LINGUIST List 2.49

Saturday, 23 Feb 1991

Disc: Cognitive Linguistics

Editor for this issue: <>


  1. , Lakoff's message of Feb 7, 91
  2. David Powers, Re: Cognitive Linguistics
  3. , Brand Names for Theoretical Frameworks
  4. Vicki Fromkin, Re: Cognitive Linguistics
  5. George Lakoff, Reply to Professor Fromkin

Message 1: Lakoff's message of Feb 7, 91

Date: Sat, 23 Feb 91 13:35:20 WAT
From: <>
Subject: Lakoff's message of Feb 7, 91
Because of a gateway problem, a small but substantial number of
people did not receive a posting from George Lakoff on February 7.
Since this submission stimulated considerable discussion, we are reposting 
it as item 5 of this issue. 
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: Re: Cognitive Linguistics

Date: Wed, 20 Feb 91 16:08:17 MET
From: David Powers <>
Subject: Re: Cognitive Linguistics
I have received three LINGUIST postings on COGNITIVE LINGUISTICS,
from Gibbs, Fromkin and Everett. But I seem to be missing a comment
from Lakoff which triggered that last. Nonetheless, I now cannot let
some of those comments pass.

I read the original posting - having never heard of Gibbs before - thinking
"Good, at last a journal which will allow the interactions between
linguistics and the other cognitive sciences to be thoroughly
explored." I completely missed any overtone which might have
explained the vehemence of Fromkin's response - even after
glancing through the original piece again. And now, after this posting,
a more careful look picks up "not as if they were autonomous", and
I begin to see a glimmer of light.

I would think Cognitive Linguistics would mean precisely the
examination of language in the context of general cognition.
I would think it wrong to pursue one's research program with an a
priori assumption that language mechanisms have nothing to do with
other mechanisms, but that the argument for such modularity should
come through examination of the relationships between language and
other cognitive phenomena, and that proposed mechanisms should be
considered for applicability/application to other phenomena.

This would increase our understanding of language and cognition
irrespective of the positive or negative outcome.

It is here that I take particular exception to what amounts to
the citation of Chomsky as the father of the Cognitive Sciences.
Chomsky was ONE of the first to seek to apply mathematical rigour
to Linguistics, AND to make some proposals for how we might proceed
from here. Chomsky prefered NOT to consider what neurological and
evolutionary mechanisms could explain his mechanism, but these
proposals became very influential in Psycholinguistics as well as
Linguistics. Thus Chomsky did provide a stimulant. Whether he always
stimulated in the correct direction in every dimension is another
question. I personally think his claims were far too strong, far
too restrictive and far too tangential.

Scientists use theory to guide experiments. Chomsky's work
stimulated the discovery and refinement of linguistic universals
and developmental studies. There was a backlash from
Psycholinguistics towards the end of the 70s. And Linguistics too
forced a weaker reformulation of the original dogma.

What was unscientific, and irreconcilable with the label Cognitive
Science, was NOT so much the logical step "language learning is
impossible therefore we don't learn language" with its flawed
presuppositions, NOR the subsequent non-sequiturs leading to the
idea of a specialized innate language organ, BUT the absence of any
attempt to make predictions which would distinguish between this
organ and general cognitive capabilities. The establishment of
language universals does NOT fit this bill. And Chomsky largely
used this language organ as a marker of the end of his interests - he
has repeatedly reiterated that he has no interest in how its behaviour
actually arises, let alone how the "organ" itself arose.

Even today the "modularity" tenet of faith denies, as a matter of
faith, the relevance of other cognitive capabilities to language
per se. This belief, however, need not have a great impact on the
course of cognitive research. Cognitive Science still needs to
explain how it works. Appeal to a homuncular organ does not change
this. Moreover, new results need have no effect on the "tenet of
faith". As more and more phenomena (which I would have called
linguistic) are explained, the "module" will continue to be redefined
in terms of the "truly" linguistic behaviour for which it is
specialized. Is this not precisely what we have seen so far in the
history of TGG? Where does Cognition or Science come into this?
It is Semantics!

Whilst I do not wish to deny that he has made a significant
contribution, I tend to feel that Chomsky has in some ways led
us AWAY from Cognitive Science. And the massive acceptance of his
tenets has tended to mean that the minority get laughed out of court
without being heard. But I see certain other work as a start towards
computationally viable and psycholinguistically faithful theories
which explain learning of the "unlearnable".

I always thought Science meant openness to have a theory proved or
disproved, and the avoidence of long strings of tendencious
argument. I summarize the argument as I see it:

1. Poverty of the Stimulus (including lack of negative information)
2. Results in Complexity Theory (impossibility of learning without such info)
3. No learning
4. Specification of a more general innate Universal Grammar
5. No language except in humans
6. Acquisition by a species and modality specific innate Language Module

1, 2 and 5 have the status of axioms based on (empirical and
theoretical) results.

1. should take into account non-linguistic and implicit criticism;
2. should take into account restrictions on human cognition;
3. should take into account the development of language/organ;
4. should take into account the natural correlates of UG rules;
5. should take into account general cognitive differences between species;
6. separates language from general cognition and hence cognitive linguistics.

All of this says nothing about Universal Grammar. Obviously
something is innate, and is responsible for our ability to learn
language as well as for the similarities amongst human languages and
creoles. But all we can conclude is that we have cognitive
mechanisms which allow this, and that this is one area in which the
cognitive architecture of other species is insufficient.

What is wrong with a journal looking at Cognition and Linguistics
together? Particularly when the group "excluded" has the majority
of the other journals in the area available to them, and has
deliberately occupied an "exclusivist" position that states that
"language has nothing to do with general cognitive mechanisms".

I see every reason to start a new journal to support another
position - a minority position which needs support. I can do
without "impartial" referees who say "Have you heard of Chomsky?"
or "When I got to this [denial of Chomskian tenets] I stopped
reading?". That's hardly an argument! One can't win: Ignore the
opposition and they think you're ignorant, argue against the majority
view and they think you're stupid!

Perhaps we are wrong, who think that there is a role for learning.
But I hope that "Cognitive Linguistics" turns the other cheek and
welcomes constructive criticism and coherent argument of other

David Powers
(Computer Scientist)

Disclaimer: The university, as buildings, has no opinions, as a community,
millions; as an institution, none yet; and as me, ... perhaps this.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 3: Brand Names for Theoretical Frameworks

Date: Wed, 20 Feb 91 17:39 MDT
Subject: Brand Names for Theoretical Frameworks
 This list has seen two recent postings expressing outrage at
the *name* ("Cognitive Linguistics") of a theoretical framework (or
set of related frameworks). The first time I took to be an idiosyncrasy,
but the recurrence alternately bothers and amuses me. I thought this
list was a forum for more substantive discussions than grousing about
what someone else's name is.

 The *mode* of objection is curious. Its logic seems to be:
 1. The name A of some approach implies they study B;
 2. Other (perhaps contrary) approaches also study B;
 3. Therefore use of name A is a usurpation and derogation of others
who study B.

 On its face, this seems plausible, but in light of the past
80 years or so of the history of linguistics, it is a strange
turnabout in standards of naming. Examples of "violations" of the same
sort would have to include "Structuralism", "Functionalism",
"Transformational Grammar", "Relational Grammar", "Lexical-Functional
Grammar", "Government-Binding Theory", and many many more. Note that
it would be circular to claim exemption by assigning the pseudo-
descriptive label the particular technical interpretation that
practitioners of the approach so named wish it to have -- the same
defense would suffice, as previous posters have indeed indicated, for
"Cognitive Linguistics".

 At best, such pseudo-descriptive brand names indicate that the
approaches so named give (or at least see themselves as giving?)
greater centrality to something their names indicate than do other/
most contemporaries. By that criterion, "Cognitive Linguistics" (in
the broad construal indicated in the charter of the ICLA, for example)
is in the tradition, since the approaches using that name give greater
prominence to (general) cognition in linguistics.

 I hasten to add that the long history of libertarian naming
of theoretical approaches includes acceptance of more evaluative labels
such as "Natural...", "Standard Theory" (and its etymological heirs),
etc. Not to mention, by the obsessively picayune sensibilities
recently shown, such out-and-out misnomers as "Generative Phonology"
(since most practitioners regard phonology as *interpretive*). About
the *only* names that might not be objected to on the grounds
recently unearthed would those of geographic or personal origins (e.g.
"Prague School" or "Bloomfieldian"), though even these are usually
based on stereotypes; or those derived from a technical term that has
no homonyms in other approaches (e.g. "Tagmemics").

 So why not let's get back to doing some *linguistics* instead
of prescriptive metametalinguistics?
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 4: Re: Cognitive Linguistics

Date: Wed, 20 Feb 91 21:45 PST
From: Vicki Fromkin <>
Subject: Re: Cognitive Linguistics
TO: Daniel L. Everett -- How nice to know there is someone else out there
who agrees with some of us out here. I hope you read my reply to
the Cog/Ling announcement. Let's keep in touch. VGicki Fromkin
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 5: Reply to Professor Fromkin

Date: Thu, 7 Feb 91 03:34:47 -0800
From: George Lakoff <>
Subject: Reply to Professor Fromkin
From: George Lakoff <>

Re: Reply to Professor Fromkin

As President of the ICLA, I'd like to take this opportunity
to reply to Professor Fromkin's note. 

Professor Fromkin raises an important issue: Exactly what evidence is
there to support the generativists' contention that language is
autonomous? The field of cognitive linguistics has developed over the
past decade and a half in response to massive evidence that language is 
anything but autonomous; rather it is, in very signifant ways,
a product of of general cognitive mechanisms of a variety of types.
The literature in the field as a whole supports this view.
Generative linguists tend not to be conversant with cognitive
linguistics literature, and perhaps it would be a good thing if
the whole matter were taken up in this forum --
preferably in a systematic way,
rather than just throwing bibliographies at one another.
That is what I tried to do, in small measure, in my book
WOMEN, FIRE, AND DANGEROUS THINGS, which surveys some of the
relevant evidence. But there is no lack of other things to
read in this field.

As for the opinions of those doing brain research, there is
no lack of research pointing in the anti-modularity direction.
A good place to get details would be from the UCSD
Cognitive Science group, in particular, Elizabeth Bates,
Marty Sereno, Marta Kutas, and Rob Kluender.

Perhaps this is the only forum where such a discussion
could take place across the cognitivist-generativist
divide. It is impossible at the LSA, which
has a conservative, generatively-oriented program committee
and which has refused to permit paper sessions devoted to
results in cognitive linguistics. The International
Cognitive Linguistics Association, which is only one year old,
was formed partly because there was no other general
forum for the discussion of these results. The Association,
and its new journal, Cognitive Linguistics, has been extremely
successful and we are grateful to the Summer Institute
at UC Santa Cruz for hosting our conference (though
it is unfortunate that no courses at all in this field are
being offered there, just as none have been offered
at other LSA Summer Institutes). One of the nice things
about this means of communication is that it is open,
and openness of communication is sorely needed in a
field as conservative as linguistics. Other professional organizations
have been far more receptive to cognitive linguistics as a discipline,
and cognitive linguists are regularly invited to address
major meetings in Cognitive Science, Psychology, 
Computer Science, Anthropology, etc. 
And the literature in Cognitive Linguistics itself is
growing so fast that it is virtually impossible to keep up with
all of it. 

To the current generation of linguistics students,
I recommend subscribing to our journal, Cognitive Linguistics, 
by joining the society at the bargain rate of $18.
Just send a check made out to the International
Cognitive Linguistics Association (ICLA) to

Eugene Casad
P.O. Box 8987 CRB
Tucson, AZ 85738

Nonstudent memberships are $55.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue