Some questions and comments regarding the following recent exchange on this
list:
Christoph Unger wrote:
"If language is a tool like any other tool then it is most appropriate and
necessary to study the structure of the tool itself and not only the
purpose that it is used for. In other words: not studying the formal
properties of
language as a formal system is certainly inadequate." [End Unger quote]
Randy LaPolla wrote:
I don't understand this comment. No functional approach only studies the
purpose and not the structure of the tool. The difference between a
functional and purely formal approach is in terms of explanation. A
functional approach attempts to explain the structures found by reference
to their function in the activity the tool is used in. A purely formal
approach ignores the use of the tool and studies the form in the abstract.
The latter sort of analysis doesn't really tell you much, though. For
example, if we try to analyze the structure of a hammer without reference
to its use in hammering and pulling out nails, then there isn't much we can
explain about why it's structure is the way it is. All we can say is
"Hammer --> stick-like-part + heavy-part". Not very informative (cf. "S -->
NP + VP"). Even using the term "handle" would bias the analysis, as the
label derives
from its function. Personally, I am not interested in form for form's
sake, or "explanations" where formal structure is "explained" using formal
structure (thereby being circular--you observe a pattern, call it a rule,
then
"explain" the pattern by saying it conforms to the rule). I am interested
in understanding communication, so factoring out communication from the
study of language doesn't appeal to me." [End LaPolla quote]
I'd like to suggest that LaPolla, or somebody who sympathizes with his
views, provide a more suitable analogy than the hammer. I actually think
that the hammer example could be instructive---that a proper explanation of
a hammer would include a specification of its constituent structure and
form, and that this is not as trivial a task as LaPolla makes it out to be
if we don't presuppose our prior familiarity with hammers. Nevertheless, it
would be fair for a defender of the utility of a formalist approach to
demand an example of explanation of an artifact whose complexity is at
least approximately of the right order of magnitude to serve as an analogy
to language. I'm not sure what this would be. But whenever I try to imagine
what it would be like to provide a proper, detailed explanation of any
reasonably complex tool or machine with many interacting parts, interacting
in complex enough ways to be considered appropriate as an analogy to
language, I find that some interesting things happen. First, the
formulation of a principled functionalist/formalist distinction of the sort
we see in linguistics seems more or less absurd; second, modularity of
function and form seems inescapable; and third, while explanation of
function is central to the explanatory task, the articulation of structure
and form in their own right seems to be an essential component of that
explanation.
As mentioned above, I'm not sure what a suitable analogy would be. But out
of respect for the intelligence of all the linguists I know (functionalists
and formalists) I would be inclined to pick something quite complex indeed,
on the assumption that if human language were simple, it would have been
fully figured out by now. A Boeing 747 might be a good starting point. In
order for the analogy to work, we would have to suppose that we initially
know nothing about a 747 except vaguely that it is a machine for flying.
Our task would be to study 747s and figure out what we can about them in an
attempt to understand how they work. I admit that in some cases, structure
will transparently relate in an immediate way to function, as in the layout
of cables connecting controls and airfoils. But it seems quite clear that
there are other domains in which structure would have to be studied at
greater length in its own right, particularly in the investigation of the
electronics of the instruments and the contours of the airfoils. We might
even study the geometry and topology of the airfoils and manifolds as an
end in itself, using complex mathematical equations and computer models,
as, in fact, aerodynamic engineers do in designing them. The study of the
manifolds in this way would, of course, in some ultimate sense "attempt[]
to explain the structures found by reference to their function in the
activity the tool is used in"; but the problem in this case is so hard that
it demands as well some elements of "[a] purely formal approach [which]
ignores the use of the tool and studies the form in the abstract."
[adapting quotes from the LaPolla posting, excerpted above]
Our theory of the 747 would presumably exhibit some degree of functional
and structural modularity. But the most striking facet of this analogy, in
my view, is that the whole idea of a functionalist/formalist dispute or
divide would probably be considered absurd. Everyone would agree that our
ultimate aim is to understand the functioning of the 747, and that
considerations of function infuse much of our theorizing about the
structure of the plane. But anyone with a clear appreciation of the
difficulty of the task should admit that we also need an extensive "formal"
theory of some facets of the plane, such as the shape of manifolds,
mentioned above, or its instrumentation, where the level of causal
embedding makes immediate explanations in terms of function less accessible
or less insightful than explication in terms of the internal workings of
the mechanisms.
A functionalist might reply that the points made above about our
investigation of the 747 fit fully into the functionalist view since, after
all, "[n]o functional approach only studies the purpose and not the
structure of the tool." [From LaPolla excerpt above.] This sounds
reasonable enough. But LaPolla continues: "A purely formal approach ignores
the use of the tool and studies the form in the abstract. The latter sort
of analysis doesn't really tell you much, though." The example of the 747
shows that, in the right context, a purely formal analysis (of some aspects
of a system--which is all that formalists are committed to studying) can,
in fact, tell us much. Of course, the "right contexts" just mentioned
ultimately touch on the relationship of form to function. Likewise,
formalists admit that the formal machinery they find in language is
embedded in a larger system, ultimately touching on the relationship of
form to function. The formalists never claimed to be studying more than
certain subsystems of the language faculty.
Some linguists would say that, by confining ourselves to examples involving
artifacts, we concede too much to the functionalist. Some linguists, after
all, think that human language is a complex natural system, not a complex
artifact. What would be a suitable analogy in the domain of complex natural
systems? Several candidates come to mind, some of which have been pursued
in the literature: the structure and function of the blood, the structure
and function of the immune system, etc. I find a more intriguing analogy to
be the structure and function of the human genome. One virtue of this
analogy is that we are barely on the verge of knowing the answers, so we
won't commit the fallacy of underestimating the subtlety and complexity of
the system. In any event, formal tools, including tools of formal grammar,
interestingly enough, have already proven themselves to be useful in the
theory of the structure and function of genomic sequences.
Michael Hegarty
Department of English and Program in Linguistics
Louisiana State University
Baton Rouge, LA 70803
mhegar1@lsu.edu
This archive was generated by hypermail 2b29 : Mon Jun 24 2002 - 22:53:39 GMT