LINGUIST List 3.309

Fri 03 Apr 1992

Disc: Reality of rules

Editor for this issue: <>


  1. Tom Wasow, types of evidence
  2. , The mind and the brain
  3. jj36, Re: 3.297 Reality of rules
  4. Avery Andrews, reality of rules

Message 1: types of evidence

Date: Mon, 30 Mar 92 10:23:59 PStypes of evidence
From: Tom Wasow <wasowRussell.Stanford.EDU>
Subject: types of evidence

Vicki Fromkin rebuts "those who do not think Chomsky wants
to consider any evidence except intuitions of the native
speaker" with a quote ending with the following sentence:
"It just seems absurd to restrict linguistics to the study
of introspective judgments." It would be very hard to

Yet, for decades it has been the practice of most generative
grammarians to restrict their data in just this way. Other
sources of evidence are cited only when they seem to support
analyses independently arrived at on the basis of
introspective judgements. (A striking example is a 1968
_Psychology Today_ article, in which Chomsky approvingly
discusses the experiments based on derivational theory of
complexity, suggesting they support the then-current version
of transformational grammar).

In _Knowledge of Language_, Chomsky wrote, "In practice, we
tend to operate on the assumption, or pretense, that these
informant judgements give us 'direct evidence' as to the
structure of the I-language..." (p. 36). On the next page,
he expressed the "hope that such evidence will eventually
lose its uniquely privileged status."

I share that hope. In fact, I think the time has come to
expect hypotheses in grammatical theory to be supported by a
variety of different types of data.

Tom Wasow
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: The mind and the brain

Date: Tue, 31 Mar 92 19:15:11 BSThe mind and the brain
From: <>
Subject: The mind and the brain

Stavros Macrakis (in Linguist 3.297) drew an analogy of the mind as function,
while the brain plays the role of a "thing". The particular analogy employed was
that of a passenger-carrying road vehicle (mind) and, say, a car (brain). The
fault is that while the mind "grows" with time and usage, the
passenger-carrying road vehicles in my garage do not (it is only that "I" get
better at using them). If monists see the mind and brain as one, they must have
curious views of intelligent women and prodigious children.
 I am, of course, content that the mind should be seen as the term for the
functioning of the brain ("of the brain", as I suggested earlier, and not "in
the brain" - I hope we have got rid of that misapprehension) and that is what I
take Saussure to have meant (in the quotations which I cited in Linguist
 But I insist that it is the difference between the functioning and the
"thing" which is the proper study of cognitive sciences (amongst which I like
to think that linguistics may be numbered).
 Philip Swann (in Linguist 3.297) wrote exactly what I might have been
writing here at greater length. I thank him for the pertinence of his posting.
 Stavros Macrakis invited me to tell what a monist misses out on. I will not
dwell on the sadness of misunderstanding intelligent women. But surely the view
of the unitary mind-brain allows for no explanation of intelligence (for
language, let us say). I cannot see any way in which the monist can escape the
herculean labours of observation, with none of the non-beneficial excitement
even of description. Bill Bennett
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 3: Re: 3.297 Reality of rules

