COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	lewont[w85,jmc]		Notes for review of Lewontin
C00006 ENDMK
C⊗;
lewont[w85,jmc]		Notes for review of Lewontin

	I'll read the book with a hypothesis, to be confirmed or
refuted or equivocated by the experience.  The hypothesis is that
Lewontin, and the Science for the People people, are Marxist
"fundamentalists" analogous to religious fundamentalists.  Fundamentalists
according to the present idea, are people who have experienced
prolonged doctrinal defeats at the hands of reason and who
discover that they can recover all their lost ground by giving
up reason except for lawyer's reason.  Lawyer's reason is reason
in support of a cause.  It may be logical in drawing consequences
from the facts it chooses to take into account, but it selects
those facts with regard to the case to be made rather than with
a disinterested pursuit of truth.  Circumscription may be of some
help in explaining this kind of reasoning.  Perhaps it would even
be worthwhile to point out the validity of some of Lewontin's
arguments, given the facts he chooses to take into account.

1985 Jan 2 - Lynn Scarlett says Reason will probably want the review.

mention the ideology of SftP.

	Ordinary deductive reasoning, as studied by mathematicians
and philosophers, has the following monotonic property.  Suppose
a conclusion  p  can be deduced from a collection  A  of statements,
taken as premises.
Suppose  B  is a more inclusive set of premises, i.e. includes all
the statements in  A.  Then  p  can be deduced from  B.  Indeed the
same proof that proves  p  from  A  will serve as a proof of  p
from  B.  The word monotonic, in accordance with its common use
in mathematics, refers to the fact that when the set of premises
is increased, the set of conclusions is also increased.

	What has been discovered, perhaps we should say emphasized,
in the research in artificial intelligence of the last ten years
is that human reason, and the reason required for intelligent
computer programs is not always monotonic.  We often draw a conclusion
from a set of facts that we would not draw from some more extensive
collection of facts, i.e. the conclusions we draw depends on the
facts we take into account.  What's new, and what I will spare you,
is the mathematics of how conclusions are drawn from a set of
premises that might not be drawn when additional premises are added.
However, the general idea is simple, namely to take the simplest
or the ``standard'' interpretation of the facts.