On "Thick Description"

John L.McCreery (jlm@TANUKI.TWICS.CO.JP)
Wed, 15 Dec 1993 11:50:37 JST

Bob Graber writes, "Geertz, in his introduction to*The Interpretation of
Cultures*, pays lip service to science; but this unjustly famous essay,
"Thick Description," on close reading proves sloppy and hypocritical." I
offer here the alternative view that what Geertz does in this essay is to
outline a project but leave his outline incomplete. He argues the need
for "thick description" but leaves unspecified how we should do it. To
reject what he says wholesale is to throw out the baby with the bath
water. It is far more interesting to consider (1) how "thick description"
might be done and (2) how competing descriptions might be evaluated.

First, a look at Geertz' own argument. As I read the essay, he argues
against both "behaviorist" and "mentalist" views of culture. The first
sees culture as patterns in behavior that could, in principle, be
described in purely physical terms. God might be able to do it, but most
anthropologist proceed instead by assuming "meaning" in what their
informants do and say. The argument that meaning is inherent in
behavior collapses when confronted with the fact that the same
physical movementQthe famous twitch which may or may not be a
winkQcan mean many different things: a twitch, a wink, a parody of a
wink, a rehearsal for a parody of a wink, etc. The meaning changes
with the situation, and the situation is specified in terms of someone's
meaning: the native's or he anthropologist's; but there comes "meaning"
again. The attempt to specify meaning in purely physical terms
becomes an infinite regress. Where then is meaning located? The
"mentalist" view says that it's place is inside human minds, but if
minds are distinct from behavior, then how do we get from one to
another. Now we're caught on the other horn of Cartesian dualism. So
back we go to what it is that most cultural anthropologists actually do:
watching people and listening to what they have to say, then
considering what we see and here in terms of our own theoretical
agendas; our own "meanings." That brings us back to "meaning," a
theoretical construct which like, say, an "electron" has to be inferred
from observations but cannot be reduced to them. We note that in most,
if not all, human situations multiple meanings come into play and the
commonsense principle that to understand what's going on we need to
sort them out as much as we can. That we shall ever have them totally
sorted out is impossible. None of us every "gets to the bottom" of what
we study; there's always something else to say. Which doesn't imply at
all that we can't evaluate interpretations as better than others. We do
that all the timeQin business, politics, kindergartens and scholarly
quarrels. The problem isQand here is where Geertz fails usQthat we
don't have the same kind of clearly defined criteria for assessing
attributions of meaning that we have for evaluating experiments or
rejecting the null hypothesis when we've done the statistical sampling
right. Developing those criteria? Now that's an interesing problem. I
await your suggestions.

John McCreery (JLM@TWICS.CO.JP)