The "bad" functionalism in question here is (mainly) American sociological
functionalism, whose origins Leigh Star already sketched and critiqued. In
its day it was perhaps rather progressive, at least insofar as it led to
some of the only systematic theory in the field (by Talcott Parsons, whose
brand of functionalism is probably sophisticated enough to elude the lapses
of most other American sociological functionalists).
What is generally seen to be "bad" about this functionalism is that it
basically says: if it happens it serves some necessary social purpose. It's
a bit like natural selection "explanations": if it's in our genes it must
confer some selective advantage. It's easy to distort this weak explanatory
argument into a more dangerous legitimation argument: our current social
institutions exist because they serve necessary functions. The functions,
being necessary, are then morally neutral, like "nature", and the
institutions, no matter how nasty, are still "good" in some sense because
they serve the necessary functions, and without them we'd be in trouble.
This can become of course a very conservative argument, which goes back
probably to Burke for its classic form: if an institution has evolved a
longstanding social form, don't mess with it because it must be serving
necessary social purposes that you probably don't understand. (Notice that
this is exactly the same argument that most people today accept when
applied to the natural environment. Given the inseparability of the natural
and the social in practice, if not in our discourses about them, this
should give pause ... we may be able to see the misuse of functionalist
arguments in the social domain, but we may not really understand in just
what sense they have a point and in what sense they are basically
misleading. This is not an easy question.) This is the functionalism that
Lave implicitly criticizes.
There is also, however, a different or "good" functionalism -- probably the
one that Mary Douglas is alluding to -- which has some common origins in
the very early British functionalist tradition (which does come from
biological organicism, which has given us everything from Nazi social
theory to my own ecosocial dynamics, and a lot in between ... partly
because social systems _are_ biological systems, just biosystems of a
special and distinctive kind). In British social anthropology (e.g.
Malinowski, who wasn't British, but was the leading figure anyway) there
was a variant of functionalism that came to oppose itself to "formalism",
especially in the study of language and sign systems (Mary Douglas is in
this broadly semiotic anthropological tradition). It argues that it is not
enough to simply map out, as formal syntax theory in linguistics or the
narrower versions of French structuralism in anthropology do, what possible
terms/items/acts and combinations occur: you cannot understand patterns of
forms unless you relate those patterns to their wider social functions
(language in use, etc.). This is the minimal or weak functionalist
principle in this other functionalism.
There is also a stronger version which says that the patterns have evolved
to fulfill these social functions -- and this is quite like the American
cousins applying this principle to other social institutions (language is
an "institution" in Parsonian sociology). The subtlest version of this that
I know is late British functional linguistics (Firth, Halliday, Sinclair,
etc.), which argues that it is not the patterns per se that serve a social
function (except locally: the utterance has an immediate social function),
but the system of resources with which we construct these patterns (the
system is the institution now), which has evolved to fulfill much more
general functions (common to many or most immediate social situations).
Acts or utterances have functional meaning only in a specific context of
use, but the system of meaning-making resources as a whole serves a very
wide range of social functions. In my own current thinking, these two
'functions' operate on very different time scales, and the forms which
serve such functions change on very different time scales ... and there
must be many functional units of analysis in between (e.g. between the
utterance and the language we have idioms, templates for common metaphors,
rhetorical styles, genres, etc.) corresponding to the intermediate time
scales.
So "good" functionalism says: if you want to understand something, look at
how it functions in its normal contexts of use; always try to explain
patterns of forms in terms of patterns of functions. "Bad" functionalism
says: patterns are as they are because they must be that way in order to
serve their necessary functions. A good functionalist would mainly disagree
with the "must be", and not just for the historicist reasons that much of
the patterns are contingent and accidental and could have been otherwise,
but also because good functionalism is a many-to-many mapping of form and
function, while bad functionalism tends toward a one-to-one relationship
(or even the many-to-one overdetermination of multiple functions). In good
functionalism there is always _mediation_ between form and function (e.g.
in functional linguistics there are grammar and context -- in different
mediations -- between utterance and meaning).
Contingency is not just historical, it is also characteristic of every
functional act ... there are always different possible ways to get
something done .... and this voids the necessitarian basis of "bad"
functionalism ... even if we agree that there are necessary social
functions, there is no reason why institutions have to have precisely the
form they do (or even remotely the form they do) in order to fill those
functions. Concommitantly, if you do want to identify _necessary_ formal
properties of institutions, you have to go to extremely abstract kinds of
formal properties (as in Chomskyan formal syntax theory), or in my view of
things, to very long timescales. While there may be such properties (a)
they are often too abstract to be of much interest, and (b) they apply to
phenomena at such long timescales that they are irrelevant for purposes of
human action (i.e. forget about changing or manipulating them over the
course of a century or two!). And of course in the long run you will always
be defeated by mediation ... trying to control longterm phenomena, you will
be undone by all the intermediate term phenomena which are changing on
faster timescales ... and which, collectively and cumulatively, constitute
the longer term phenomenon you are trying to change. Change itself serves
functions on ALL time scales ... and so we come to a not very conservative
non-conclusion ... JAY.
---------------------------
JAY L. LEMKE
PROFESSOR OF EDUCATION
CITY UNIVERSITY OF NEW YORK
JLLBC who-is-at CUNYVM.CUNY.EDU
<http://academic.brooklyn.cuny.edu/education/jlemke/index.htm>
---------------------------