Helena puts the practical critique of 'transfer' into an extreme
form of theoretical dilemma: why should I believe that practices
in different activities-cum-situations have anything in common,
other than my ability as an analyst to construct some
similarities? We all know how we construct these similarities:
you build a ladder of abstractions, categories of similarity
built on various principles which are salient because the theory
we use makes them salient, until you reach a point where
descriptions of the two activities can be mapped one-to-one onto
each other. We make them sound the same. The participants in the
primary activities, however, do not participate in our
abstractive exercises and are often quite unfamiliar with the
principles we use to unify the disparate. Sometimes, by teaching,
explicitly or even implicitly, our culturally privileged
abstractive categories, we can get people to 'transfer' they way
they ought to do if there "are" similarities between the
activities. In effect we teach them how to "make" similarities.
But they have to do a lot more than we do. They also have to
climb back down the ladder of abstraction, and we all know that
is a lot harder to do in practice because of the number of
unspecified choices to be made along the way. The theoretical
model of similarity underdetermines the practice of transfer; it
is much harder to 'rise to the concrete'.
If of course we believe that just because we can reliably and
reconstitutably construct the abstract similarities that
therefore the activities really do have instrinsic properties in
common ... then we are going to get other people in trouble.
So perhaps my answer to the original question which Helena is
puzzling over: "What to you seem difficulties and opportunities
in obtaining ecologically valid results concerning cognitive
practices?" is that a key difficulty is that we may be defining
'cognitive practices' by one-way ladders, and the mere fact that
we can do so 'on the way up' (i.e. producing similarity by
abstracting away from naturalistic activities) is not enough. We
must also require that the similarity maintain itself, and that
we specify the procedures for, going back down the ladder (i.e.
re-contextualizing cognitive practices as integral constituents
of functionally and situationally distinct activities).
It is indeed a serious question whether this can _ever_ be done.
That is, whether generalizations across significantly different
activities, as being constituted from the 'same' practices, hold
up both upwards and downwards. Presumably it is possible to teach
people in some cases to behave in ways that make our theories
appear to be correct. Is this the _only_ way in which to validate
such abstractive generalizations? How can we determine the extent
to which we _have_ so trained people we observe? (i.e. before
testing the transfer). Or the extent to which they already share
some of the same cultural assumptions we employ in our upwards
constructions? (but without which the practices could still be
performed, but the similarity no longer be validated).
And what is the theoretical alternative? At the far opposite
extreme is particularism: each text or instance is unique,
happens in a different way than every other instance, the
uniquenesses are explanatory, the differences are interesting,
and similarities are strictly analytical artifacts. This is akin
to how we study startlingly original works of art (including
textual art). Since difference and similarity basically operate
in terms of the same practices of categorization, there is a
certain paradox, perhaps inconsistency in this view. But it is
still possible to focus on uniqueness, or at least on the
instantial vs. the typical. One always perhaps needs both, but
the question is what is of greater interest and value.
Hermeneutics vs. formalism. One can argue, I think, that
different kinds of systems require that we strike a different
balance between these poles. In some sorts of phenomena, in
physics and chemistry, uniqueness is pretty much impossible, or
irrelevant to our concerns. In other domains, the reverse. In the
case described by Helena, it looks likes the relevant concrete
practices are pretty situation-specific, and an analysis in terms
of abstract, more context-independent practices not so useful.
So perhaps another way of posing the issue of difficulties of
naturalistic investigation is to ask: How do we determine to what
extent it is useful to make our terms of description specific to
an event or situation-type or activity-type (as say with Geertz'
thick description or particularistic ethnography generally), and
to what extent to try to make them abstractly general across a
wide range of activities? One possibility here is that we always
need to study one activity in relation to others, cross-checking
the usefulness of our terms of analysis across two or more
activities, and always doing so both up (how comparable do they
make events of different types appear?), and down (how useful are
such formulations concretely in practice in each activity?).
Finally, a consistent naturalistic stance would suggest that we
should not be working back and forth with our terms of analysis
between arbitrary activities, or even ones that are presumptively
similar on our abstract theoretical or cultural criteria, but, as
in Helena's case, ones that are in fact interdependent on one
another as part of larger social-institutional networks of
activities. This might also ultimately lead to a different and
more viable kind of generalizability: not an abstract one that
posits universals, or even species-specific, or culture-specific
categories of analysis, but one that is specific to such
networks, and where the categories mutate in comprehensible ways
as one traverses the links in the network from activity to
activity. They could thus remain activity-specific (and so
useful) while also not being purely _ad hoc_ for each different
activity-type.
JAY.
------------
JAY LEMKE.
City University of New York.
BITNET: JLLBC who-is-at CUNYVM
INTERNET: JLLBC who-is-at CUNYVM.CUNY.EDU