Date: Tue, 31 Mar 92 22:50 EST
From: jj36 <>
Subject: Re: 3.297 Reality of rules

 Bill Bennett is under the mistaken impression that in my earlier
posting [3.276/1] I "argued that 'langue' was for Saussure a stabilizing
factor". I actually said: "As I read Saussure, the social nature of
language accounts for its stability, not its instability... it is the
arbitrariness of language that accounts for its variability and
changeability, and...this is always in tension with the stability inherent
to language by virtue of its being a social fact". The point is not that
langue is a stabilizing factor, but that it displays a tension between
variability and stability, and its status as a "fait social" is connected
with the latter rather than the former (contrary to what Martti Nyman had
 The passage I cited in support of this view is not an isolated one,
but representative of many such statements in the _Cours_ and its source
materials. Here is another, taken from the final course given by Saussure
(CLG/E III C 317, 1232-33): "La circonstance que la langue est un fait
social lui cree un centre de gravite. Mais nous avons admis des le debut ce
fait... Les forces sociales agissent en fonction du temps et nous montrent
en quoi la langue n'est pas _libre_. En effet la langue est <tout le temps>
solidaire du passe, c'est ce qui lui ote sa liberte, et elle ne le serait
pas, si elle n'etait pas sociale" ("The circumstance that langue is a social
fact creates a center of gravity for it. But we have admitted this fact
from the start... The social forces act as a function of time and show us
wherein langue is not _free_. In effect langue is <always> in solidarity
with the past -- that is what takes away its freedom -- and it would not be
so if it were not social".)
 Bennett also raises an interesting question about Saussure's
"antinomy between a 'force du clocher...agissant comme principe divisant'
('parish influence...acting as divisive principle') and 'intercourse...force
unifiante' ('general communication...unifying influence')." The antinomy is
raised toward the end of the part of the _Cours_ on geographical
linguistics; clearly Saussure did not see it as directly connected
with his characterization of langue as a social fact, since the two are
never mentioned in the same context. Indeed, to suggest that they be
considered in tandem risks ignoring the separation between "internal" and
"external" linguistics made at the start of the geographical section. Langue as
"fait social" belongs to the former realm, clocher/intercourse to the latter.
 The "esprit de clocher" and "intercourse" (a borrowing from English)
are cited by Saussure as general factors in human behavior to which "the
propagation of facts of langue is submitted". Behavior is subject to
conflicting social centers of gravity, local vs. national. In a village or
canton, local social bonds are very powerful, but cannot become absolute
because of the needs of intercourse with other neighboring communities. Far
from implying that the social nature of langue explains its variability and
changeability, the clocher/intercourse antinomy means that langue will
SPEAKERS, both local and national. This is the point of the final statement
on the subject in the _Cours_ (p.285): "...on peut tout ramener a la seule
force unifiante sans faire intervenir l'esprit de clocher, celui-ci n'etant
pas autre chose que la force d'intercourse propre a chaque region" ("one can
trace everything back to the unifying force alone, without bringing in the
esprit de clocher, since the latter is nothing other than the force of
intercourse proper to each region").
 The rest of Bennett's ideas about Saussure I find, with all due
respect, quite incredible. I'll limit my comments to 6 points on which I
strongly take issue:
 (1) "faculte de langage" (defined by Saussure as "langue in the
individual") is not the same as "langage";
 (2) "langue" cannot be equated with Chomsky's "competence"; doing so
leads Bennett into some major anachronisms (see further my chapter
"Ideologizing Saussure: Bloomfield's and Chomsky's Readings of the _Cours_"
in _Ideologies of Language_, ed. Joseph & Taylor, Routledge 1990);
 (3) the statement that langue "was defined by Saussure as a 'fait
social' only as far as it was the trigger of the universal potential for
language" has no textual support, nor is it supported by Bennett's following
comments, which err on point (1) above and (4) below;
 (4) the quote from Saussure that "'langue' and 'langage' are but one
single thing..." dates from 1891, well before he had worked out any details
of his system, and fully 16 years before the second sentence with which
Bennett joins it as a single statement;
 (5) the idea that "Saussure was thinking of mind" has no obvious
support; and to say "It is beyond doubt that he meant [langage] to apply to
the mind and not to the brain" is an exaggeration, inasmuch as
references to the brain are much more common in Saussure than those to the
mind, which had undesirable metaphysical connotations in his positivistic day;
 (6) the plan of the third course cited by Bennett omits the
definition of langue which directly follows it -- "produit social,
institution semiologique" -- and which severely limits what one might
otherwise be tempted to read into the characterization of langue as "what is
--John E. Joseph, Dept. of French & Ital., Univ. of Maryland, College Park
MD 20742 USA
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 4: reality of rules

Date: Thu, 2 Apr 92 11:40:23 PSTreality of rules
From: Avery Andrews <andrewsCsli.Stanford.EDU>
Subject: reality of rules

Re Helge Dyvik (linguist 3.297):

>Grammars do not describe what *must* happen. It is possible to choose not
>to conform to the norms for syntactic well-formedness described by a
>grammar. Since it is possible to know (in the sense of being able to
>follow) a grammatical rule without actually following it, the fact that
>someone knows a rule cannot by itself explain the fact that he *does*
>follow it.

I don't see the relevance of this, other than that it illustrates
the fact that we don't know much about how whatever it is that
grammatical rules are describing is integrated into the organization
of behavior. E.g. how is it than when we just talk, we mostly
follow them, but we can put verbs at the end, etc., if we want to.
This is all quite mysterious, but it doesn't bear on the existence
of rules, but how they are integrated into everything else.

>I suppose the crucial question is: is it plausible to assume that whatever
>causal mechanisms in there enable us to conform to the grammatical norms
>must have the same structure as a grammarian's description of the norms

In principle, perhaps not, but its hard to see any concrete
alternative for the basic generalizations of syntax. How could there
be all the evidence there is for noun phrases if there wasn't some
single facet of mental structure responsible for the (Det) Adj* N PP*
etc. sequences showing up all over the place? How there be all the
evidence there is for (some kind of) Wh-movement (a wide range
of constituent types appearing in a normal version, and another
differing for the most part only in the presence of a hole), if
there wasn't some sort of rule or principle capable of making holes?

Although I strongly advocate the reality of rules (given due caution
in cases where good alternative explanations for generalizations
exist, such as historical ones), I would not of course argue that
current linguistic theories are particularly accurate - they have
enough problems on `strictly linguistic' grounds so that I wouldn't
want to take them as providing more than the vaguest indications
of what is going on inside the head, and even the best grammatical
theory will surely provide only very partial evidence about the
mental structures underlying language use. But current ones to
provide some, and better ones will provide more.
 (currently also
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue