Churchland and prediction

Vance Geiger (geiger@PEGASUS.CC.UCF.EDU)
Sun, 17 Sep 1995 16:58:37 -0400

Churchland's _Neurophilosophy: Toward a unified science
of the mind/brain.


In trying to understand the functional principles governing the
human nervous system, we must remind ourselves that our brain has
evolved from earlier kinds of brains--that our kind of brain was
not built from scratch especially for us, but has capacities and
limitations that are due to its historical origins. The pressure
for nervous systems to evolve has derived not from the intrinsic
beauty of rationality or from some indwelling goodness attaching
to cognition, but primarily from the need for animals
successfully to predict events in their evironment, including of
course events originating in other organisms....The fundamental
nature of cognition is rooted in the tricks by which assorted
representational schemas give organisms a competitive advantage
in predicting.


Comment: Jeeesh does this sound familiar. Having thought about
and argued about this for a long time, a couple of questions.

1. How does Churchland deal with the confounding of prediction
as an end in itself and prediction as a process. Culture as we
know it, is possible because we can create and act on predictions
as if they were valid, and thus to some extent make the
predictions come true. Driving a car is a perfect example of a
human activity that is entirely dependent on interpersonal
predictability. Another example, I can still elicit the response
from undergraduates (we are talking USA here) that it is
"inappropriate" for a Ken doll to wear a dress to class (I say
"still" as I sort of expect that at some time this will not work
anymore) even when the clothes one wears to class have no bearing
on a person's ability to do, i.e. the physical behaviors, what
someone does in class. This is a judgement entirely derived from
predictions about what people wear and do not wear to class.
Further, if you ask students if they would sit next to someone
who behaved this way they tend to say no, thus their predictions
also affect their behavior.

To continue, these are example of culture where the ability to
create predictions makes knowledge of what to do possible. These
predictions are arbitrary however, as they are not derived from
any physical constraints. Consequently, much of culture as
studied by anthropologists (much, not all) is an artifact of the
capability to create predictions of events that are not directly
related to the purely physical or material business of surviving
and reproducing and thus arbitrary as they are not subject to
direct testing in an evolutionary sense. In fact, many aspects
of culture are very probably contingent on a whole range of
historical factors, that could have yielded other equally
arbitrary alternatives. The persistance of such materially
detached behaviors can persist in a population entirely as a
result of the interpersonal predictability or knowledge about
other people they provide to predictability hungry big brains.

Does Churchland address this?

There was this discussion on the list a while ago about whether
culturally based predictions which tend toward short term
validity could yield long-term maladaptive behaviors.

Does Churchland address this possibility?


2. The split between neurobiology and cognitive psychology may
not be all that great. Ever since Herbert Simon and
"satisficing" and the heuristics of Tversky, Khnemann and such,
there has been an interest in the evolutionary consequences of
analyzing behaviors cognitive processing in terms if making lots
of decisions, i.e. doing a lot of trial and error, versus
optimizing in a world of imperfect information. Much of
cognitive psychology addressing typological thinking and
categorization is directed at how humans constrain the amount of
information (through attribution biases and counterfactuals, for
example) that has to be processed and thus the predictions and
consequences that have to be considered in making a decision. In
essence, culture here is an artifact of the constraints on the
amount of information a human mind can process and represents a
choice in limiting the amount of information that has to be dealt
with to make a decision about, for example, what another person
will do.

Does Churchland go into this?

vance geiger