play, rules

Jay Lemke (JLLBC who-is-at CUNYVM.CUNY.EDU)
Mon, 12 Feb 96 22:12:15 EST

Play, rules, learning

I have been fascinated with many of the contributions to the new
theme of play and rules that's emerged in our (and Vygotsky-L's
discussions).

My own notions about play come first of all from Bateson, who
emphasized the need to take a 'meta' or multiple levels of
abstraction view of play. In such a view, at the very least, we
have the Played Activity (what we're playing at) and the Playing
Activity (i.e. play as itself a kind of activity). For Bateson
the message 'this is play' is a meta-communication, a commentary
on the Played Activity which classifies it for social purposes as
also a Playing Activity. This need not be done, and often is not
done, explicitly. When a play-act is misinterpreted as not-play
do may say, 'but we were only playing', and there may be explicit
entry cues (e.g. 'lets play ...'), but while the playing is
happening, it can be very subtle how it distinguishes itself from
non-play behavior (if it does).

Part of the point here is that 'play' is a _category_ of activity
that is construed for certain activities. It may or may not make
sense to actually consider it an activity-type in its own right
(and it can certainly lead to classic confusions if we consider
action-as-play and action to be distinguished in the same ways we
distinguish activity-1 from activity-2).

So a key feature of play is how we can move into and out of it,
and negotiate about it (even about whether it's happening or
not), as in some examples here lately, esp. Ana Shane's. We could
see this as another of the wonderful semiotic 'doublings' of our
Real. Stanton Wortham's work on 'participant examples' has shown
the interesting effects of this doubling of
situations/activities: the activity of discussing an academic
classroom topic (here the analogue of play, the embedded
activity) conflated with the activity of negotiating the 'real'
social relations of the participants in the discussion (the
embedding, non-play or frame activity).

We humans (at least) can enact two activities simultaneously, and
we can mark one as 'play', or as being in some other meta-
relation to the first. Probably we must learn to do this, moving
from no meta-distinction made (play is a kind of real action), to
a simple extra marking (it's real action _and_ it's 'play'), to a
contrastive distinction (it's like real action, _but_ it's only
'play'). As with all development, we never quite lose what may be
developmentally earlier practices (development as cumulation, not
replacement). I think in some ways Bateson's model, and
Vygotsky's, are both trying to capture our capacity to 'play' in
all three modes. I often find myself shuttling among these modes
in the theatre or when reading a book: complete immersion as if
in reality (a 'trance' ala Milton Erickson?); participation but
with an awareness on the side that it's, say, a movie or a story;
and a distancing relation, that it's 'only a story', etc.

Play has rules, and meta-rules. It would not be recognizable or
playable as a distinct type of activity (playing House, playing
Doctor, playing Daddy & Baby, playing Hamlet) without the first-
order rules it shares with all activity. But its relations to the
non-play situation in which it is occuring can also be
foregrounded (as in shifts in and out, negotiations and disputes
'about' the play-reality) and these are enacted by meta-rules
(which are meta with respect to the play-reality, but first-order
with respect to the embedding situation). If the rules of the
play-world seem more fluid, it is because the meta-rules enable
us to intervene in and change them, subject to the ecological and
social constaints not of the play-world itself, but of the
embedding non-play world. The meta-relation frees the play-world
from some of its connectivity to our other worlds.

But, as Stanton pointed out, a lot of us have serious
reservations about many uses of the 'rules' metaphor for human
action. Neither play nor non-play activity (and the same activity
can in principle be both, though not usually at the same time and
in the same respect; Stanton's research examples come close, with
small critical shifts in the 'respect') can be 'generated' by
fixed or context-independent rules, and so neither can it be
explained in this way, at least not to my satisfaction, except
for limited, ritualized cases or highly regular 'respects'. A
more modest claim, with some cogency, is that we are in the habit
of explaining and regulating our behavior _with reference to_
rules, but do not act _from_ rules (rule-referencing vs rule-
governed behavior).

Perhaps most useful for me is the notion that we make up a lot of
the rules as we go along. Not, generally, the top-level rules,
but the interpretations of these rules for specific situations
that must always be generated to link abstract rule to concrete
instance, and the intermediate level rules that are part of this
linking work. We do this not just in play, but in all (or almost
all; sometimes it's nice to be relieved of this job) activity.
And it is very important that we do. We may not behave some way
because of these rules, but what we do makes sense to us and to
others in part through the application of these rules. Play is
not just about learning how to do activity-X, it is also about
learning how to negotiate the fluidity of 'rules' generally. In
our culture is significantly about learning rule-referencing
habits, and using these internally to (weakly, or coarsely)
regulate our behavior in different activities and (more
significantly, I think) to explain and interpret our own and
others' behavior, whatever it may actually be and from whatever
(spontaneous, emergent, situation-specific, context-generated,
self-organizing -- but not rule-governed) sources it may arise.

It may not be necessary to point out that this view is totally at
odds with the algorithmic (rule-generative) paradigm of cognitive
science explanations and models for human behavior (esp. the
'classical' ones of 1960s/70s Chomsky, H Simon, and AI
computation). It would propose that these models project onto
organisms (and design into machines) historically and culturally
specific models of rule-referencing which developed much later
than the underlying organismic and social-interaction capacities
of the systems which enact these behaviors by means we still
understand nearly nothing about. That may be a harsh judgment
(and does not imply that much useful work was not still done
under those rubrics), but I think it needs to be explicitly said.
Too many students and new researchers still think we already have
a basic, viable model of human behavior as cognitively governed,
when we are only just beginning to appreciate what kinds of
phenomena such a model should try to make sense of and what
capacities might underly them. JAY.

----------

JAY LEMKE.
City University of New York.
BITNET: JLLBC who-is-at CUNYVM
INTERNET: JLLBC who-is-at CUNYVM.CUNY.EDU