LINGUIST List 4.920

Fri 05 Nov 1993

Disc: Psycholinguistics

Editor for this issue: <>

Directory

1. Vicki Fromkin, Re: 4.913 Psycholinguistics
2. Bill Bennett, Re: [4.913 Psycholinguistics]
3. Sam Wang, Re: 4.903 Psycholinguistics

Message 1: Re: 4.913 Psycholinguistics

re the deabate or rather discussion on psychlinguistics vis a vis linguistics.
The question is asked whether, since according to 'some' (since Chomsky
put the mind back into the brain and language in the mind) linguitics
may be a sub branch of psychology, why has it taken so long to accept
psycholinguistic evidence. I think this reflects the 'definition' of
psycholinguistics as having to do with processing rather than representation.
Grammaticality judgements are of course performance judgements but
must in a very direct way reflect stored knowledge or I-Language. Experimental
 results may or may not reflect such knowledge and it is important to
understand what other factors may be involved in obtaining such data,
including short term memory, attentional aspects, real-time processing
factors, etc etc etc. But we are beginning to see interesting experiments
that are able to pull apart these different aspects of linguistic performance.

What has been a subject for debate however is whether linguistic theory
based on what Zwicky once called 'internal' evidence is as 'good' as
'external' evidence such as psycholinguistic experimental results, or aphasia
breakdowns or speech errors or.... Some of this evidence clearly argues
in favor of one rather than another hypothesis in linguistics. For example,
at the recent Academy of Aphasia meeting, Eleanor Saffran reported on a
case of a brain damaged patient who has a clear dissociation between semantics
and syntax -- is unable to look at a picture of a cow or a horse and
tell you what it is, or anything about a cow or horse. But if shown
two pictures, one where a horse is kicking a cow and another where the
cow is kicking the horse, can use syntax to point to the correct animal,
i.e. if presented with the sentence "The cow is kicked by the horse" and
asked to point to the 'cow' or the 'horse' will do this 100% correctly.
(other such syntactic structures were also presented). Thus any theory of
grammar which does not separate syntax from semantics is unable to
account for such data. On the other hand one can present equally valid
evidence for this position based on linguistic evidence.

I think Chomsky argues for the equal status of data very well in 'On the biolo-
gical basis of lg capacities' in Miller and Lenneberg's 1978 book:

"Suppose that someone were to discover a certain pattern of electrical
activity in the brain that correlated in clear cases with the presence
of wh-clauses, relative clauses (finite and infinitival) and wh-questions (dir-
ect and indirect). Suppose that this pattern of electrical activity is
observed when a person speaks or understand a particular sentence. Would
we now have evidence for the psychological reality of the postulated
mental representations? We would now have a new kind of 'evidence' but
I see no merit to the contention that this new evidence bears...reality
whereas the old evidence only relates to hypothetical constructions. The new
evidence maight or might not be more persuasive than the old; that would
depend on its character and reliability, the degree to which the prinnciples
dealing with this evidence are tenable, intelligible, compelling and so on."

Vicki Fromkin

Please: what exactly is linguistics without a psycho-/socio- component?

Bill Bennett.
du

Message 3: Re: 4.903 Psycholinguistics

> From: wendy sandler <RHLE702UVM.HAIFA.AC.IL>
> a lot of psycholinguistic work tests
> hypotheses about language performance ("processing"), rather
> than using performance to test proposals about fundamental linguistic
> structure (competence).
>
As a matter of fact, quite a lot of experimental work has been done
on some linguistic assumptions, e.g. the vowel shift rules proposed
by Chomsky and Halle. It has been shown time and again that
the vowel alternation phenomena do not work as suggested in SPE,
nor in any formal, featural accounts. But as late as 1985, Halle
and Mohanan still pursue this line of work, without mentioning
any single experimental result that has been done on this topic.
And in Durand's textbook 'Generative and non-linear phonology' (1990)
he gives up the experimental counter-evidence of vowel shift rule
and opt for a systematic 'package deal' (p.152). Is system all that
matters? If all this experimental evidence cannot count, what can?

> From Vicky Fromkin:
>
> I don't think there is a problem re accepting experimental (or other
> real-time production/perception data such as speech errors) when such
> supports particular linguistics hypotheses and ignoring them when they don't
> if one accepts the separation between representation (competence) and
> processing (performance). For example, many years ago , examples of

Alas, the division between competence and perfromance is an easy escape
from counter-evidence in experimental results. One will of course
accept experimental evidence when it supports linguistic hypothesis,
but if one ignores the same type of evidence on the ground of
the division between competence and performance just because
the evidence does not accord with the hypothesis, that's totally
irresponsible. I think only linguists will do that. Of course
one can question the validity of the experimental methods, but
shouldn't one also have to examine the validity of the theoretical
hypotheses as well?

..............................................................................
. H. Samuel Wang . EMAIL: onghiokling.nthu.edu.tw .
. Department of Foreign Languages . TEL: 886 35 715131 ext 4398 .
. National Tsing Hua University . FAX: 885 35 718977; 886 35 725994 .
. Hsin-chu, Taiwan . .
..............................................................................