Comments: To: lfg at list.stanford.edu

[LFG List]

This continues discussion of issues raised by Joan's note on LFG and
functionalism.

Everyone would probably agree that the terms "functionalism" and
"formalism" can mean different things to different people. It might be
useful therefore to separate these different meanings before answering
the question, "Is LFG functionalist"? We would like to make an initial
attempt, and invite others
to add to or modify our formulations.

Formal-1, Functional-1: A formal-1 theory or grammar is one that makes
use of a logico-mathematical language to formulate the propositions of
the theory/grammar and calculate their consequences. In this sense,
the theory of relativity is a formal-1 theory: it makes use of the
formalism of Reinmanian geometry. Marxism is not a formal-1
theory. LFG is formal-1, but Li and Thompson's Mandarin Syntax is
not. We may say that Li and Thompson's work is functional-1 because of
the absence of logico-mathematical formalism.

Formal-2, Functional-2: A formal-2 theory or grammar is one that is
concered with structure/form of something. A functional-2 theory or
grammar is one that is concerned with the function/role/purpose of
something in a larger context, where "function" means "what something
does" (purpose). In this sense, anatomy is formal, but physiology is
functional. (e.g. structure/form of the heart, vs. what the heart
does). LFG is clearly formal-2. We suspect that it is not
functional-2. Halliday's linguistics is clearly functional-2, because
it is concerned with the functions of language in human communities.
(Of course, a theory can be both formal-2 and functional-2)

Functional-3, Formal-3. A functional-3 theory or grammar is one that
allows "external explanations" of internal phenomena. A formal-3
theory or grammar disallows external explanations. Sociobiology is
functional-3 because it provides exlanations of social patterns in
terms of biology. Francis Crick's approach to vision is functional-3
because he tries to explain vision in terms of neurons, but David
Marr's approach is not functional-3. Explaining phonological patterns
in phonology in terms of human physiology or communication is
functional-3, hence grounded phonology is functional-3.  Explaining
syntactic patterns in terms meaning or language interaction is
functional-3. LFG is open to the functional-3 approach, but we believe
that a lot more work needs to be done in this area.

Formal-4, Functional-4. A formal theory or grammar is one that allows
internal explanations. Functional-4 disallows internal
explanations. LFG is formal-4. A number of "functionalists" that we
have talked to behave as if they disallowed internal explanations. We
have a feeling that John Ohala and Sandy Thompson are functional-4: JO
seems to think that all phonology can be reduced to phonetics, and ST
seems to think that all syntax can be reduced to non-syntax. (We are
sure that both JO and ST would reject this extreme position when the
issue is explicitly stated like this.) LFG is clearly not
functional-4.

The terms "formalist" and "functionalist" would apply to all the
meanings listed above. There is yet another meaning of "function"
which doesn't seem to be related to the word "functionalist". This is
the use of the term to refer to grammatical relations such as subject
and object. LFG expresses these relational concepts in "functional
structure", separate from the representation of categories in phrase
structure. However, as everyone would agree, the inclusion of a
"functional structure" by itself doesn't make LFG functionalist.

Our functional-3 is the same as Joan's point 3. We believe that LFG
is formal-1, formal-2, functional-3, and formal-4, and that in fact,
it is opposed to formal-3 and functional-4.

Joan's point (2) is strictly speaking about the model of grammar, not
about functionalism per se. As Joan points out, the multidimensional
view of the organization of language (consisting of parallel
dimensions of structure related through correspondence rules or
constraints) is common to Halliday and LFG. This view opens up the
possiblity of various kinds of interactions (e.g. syntax-semantics
interaction, semantics-phonology interaction, syntax-discourse
interaction, etc). Hence, it lends itself to external explanations
(functional-3). However, multidimensionality by itself does not
necessarily make the theory functionalist.

 K.P.Mohanan, Tara Mohanan

----
Tara W Mohanan, Dept of English & Literature, NUS