LINGUIST List 5.477

Mon 25 Apr 1994

Disc: Cognitive grammar

Editor for this issue: <>


  1. Bill Croft, Do cognitive grammarians distinguish syntax from phonology from
  2. Johanna Rubba, Langacker's Cognitive Grammar

Message 1: Do cognitive grammarians distinguish syntax from phonology from

Date: Fri, 22 Apr 94 16:35:19 BSDo cognitive grammarians distinguish syntax from phonology from
From: Bill Croft <>
Subject: Do cognitive grammarians distinguish syntax from phonology from

In LINGUIST 5-461, Larry Gorbet takes issue with Paul Deane's
characterization of one of the principles of cognitive grammar. I believe
Gorbet is right to do so. Contrary to what many nonfunctionalists
appear to believe, cognitive grammarians (and functionalists in
general) do distinguish between form and meaning, though they claim a
stronger relation between the two than do formalists.
 Unfortunately, Gorbet's post makes it appear that cognitive
grammarians instead abolish the distinction between phonology and
syntax. I would like to nip this misimpression in the bud (with Larry's
approval)---functionalists have enough problems to deal with already!

 First, an abbreviated version of the Deane-Gorbet interchange, to provide
the context:

>In his generally accurate post in LINGUIST 5.418, Paul Deane appears to
>accept a presupposition that I am pretty sure is inaccurate.
>The relevant paragraph begins:
>> Langacker's Cognitive Grammar is unique in its attempt to
>> REDUCE syntax to abstract semantic patterns.
>At least as most readers of the list are likely to interpret "semantic",
>this is almost antithetical to a central Cognitive Grammmar premise: that
>grammar is essentially symbolic.... In his most
>recent major work (Foundations of Cognitive Grammar II, 1991) he says (p.
>"The central claim of cognitive grammar is that language is fully
>describable is terms of semantic structures, phonological structures, and
>symbolic links between the two. Only symbolic structures
>need be posited for the characterization of lexicon, morphology, and
>syntax, which form a gradation that can be divided only arbitrarily into
>discrete components."
>... Langacker goes to considerable effort to
>support the claim that linguistic forms (including the form of
>grammatical constructions) are essentially phonological...But
>the claim that constructions (etc.) have meanings is fully consistent with
>regarding their forms as phonological. I think few Langackerian
>cognitive grammarians (if any) would claim that constructions consist
>solely of or reduce to their meanings.

 Langacker uses the term "phonological structure" to describe the form side
---"pole"---of a symbolic unit; the meaning side is less controversially
the "semantic pole". However, he speaks of "unipolar organization" vs. "bipolar
organization" of phonological structure. The unipolar organization is what most
linguists would call "phonology". The bipolar organization is what most
linguists would call "syntax"---or more precisely, morphology/lexicon/syntax,
since as Gorbet notes cognitive grammar takes that to be a continuum (but
another issue).
 Langacker uses the example of "tables". The unipolar organization of its
phonological structure would segment this unit tey-blz ---by syllable
The bipolar organization of its phonological structure would segment this unit
teybl-z ---by its morphological structure as an instance of the schema [NOUN-
PLURAL]. These two organizations can exist simultaneously in cognitive
grammar, as they do in formal theories of language. "Phonology" and "syntax"
are still distinct. This is something that I think formalists and
CAN agree on.*
 The claim of cognitive grammar, and I think, of functionalism in
controversial one that formalists take issue with---is that syntactic
organization exists
only by virtue of its being the form side of the form-meaning or symbolic
organization of language. In cognitive grammar, the symbolic structures
that [NOUN]
and [PLURAL] are part of are the fundamental units of grammar; the form side is
derivative of that. In cognitive grammar, syntax isn't reduced to
phonology, nor to
semantics; it is reduced to semiology, in the Saussurean sense of that term.
 Of course, this means demonstrating that most "formal syntactic" units,
particularly the important ones like 'noun', 'subject', 'head',
'constituent' etc.,
have some kind of semantic content so that they can be taken seriously as
symbolic units. So cognitive grammarians do a lot of semantics in order to find
evidence for their view of "syntax" (= bipolar organization of symbolic units).
That was what the original Deane-Gorbet discussion was about.

Bill Croft

*Hockett gives a straightforward functional explanation for this "dual
patterning" of language on p. 16 of "The problem of universals in language", in
J. H. Greenberg (ed.), "Universals of Language", MIT Press, 2nd ed., 1966.)

Dept of Linguistics, U Manchester, Oxford Rd, Manchester M13 9PL, UK Phone: +44-161-275 3188 FAX: +44-161-275 3187
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: Langacker's Cognitive Grammar

Date: Sat, 23 Apr 1994 08:56:14 Langacker's Cognitive Grammar
From: Johanna Rubba <>
Subject: Langacker's Cognitive Grammar

Larry Gorbet recently made a posting correcting Paul Deane's posting
about the reduction of syntax to semantics in Cognitive Grammar as
formulated by Ron Langacker. Larry's claim was that Langacker does
not in fact reduce syntax to semantics in this way, because syntax is
an aspect of linguistic form, which consists of phonological structures.
And phonological structures are not semantic structures, but are

I'm afraid Larry has this last point wrong. Langacker does posit syntactic
form as one kind of phonological structure; but phonological structures are
in fact claimed to be 'located' in 'semantic space'. I quote from Vol. I
of Foundations of Cognitive Grammar (1987, pp.78-79):

"sounds (at least for many linguisic purposes) are really concepts ...
If sounds are conceptual entities, our previous characterization of
symbolic space was oversimplified in treating semantic and phonological
space as disjoint fields of cognitive potential; phonological space
should instead be regarded as a subregion of semantic space ... Locating
phonological space within semantic space is more than a terminological
nicety, for it resolves certain actual or potential conceptual problems."

All of language is reduced to semantics under this view. Knowledge of
language is viewed as those portions of our general conceptual inventory
devoted to symbolic expression; these portions include the narrower notion of
semantics as the meanings of linguistic expressions, plus the phonological
forms which symbolize same, plus the links between the two which accomplish
the symbolic function of language.

Jo Rubba
The University of Montana
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue