>
> Yes, he does appear to use 'focus' as the label for a 'case'. I think,
> however, his use of 'patient' and 'agent' are clear indicators that he
> is not talking about surface grammatical cases but about 'deep cases'
> i.e. semantic roles. This makes sense, I think, in discussing an
> interlingua for (universal) machine translation: we need to get at what
> a sentence/utterance etc _means_.
Yes.

> Even so, I find his use of 'focus' unhelpful in that it already has
> another linguistic use.
Yes.  "Focus", as it is usually understood in linguistics, has nothing to do
with either cases or semantic roles.  It is a *pragmatic* cateogory.
(There is a second meaning of "focus", though.  In Austronesian linguistics,
the word is sometimes used for a noun case whose semantic function is marked
on the verb.  But that isn't what Morneau means, either.)

> Personally, i think it can (and does) cause confusion to use 'case' to
> denote both surface, grammatical features and semantic roles. Although
> there is some correspondence between the two, it is very far from being
> identical.
Very much so.  In most languages, the connection between grammatical case
and semantic role is at best a very loose one.

> As I said, I agree with you that Rye is using labels wrongly in
> http://www.xibalba.demon.co.uk/jbr/ranto/r.html
>
> His error IMO is his apparent 1 to 1 mapping of grammatical case to
> semantic role, e.g. that the subject of a transitive verb is always the
> patient - it ain't.
AND what is especially wrong in Rye's article is to call an intransitive
subject an "experiencer".  Intransitive subjects can be just about
anything, and experiencers are not typically intransitive subjects
- the archetypical experiencer is the subject of a verb of perception
or emotion; some of these verbs are intransitive, others not.

Also, a transitive subject isn't always an agent.

Unfortunately, I have seen Rye's "nonstandard" terminology on many webpages,
usually ones describing ergative or other non-accusative conlangs.
If I could have $100 for each repetition of this mistake ...

> BTW - Rick Morneau says "All verbs have a patient, whether stated or
> implied." Is that in fact true?
>
> What is the patient of the following
> LATIN   SPANISH   ESPERANTO   ENGLISH
> pluit   llueve    pluvas      it's raining
> niuit   nieva     neghas      it's snowing
>
> Of the languages above, only English gives the verb a grammatical
> subject - the dummy 'it'. What is the patient implied in those and
> similar verbs?
There is none.

... brought to you by the Weeping Elf

Reply