It's gratifying to see a discussion that engages so many of us, as well as
the different saliences and concerns with which each of us comes to it.
My own initial objective was to push the classical zpd paradigm for social
learning so far in various directions that one might begin to see where its
limits were. That doesn't mean I doubt its usefulness in analyzing the core
types of learning for which it was originated. Or that I'm not hoping some
of LSV's insights and the other wisdom that has accreted to the model won't
still offer us assistance as we stretch it out of shape and out of its
original sites of application.
It seems that most of what I hear regarding limitations in the way it is
used, even in its core domain of interpersonal, constructive, scaffolded
learning, are that its level of analysis, the dyadic, does not reach far
enough toward the organismic and human-individuality levels (bodies and
humans who care for one another) on one side, nor far enough toward the
macro-social and historical context on the other side. I think we have all
figured out well before now that the core problem of activity theory
analysis, or any other approach to meaningful human social behavior, is the
integration across levels and scales of this kind. I have been thinking and
writing a bit about this lately, struck by the ways in which cross-scale
integration is _necessary_ for eco-social-semiotic systems (places with
people who mean and care) in ways it is either not necessary, or quite
pointless to do for many sorts of elementary physical systems.
The fascination with pushing the limits of the zpd notion seems in turn to
center around the issues of 'negative learning' and 'learning with
nonhumans'.
There was perhaps some misunderstanding of my meaning in proposing the
first of these issues: I really meant some effect of developmental
retrogression, or stunted, unrealized development as a result of conditions
of social interaction -- and not merely the cultural value attached to what
was being taught. That's why I thought of the double-bind as a possible
instance. Most environments said to inhibit normal development are
described as depriving the organism of necessary stimulation or
information; I was looking for 'positive' features of an interaction that
could also do this, or worse. I am not sure I believe in the possibility of
genuine retrogression; development is an irreversible process. But there
are some terrible possible outcomes that fail to realize significant normal
species-specific potential, and no doubt also whatever cultural learning
promises beyond that.
Learning with nonhumans, however, is less an extremal condition and more
nearly a universal one. So the issue is what do we mean by it in direct
comparison with what happens in the zpd? Mediated social learning models
certainly allow for artifacts, and sign-displays in some role in
interpersonal social learning. We learn 'with' or 'through' them, but still
in some sense 'from' another human. Or so the usual story goes, but I don't
think this simple model of the separability of the role of the human other
from texts and tools in learning will hold up under critical scrutiny and
in the analysis of many actual learning examples. The dynamics of the total
system ties the interdependencies of actions and processes and their human
and nonhuman participants too tightly together for neat separability.
How is learning by reading a book different from learning by listening to
someone lecture? How is learning to operate a piece of exercise equipment
that only affords one mode of operation consistent with cultural
assumptions about its purpose different from being scaffolded nonverbally
by a trainer through the steps of a motor skill? How is learning by
interacting with an intelligent tutoring program different from learning by
interacting with a not very intelligent tutor?
We really do need to have much more useful things to say about such
questions than we can presently muster.
Some speculation based on the core insight of LSV about human learning:
that we learn first to do something in interpersonal interaction, and then
learn to recreate that inter-action 'internally'. We first learn to learn
from others, and then we learn to learn without others by an
'internalization' of the schema of the social learning process. We learn
from nonhumans by interacting with them in the same way we have learned to
learn by interacting with humans. In effect we pretend that they are people.
This makes more sense if we put back the rest of the system: the signs and
texts, the tools and artifacts, as well as the bodies and personas of
others. But we need a clearer idea just what 'internalization' might really
mean as well. I don't believe in 'minds' as a separate explanatory
principle, and insofar as my notions are equivalent to what other people
mean by the term, a 'mind', or meaning-making processes, are not localized
internally within the organism; they take place in systems across many
levels, including neurological levels, bodily levels, the level of the
micro-ecological situation in which the organism participates, the larger
scale systems linked to these through meaning systems of culture, etc. So
what I take LSV to have meant is simply that you and the rest of the system
can get along without the Other being physically present and active in the
process because all the other elements are still present, in matter or in
imagination, and/or by substitution and reinterpretation. That is, we
continue to do in the system without the other as much as possible as if
the other were present, including very often at first, imagining the other
present and speaking and guiding us, except that now we are do this
speaking, etc.
True theoretical individualism would probably say that we are learning from
ourselves at this point, or teaching ourselves, or scaffolding ourselves.
But when we make salient all the signs and texts and tools and artifacts,
it is clear that they are doing quite a bit of the scaffolding along with
us. And so we come to be able to have a conversation with a book, or to
react to a book as we make the meanings its words suggest to us (according
always to past external social scaffolding for how to read and interpret),
as if we were, in some respects, in conversation with a human. Many humans
go on and on and we only comment silently to ourselves, but we comment
according to many of the same conventions as when we might comment aloud.
It is a small step to let a part of ourselves fill in for the monologist,
guided by the text.
This model works best where sign mediation is clearly important; it is
easier to transpose the material source of the signs from a material Other
to an imaginary one. And we can probably be safe in saying that most
learning of interest beyond a very early age is sign-mediated. But the
model would still need a lot more development in cases like learning from
the landscape, where there may either be a non-semiotically mediated
component or the semiotic systems that mediate are ones we don't clearly
recognize (as when we learn a path through the landscape without either
making verbal directions for ourselves or visualizing a map or route, and
we certainly sometimes do this).
At some basic level I imagine that not all human learning is an extension
of the basic social-interpersonal learning paradigm; evolution has probably
also provided us with some forms of motor learning that adapt directly to
the nonhuman environment without the mediation of other persons, and
perhaps without the mediation of signs. Here is perhaps where the strange
relationship of tools and signs has its origins: tool-mediated learning has
some roots in the nonsocial learning capacities as well as in cultural
modes of learning, and there is a sort of continuum from the cultural tool
to the cultural artifact to the non-artifactual elements of the
environment. Even if sign-mediations are pretty well pervasive, we do still
interact with objects, and bodies, in non-sign-mediated ways as well (i.e.
simultaneously).
Enough. If people are interested in discussing particular cases, learning
from dogs, hot radiators, landscapes, machine-use, books, music, computer
programs, etc., we can see what such a model might say and what else needs
to be said too.
JAY.
---------------------------
JAY L. LEMKE
CITY UNIVERSITY OF NEW YORK
JLLBC who-is-at CUNYVM.CUNY.EDU
---------------------------