246
PARTISAN REVIEW
sive gambler. First , any contradiction between experience and one magical
notion is explained by reference to other magical notions . Thus the entire
structure of the magical system of beliefs is supported by its very circularity.
This way of protecting the system against assaults upon it that come from
reality is especially effective if objections are always met one at a time. For
then , the very demonstration that an apparently anomalous fact can be incor–
porated in the system serves to validate the system. The gambler may, for
example , appeal to the fact that he didn't tie his shoelace , as he knew he
should have, to account for his" bad luck " on a particular day . That sort of
explanation is formally equivalent to the compulsive programmer's assump–
tion that his program's misbehavior is caused by a merely technical program–
ming error.
A second way the conceptual frameworks of gamblers and programmers
are protected is by cyclical elaboration. The gambler who suddenly realizes
that certain of his tricks work only on Thursdays simply incorporates this new
" insight" into his already existing framework ofsuperstitions , thus, in effect,
adding an epicycle to its structure . The programmer is free to convert every
new embarrassment into a special case , to be handled by a specially con–
structed
ad hoc
subprogram and thus incorporated into his overall system.
The possibility of this kind of unbounded epicyclic elaborations of their
systems provides both programmers and gamblers with an inexhaustible
reserve of subsidiary explanations for even the gravest difficulties .
Finally , the conceptual stability ofa magical or of a programming system
may be protected by denying, as Michael Polanyi writes in
Personal Knowl–
edge,
" to any rival conception the ground in which it might take root.
Experiences which support [the rival conception] may be adduced only one by
one . But a new conception . .. which could take the place off the one held],
can be established only by a whole series of relevant instances, and such evi–
dence cannot accumulate in the minds of [gamblers or programmers] if each
of them is disregarded in its turn for lack of the concept which would lend
significance to it." The gambler constantly defies the laws of probability . But
he refuses to recognize the operational significance of these laws . He can,
therefore, not permit them to become a kernel of a realistic insight. A
particular program may be foundering on deep structural , mathematical, or
linguistic difficulties about which relevant theories exist . But the compulsive
programmer meets most manifestations of trouble with still more program–
ming tricks and thus, like the gambler, refuses to permit them to nucleate
relevant theories in his mind . Compulsive programmers are notorious for not
reading the literature of the substantive fields in which they are nominally
working.
These three mechanisms , called by Polanyi , circularity, self-expansion ,
