LINGUIST List 4.1013

Tue 30 Nov 1993

Disc: Linguistics as Psychology

Editor for this issue: <>


Directory

  1. Jack, Linguistics as Psychology
  2. Esa Itkonen, (psycho-)linguistics

Message 1: Linguistics as Psychology

Date: Thu, 25 Nov 93 20:11:31 ESLinguistics as Psychology
From: Jack <JAREAukcc.uky.edu>
Subject: Linguistics as Psychology

The discussion of Bloomfield versus cognitive psychology is wildly
anachronistic. M Haspelmeth has suggested, "The only anti-cognitive
anti psychological school was Bloomfield and the post-Bloomfieldians."
And R. Wojcik has added, "Altho the post-Bloomfieldians represented a
rather extreme anti-cognitive position...."

It is my suggestion that these people should first make themselves at home
in the principal psychological enterprises of the 1930's before passing
judgment on Bloomfield versus psychologists. Had Bloomfield decided to
associate his linguistuc approaches with any dominant, accepted approach
to "psychology" in his day, rather than sticking essentially to his last,
we might find strong reason to take issue with his "psycholinguistics".
Indeed we must remember that (aside from psycho-analytic approaches of
one school and another) the "standard" view was a rather strictly
behavioristic one, where S-O-R was the shibboleth, and any attempt to
pry into the presumed inner workings of the "mind" while discussing
human behavior was eschewed (above paranthesis excepted). It is silly
to castigate important workers of the past for not basing their work
on the theoretical approaches of this moment: and who knows where they
will be in tomorrowns ten thousand years! An easy start can be made by
looking up "cognitive psychology" (or even just "cognitive" in a desk
dictionary of the early forties.

 Kentu konkaza, kentu beritaza!
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: (psycho-)linguistics

Date: Tue, 30 Nov 1993 14:39:22 (psycho-)linguistics
From: Esa Itkonen <EITKONENsara.cc.utu.fi>
Subject: (psycho-)linguistics

Vicki Fromkin suggests (21 NOV) that it does not really matter whether
or not linguistics is taken to be a branch of psychology. But of course
it matters. Let's start by quoting Kiparsky 1993: "Modern lingustics
acknowledges Panini's grammar as the most complete generative grammar
of any language yet written, and continues to adopt technical ideas
from it." However, Panini's grammar deserves such high praise only on
condition that it is considered as a NON-psychological undertaking
(a point which Kiparsky agreed to when I raised it at UCLA, May 1982).
If Panini's grammar is considered as a psychological undertaking, i.e.
as striving after psychological reality, then it is immediately
falsified on a number of points (e.g. certain underlying forms in
morphology and the strict rule-ordering in phonology). Similar comments
apply, even more forcefully, to the Montague grammar. Now, it clearly
MATTERS whether a theory is true or false. Therefore it also matters,
contra Fromkin, whether or not linguistics is considered as just
a branch of psychology.

Anjum Saleemi claims (10 NOV) that "all evidence is external" (i.e.
homogeneous). This is not true. A formal logician who constructs e.g.
axiomatizations of deontic logic relies ONLY on his (logical) intuition,
which means that he NEVER uses experimental evidence.While it is
conceivable that one could try to practise psychology of logic without
experimental evidence, it would be unreasonable to do so. (Thus both
internal and external evidence is needed; it is never possible to
use external evidence only.) And in any case, the resulting descriptions
are quite different (witness Aristotle - or modern predicate logic - on
syllogistics vs. Johnson-Laird on 'mental models' for syllogistic
reasoning). Therefore the evidence is not homogeneous. I claim that
similar comments apply to the linguistics/psycholinguistics divide
(recall Panini & Montague).

I agree with Richard Wojcik (19 NOV) that Trubetzkoy should not be
considered a cognitivist. In this respect he was more consistent
than e.g. Saussure or Sapir. According to Saussure, 'langue' qua
system of linguistic signs is social; and yet linguistic signs
themselves were defined in psychological terms. For Sapir, we should
take language "as an institutional or cultural entity, leaving organic
and psychological mechanisms back of it as something to be taken for
granted" ('Language', p. 11); and yet he was ALSO striving after
'psychological reality', i.e. did NOT take the psychological mechanisms
for granted (p. 33). As I said before, Trubetzkoy was more consistent.
A Finnish phonetician Arvo Sotavalta had suggested that it might be
possible to 'ascend' from particular occurrences of speech (= 'parole',
'Sprechakt') to the linguistic system (= 'langue', 'Sprachgebilde'). To
this Trubetzkoy replied ('Grundzuge', pp. 15-16) that this kind of
transition from particular to general, which is the method of the natural
sciences, is not possible within linguistics, because linguistics is
not a natural science. The linguistic system (= 'langue') is a social
(and normative) institution, and an institution necessarily precedes,
and is presupposed by, any particular actions performed within it.
Trubetzkoy was right, of course. (To be sure, the institutional aspect
can be complemented, but not replaced, by the psychological aspect;
cf. above the impossibility of an exclusive use of external evidence.)

It is interesting to note that more recently Bromberger and Halle (in
'The ontology of phonology', 1992) have tried to reimplement Sotavalta's
program, in the following way. First, a generative phonological analysis
is performed, i.e. an analysis based on nothing but intuitive evidence
of 'our language'. (Notice how the reference to [the institution of]
'our language' takes for granted the easy identifiability of what is or
is not a correct phonological sequence.) Second, this analysis is
'psychologized', by postulating a 'competence' (or 'I-language') identical
with it. Third, the 'competence-type' analysis is made into an analysis
of 'real-time processes', by postulating two types of intentions (i.e.
word intentions and phonetic intentions) behind the descriptive units.
All this presumably contains no reference to anything else than particular
occurrences of speech, i.e. it is pure natural science.

This is a nice example of what many of us regard as the pseudo-psychological
attitude pervading the generative enterprise. That is, if you really
want to construct a psychological theory of phonology (or of syntax),
you have to use experimental evidence, and you are bound to come up with
results that you could have not reached on the basis of (conscious)
intuition alone. For instance, Kari Suomi (anothr Finnish phonetician,
by the way) uses experimental evidence to construct a model where production
and recognition deal with distinct types of units (phonemes and syllables
vs. word-size gestalts); and assuming that this difference is psycholo-
gically real, any model of real-time processes - and even any realistic
model of competence - has to accommodate it (Journal of Phonetics, 1993).

The existence of 'make-believe psychology' within linguistics, as well
as the distinction between (autonomous) linguistics and psycholinguistics,
was well known and well-established by the end of the seventies. Anjum
Saleemi (10 NOV) disagrees with this assessment. According to him, "no
wheel was invented" during the decade 1973-1983. I claim that Saleemi is
wrong. The wheel was invented, but Saleemi just does not know it. It is
conceivable, of course, that I am wrong. But this has to be proved.
In order to prove me wrong, Saleemi cannot just rely on hear-say, but
has actually to read the relevant literature. This may sound simple,
but it isn't. It has turned out that, aspart of their biological endow-
ment, some members of the field are innately predisposed not to read
anything writen by the 'bad guys'. (I am referring to an expression
that was in current use at MIT in the end of the sixties when I spent
one academic year at the said institution.)

Esa Itkonen
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue