functionalism vs generativism vs ...

Scott Delancey delancey at OREGON.UOREGON.EDU
Fri Jan 1 19:49:24 UTC 1999

On Thu, 31 Dec 1998, Dick Hudson wrote:

> Guy Deutscher's query implies that there are only two positions:
> functionalism or generativism. Those who align themselves with one of these
> positions may like to think this is so, but it ain't. There's also
> cognitivism (non-modular, but also not functionalist in the `discourse'
> sense), and Labov's view (explicitly non-functional but not generativist in
> the normal sense either), and ....

I have no argument with Dick's larger point, but I want to take issue
with the definition of functionalism which is implied here.  The
differences between functional and formal approaches have always been
described, and in part defined, by arguments about what counts as data.
As generative grammar has come to take more and more cognizance of
typological and historical data, it may be true, by this point in time,
that functionalism defined by the database differs from formalism
primarily in its inclusion of discourse structure as both explanans and
explanandum.  But while there are functionalists who attempt to explain
everything in discourse terms, this can't be taken as a definition
of functionalism, unless we want to exclude the likes of Givon or
Bybee or Heine from that label.  Such a restriction would certainly be
        The basic difference between functionalism and formalism is in
where explanations are lodged, and what counts as an explanation.
Formal linguistics generates explanations out of structure--so
that a structural category like Subjacency counts as an
explanation for certain facts about various syntactic structures
and constructions.  Most contemporary formal theories, certainly
generative grammar, provide ontological grounding for these
explanations in a hypothesized, but  unexplored and unexplained,
biologically based universal language faculty.
     Functionalists, in contrast, find explanations in function.
Formal principles can be no more than generalizations over data, so that
most generative "explanation" seems to functionalists to proceed on the
dormitive principle.
     There is a range of different functional arenas in which
explanation can be sought--sentence processing / memory
constraints, discourse functions, cognitive structure, even historical
tendencies.  Of course most researchers specialize in one of these areas,
(or at least in one at a time; think of Chafe's earlier work on
semantic patterns as explanation for structure, and his later work
primarily on discourse).  And sometimes--as everywhere in science and
scholarship--some of us come to believe that all real explanation lies in
our own bailiwick.  But in truth the different domains of functional
explanation don't separate out all that easily.
     In particular, when we talk about cognitive explanations, we're
explicitly claiming that linguistic structure is informed by general
patterns of thought.  Obviously these same cognitive factors must inform
other domains such as discourse structure as well, so that at a
sufficiently deep level of analysis "cognitive" and "discourse-functional"
theories are complementary.  While it is certainly true that, on the
contemporary linguistic scene, "cognitive" and "functional" linguists
represent distinct (though overlapping) social-interactional sets,
I think it is a mistake to regard them as competing theoretical

Scott DeLancey
Department of Linguistics
University of Oregon
Eugene, OR 97403, USA

delancey at

More information about the Funknet mailing list