Re: A talk with Michael Halliday

From: Jay Lemke (
Date: Tue Sep 21 2004 - 18:18:17 PDT

This is an interesting account of what must have been quite a good

We all know how difficult it is to articulate complex issues like these in
a dialogue, especially if we don't already share a lot of past history with
the person we're speaking to. I've certainly also had quite a few
conversations with Halliday about these topics, though not for about five
years now. He and I do share more than 20 years of on-and-off dialogue,

Halliday starts from a materialist dialectic view, and his approach to
language as social semiotic was built out of an original project for a
Marxist social linguistics. So we ought to expect many points of
convergence with Vygotsky. Halliday of course had also read LSV.

His view is indeed anti-dualist and anti-Cartesian, and he takes this so
far as to reject mentalism altogether (and with it much of the language of
cognitive science and cognitive linguistics) in favor of having just
meaning-making (social semiotics) and its material base (i.e. brains and
bodies and tools, but not separately "minds"). Our agreement about this was
the beginning of our collaboration (I gave a paper saying much the same
thing at a seminar he organized, not knowing it was also his view).

What is much harder to articulate is the relationship between meaning and
matter in this perspective. My version is in my book Textual Politics
(mainly chapter 5, adapted from a special issue of a journal that Halliday
was guest editor for). It has a lot in common with some of what David
reports about the relations of physical, biological, and social systems.

Halliday's developmental views are not ones that I as fully understand,
though I find them very interesting. I don't think that when he says the
child re-invents language that he means this in a Piagetian sense. He also
doesn't agree at all with the Chomsky-Pinker innatist views about language
development. I think (I may be wrong) that his view is more that the leap
to a fully semiotic use of language (the famous "double articulation":
sounds (signs) construed as wordings (linguistic forms) construed as
meanings -- i.e. semantic sense that is still specifically linguistic) is a
recapitulation or re-emergence of the historical-evolutionary leap, and
arises from the same embodied and social-interpersonal functional
communicative needs within an already-language-using community. I know that
the central force leading to this for him is the social-interactional
function of language use (and not, for instance, its designative or
denotative-informational functions). So this seems to me pretty close also
to LSV.

Perhaps the hope that Halliday expressed to David about physics someday
getting around to a meaning/matter conflation comes mostly from his
rejection of Cartesian dualism. Such a rejection does imply at least that
there is a specific material basis for the possibility of semiosis (e.g. a
simplest possible material system that could be said to make meanings or do
interpretations of signs). Depending on just how broad a definition of
semiosis you're willing to accept, that could possibly extend to a notion
of information at the quantum level. I know that Halliday has been
influenced by some of physicist David Bohm's efforts to push this idea.
It's not quite as bizarre as it seems, if we move beyond simplistic notions
of information to more sophisticated ones in which information is always
about something for someone (or for some other system that reads or uses or
interprets the information). That gets it much closer to "meaning" in
semiosis. On the other side, quantum theory has always rather fudged on the
issue of the role of the observer in determining whether a material system
carries definite information or not (basically what Bohr and Einstein
argued about for years). There is certainly a sense in which quantum theory
describes matter in informational terms. So a convergence such as Halliday
hopes for might just occur some day.

I don't think this is inconsistent with the view that matter and meaning
are united in socially meaningful action at the organizational level of
humans-in-social-material-environments, where "meaning" has its more
customary social-semiotic interpretation. I'm quite sure that Halliday
believes this too.

As the volumes of his collected works keep coming out, with a few
previously unpublished essays in many of the volumes, I hope we'll all get
a more complete picture of the development and latest versions of his
views. He's certainly not someone who believes there are simple or
definitive answers to big questions ... just many useful ways of talking
about them.


At 08:41 PM 9/16/2004, you wrote:
>David Kellog asked me to send this message on to the list. Some really
>interesting points raised.
>Last Saturday I splashed out on a plane ticket to go to Tokyo and hear Dr.
>Halliday speak, partly because I wanted to ask him about his agreements
>and his
>disagreements with socio-cultural theory.
>a) Both Vygotsky and Halliday believe in the necessity of a single, unifying
>theory of language and of learning. No more cognitivism in language but
>behaviorism in learning! Both of them believe that a steady state theory of
>language is no good at all; we need a genetic one. BUT....
>b) Halliday seems to think that children re-invent language, by discovering
>first how the "outer" world is projected onto the inner one through sense and
>then learning how to "project" the inner world onto the outer one. Vygotsky
>rejects this: childern do not "re-invent" language--instead the child's own
>line of development merges with an already developed socio-cultural one and
>both are transformed.
>c) Christie's idea of instructional registers which are "projections" of
>regulative registers (and classroom genres which are projections of
>extra-classroom ones) is clearly related (but curiously INVERTED).
>Vygotsky explicitly rejects this sort of thing in his criticisms of
>the chid's needs (defined as anything that motivates the child to action) are
>what "projects" the child's writing. Now, if Halliday really believes in a
>rather Piagetian vision of the child as unaided discoverer, wouldn't that
>him nearer Vygotsky than near Christie?
>There was a lot more, but I'm afraid I rather bungled it--I wasn't good at
>articulating all this stuff, and instead we got bogged down in a rather
>philosophical discussion. (I just got the photos of our discussion today, and
>we both look extremely tense!)
>Professor Halliday said that the world had two "realms" of phenomena:
>matter and
>meaning, and that they were incommensurable. Information was one kind of
>meaning, but only the quantifiable kind.
>This bothered me even more. Not only does it suggest a rather Popperian
>view of
>the world, it really seems to go against what Halliday himself has written.
>Halliday said that all biological systems are physical systems (but not, of
>course, vice versa). Similarly, all social systems are biological ones
>(but not
>vice versa) and all semiotic systems are social (but not vice versa).
>Assuming that "meaning" is the stuff that semiotic systems are made of (and
>"information" the stuff that physical systems are made of), this suggested to
>me that meaning was a kind of information, not vice versa.
>I know, it looks like a red herring, doesn't it? But back in Seoul, I
>wonder if
>the two problems really are unconnected. You see, it seems to me that the big
>problem with assuming a universe made up of meaning and matter on every level
>(besides the fact that it seems to lead to two, parallel, rather Cartesian
>universes) is that there's no place in this scheme of things for the thing
>really interests us: consciousness (or "cognition", as Lantolf would say).
>Maybe meaning is not a kind of information, but information that has been
>projected onto consciousness. In Tokyo, Dr. Halliday spoke of a belief that
>some day ultimate particles would be discovered that united meaning and matter
>(I think they already have in quantum mechanics, but I would say that they
>unite information and matter).
>But for Vygotsky the unity of meaning and matter takes place not at the
>elemental level. The "unit" of meaning appears at a higher level.
>The smallest unit for him is the word. Lantolf expressed this exquisitely when
>he was here in Korea by saying that a child is not a brain piloting a
>body, but
>a social being piloting both. A social being projecting itself onto both!
>David Kellogg
>Seoul National University of Education
>This message was sent using Inet-Webmail.

Jay Lemke
University of Michigan
School of Education
610 East University
Ann Arbor, MI 48109

Tel. 734-763-9276

This archive was generated by hypermail 2b29 : Tue Nov 09 2004 - 11:43:04 PST