LINGUIST List 4.903

Tue 02 Nov 1993

Disc: Psycholinguistics

Editor for this issue: <>

Directory

1. Alec Marantz, Re: 4.902 Psycholinguistics: Graduate Linguistics at MIT
2. wendy sandler, Re: 4.902 Psycholinguistics
3. Vicki Fromkin, Re: 4.902 Psycholinguistics

Message 1: Re: 4.902 Psycholinguistics: Graduate Linguistics at MIT

On the topic of experimental methods in linguistics, I thought I might
remind the readers of some changes over the last few years in the PhD
program in Linguistics at MIT. A course in language acquisition is now
required of every student in the program. A quarter to a third of our
students are enrolled in a five-year joint program with the Department
of Brain and Cognitive Sciences. Students in this program do a year's
worth of course work in experimental cognitive science and are required
to be involved in serious experimental research projects. These
students, and others not formally enrolled in the special program,
integrate their experimental and "theoretical" research in papers for
standard linguistics courses. In addition, papers reporting results
from experiments are included among the readings for introductory syntax
courses.

-Alec Marantz

Message 2: Re: 4.902 Psycholinguistics

To follow up on Joseph Stemberger's reply to Mike Hammond, I think there
is another reason for the increasing interaction between psycholinguistics
and linguistics. After the long haitus following the downfall of the
Derivational Theory of Complexity, researchers with a psychological bent have
again begun exploring their questions within current theoretical linguistic
frameworks. This has the triple advantage of ensuring rigor in the
initial hypotheses, of posing them in a language understandable to
linguists, and of testing phenomena accounted for by system-internal
evidence with so-called external evidence. This is different from the
generally prevailing situation: 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 graduate student, I used to have this unsettling feeling that I was
a cog-psychologist when reading linguistics, but that I was definitely a
linguist when reading psycholinguistics. This situation is changing, for-
tunately -- primarily in cases where psycholinguistic research questions and
/or methodology address purely linguistic theoretical proposals. The work
of Joseph Stemberger and others in acquisition of phonology exemplifies this
trend, and so does the work of a number of sign language phonologists
and syntacticians in recent years.

Wendy Sandler

Message 3: Re: 4.902 Psycholinguistics

It seems to me that an even more important change is the acceptance of
linguistic approaches in neurolinguistics. The most recent issue of
Brain and Language is a special issue on Agrammatism and Linguistic Theory,
edited by Yosef Grodzinsky. At the TENNET (Theoretical and Experimental
Neuropsychology/Neuropsychologie Experimentale et Theoretique) meeting in
Montreal (May 29-31) there will be one whole day devoted to Linguistic
Explanations of agrammatism. And despite the period of dissillusion with
linguistic theory in psycholinguistics, linguistics is certainly thriving
as an approach to understanding not only language acquisition but language
processing.

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
single feature errors were reported (by me) such as 'Cedars of Lebanon'
becoming 'Cedars of Lemadon' which can not be explained or accounted for
unless one has a theory of phonological features. Even if we did not find
such features arising in speech errors one cannot conclude that they
don't exist in phonological representations in the grammar. There is
other evidence from linguistic analysis and historical change that they
do. It may just be the case that in mapping from the representation to
the production mechanisms we do not utilize features. The relationship between
the grammar and processing is a complex one. One can make a similar case
re morphological processes, syntactic structures etc.

This does not mean that psycholinguistic or neurolinguistic (aphasic)
evidence should not be used or is not important. It just means that
evidence is evidence -- linguistic data is as good (or bad) as any other kind
of data in testing of linguistic hypotheses.

Vicki Fromkin