LINGUIST List 4.108

Wed 17 Feb 1993

Disc: Subject/Object Asymmetry and Idioms

Editor for this issue: <>

Directory

1. , Re: 4.95 Subject/Object Asymmetry and Idioms
2. , Re: S/O Asymmetries
3. Jon Aske, explaining subject-object asymmetries?

Message 1: Re: 4.95 Subject/Object Asymmetry and Idioms

Alexis Manaster-Ramer has made an important methodological obervation
here: that idioms have been used as 'evidence' for analyses within a
framework where idiomaticity has at most a marginal role, and in the
absence of any general theory about the relationship between idiomaticity
and syntactic structure. For example, the problem of phrasal/prepositional
verbs has been largely ignored in the debates on Prep-stranding and reanalysis.
 I wonder if a similar problem arises with the use of language games
as evidence for phonological analyses, which have taken on great importance
in some recent work. It is not a priori clear that backslang, for example,
is relevant to the phonology of English. Language games are a meta-
linguistic activity, like drawing tree structures or metrical grids,
and whether they reflect internalized phonological representations
is a substantive question. Has anyone addressed the methodological role
or metatheoretical status of such evidence?

Dick Hudson wrote:

>Alexis Manaster-Ramer is of course quite right in complaining about
>the way we treat our theories. There was once a theory that fixed-
>subject idioms were impossible; the facts turn out to be otherwise, so
>someone should have said that the theory concerned had been refuted,
>and needed to be replaced. Like Alexis, I don't remember reading
>anything along these lines from those who have supported the theory in
>the past. But then, I suppose this is normal practice [...]

Do you mean that proponents of the asymmetry theory should have made a
public confession of error? Are we as a field supposed to care (except
as gossip) whether X, who once believed Y, has been convinced by Z
that Y is false? It seems to me that this sort of complaint suggests
that linguistics is about winning and losing arguments rather than about
the content of arguments. The matter would be different if, in the
present case, losing the subject/object asymmetry for idioms entailed
the incorrectness of some more general claim. Since there never was a
good reason for idioms to work this way (such that if they worked
differently, we'd need to change our ideas about things other than
idioms), the claim about idioms was never more than a weak example of
the sort of subject/object asymmetry a specific, more general theory of
theta roles might lead one to find. Obviously if someone uses disproved
claims as an assumption in other work, that person is not doing his job,
but I don't know of any examples of this type in this domain.

I also wonder about the facts. Do we know whether Marantz's claim that
alleged subject-object idioms are unaccusative in nature is wrong (e.g.
[the jury/i is [t/i out on...]]? If Marantz's suggestion is correct,
then nothing has been disproved. Can we yet exclude the possibility
that subjects may participate in idioms so long as (the head of) the
object is also part of the idiom (e.g. "on")? Can any collocation of
words be an idiom -- or are there constraints, perhaps more structured
than the literature has so far uncovered?

A preliminary: do we know an idiom when we see one (vs. a quote from
Shakespeare, a piece of a proverb or an allusion to a political speech),
i.e. is there any sense of what the object of study is? If not, there
still hasn't been a refutation of any theory, since the discussion has
been muddled from the start (which I rather suspect).

-David Pesetsky

Dick Hudson writes (vol-4-105, 2/17/93):

> I agree that it would be good to have a theory about why fixed-subject
> idioms (e.g. "the jury are out on X") are so much rarer than fixed-
> object ones (e.g. "X kick the bucket"). Does anyone know of any
> candidates?
>
> This question is presumably related to the claim that theta-roles are
> assigned to objects by V, but to subjects by V-bar, on the grounds
> that the choice of object is more likely to influence the subject's
> theta-role than vice versa (e.g. if I take a punishing then I'm
> affected/patient/theme or whatever, but if I take a book then I'm the
> agent). Does anyone know of any clear counter-examples to this claim?

I guess one could also say that most idioms (of the kind being
mentioned) are are more likely to be "comments" than anything else. And
since comments tend to become grammaticalized as VPs (or topics as
subjects), that *explains* why fixed object idioms are the norm.

Notice that the subject-fixed idiom "the jury are out on X" is also a
comment on X. But then again, that is not a counterexample or a big
surprise either.

--Jon

Jon Aske
Political Science/Anthropology Home address:
Bates College Jon Aske
Lewiston, Maine 04240, USA "Aritza Enea"
 12 Bardwell St.
Work phone: (207) 786-6472 Lewiston, Maine 04240-6336
Fax number: (207) 786-6123 -Phone: (207) 786-0589

e-mail: jaskeabacus.bates.edu, or
 jonaskegarnet.berkeley.edu (still!)