LINGUIST List 3.235

Mon 09 Mar 1992

Disc: The Psychological Reality of Rules

Editor for this issue: <>

Directory

1. , 3.231 Are Rules Psychologically Real?
2. , Re: [3.231 Are Rules Psychologically Real?]
3. CHARLES LAUGHLIN, ON THE EXISTENCE OF RULES

I second Jon Aske's idea that an open discussion of the
question of how we establish the existence of rules would
be ideally suited to the LINGUIST format (or vice versa).

I would only like to suggest that we consider more possibilities
than (a) yes, there is a rule and it is in the usual "generative"
format, and (b) no, there is no rule, all the attested forms
are "in the lexicon".

Thus, if we decide, as Jon suggests, that something
(say, Spanish stresses or plurals in -es or the feminines that take
'el' instead of 'la'), that all by itself unfortunately
does not tell us whether speakers will access this information
when dealing with novel forms or not. For example, the Polish
verb meaning 'can, be able to' has no imperative, but I have
found that some speakers simply cannot form one no matter
how hard they try, others (like me) had to consciously find
a rhyming verb that does have an imperative and use simply
analogy, while yet others were not aware of how they knew
the form, they just knew it. It could be that the first and
second types both have identical lexicons, but differ on
how readily they access what's there, whereas the third group
are using a rule. But it could also be that all three have
no rule, and that even the third group used analogy to what
is in the lexicon (only did so with the conscious effort that
group two needed).

Indeed, one could have a situation in which analogy to
what is in the lexicon would always give the same results
as a system of rules.

I would also suggest that we adopt the policy "Write your every rule
like it was your first one", i.e., do not assume that something
is a rule because somebody has argued previously for a similar
rule in, say, a different language or dialect.

Finally, may I urge the importance of the point I made in an
earlier posting, namely, that phonology and morphology, no less
than syntax and semantics, must deal not merely with closed
corpora but with productivity and creativity in language use,
yet it seems obvious that an enormous chunk of what has been
written about phonology and morphology does rely on closed
corpora (of course, there are notable--and noble--exceptions,
but they ARE exceptions).

Jon Aske ia right to expect no less than a fitting MODEL of the
speaker-hearer's mental activity. But we need to sort out "real" and "the
pattern" first, don't we?

I am delighted that Jon Aske raises the question about the psychological
reality of rules in generative acccounts of language. We anthropologists
ran through this issue in the '70s and early '80s re Levi-Straussian
structuralism. The issue remains the same, and several answers are the
same. Let me briefly tic them off:

1. When the rules are determined by solely deductive methods from
observables, then there is no way within the confines of those
methods to determine whether the rule "structures" are real or not.
This is because more than one deduced structure may account for the
observables. Some independent method must be used to evaluate the
reality of the "structures."

2. Rule structures deduced from observables often bear no clear
relationship to actual physiological (neurophysiological) structures.
There is no LAD in the human brain.

3. Any explanation of "errors" in utterances based on failure to
apply rules that are presumed to be "in the speakers head" violates
obvious rules of logic -- guilty of both tautology and post hoc
fallacies.

There are other problems with the kind of methods linguists use to
generate accounts of language and adduce evidence in support of
those accounts, but these are the main ones. Generative accounts
leave themselves open to charges of anti-empiricism and
epiphenomenalism. My own bias is that any approach to the
explanation of the productions of the nervous system that do not
AT LEAST IN PRINCIPLE leave themselves open to disconfermation
via the neurosciences is increasingly obsolete.

Thanks for bringing up the question, Jon!

Charles Laughlin
Department of Sociology and Anthropology
Carleton University
Ottawa, CANADA K1S 5B6

Charles Laughlin <CHARLESLCARLETON.CA>