[net.ai] Internalized World Knowledge

chertok%ucbkim%Berkeley@sri-unix.UUCP (04/04/84)

From:  chertok%ucbkim@Berkeley (Paula Chertok)

       [Forwarded from the CSLI bboard by Laws@SRI-AI.]

             BERKELEY COGNITIVE SCIENCE PROGRAM
                        Spring 1984

            IDS 237B - Cognitive Science Seminar

          Time:        Tuesday, April 10, 1984, 11-12:30pm
          Location:    240 Bechtel


              HOW THE MIND REFLECTS THE WORLD
                      Roger N. Shepard
       Department of Psychology, Stanford  University

Through biological evolution,  enduring  characteristics  of
the  world  would  tend  to become internalized so that each
individual would not have to learn  them  de  novo,  through
trial  and possibly fatal error.  The most invariant charac-
teristics are quite abstract: (a) Space  is  locally  three-
dimensional,  Euclidean, and isotropic except for a gravita-
tionally conferred unique upright direction. (b) For any two
positions  of  a  rigid  object, there is a unique axis such
that the object can be most  simply  carried  from  the  one
position  to  the  other  by  a  rotation  around  that axis
together with a translation along it. (c) Information avail-
able  to  us about the external world and about our relation
to it is analyzable into  components  corresponding  to  the
invariants  of  significant  objects,  spatial  layouts, and
events and, also, into components corresponding to the tran-
sitory  dispositions, states, and manners of change of these
and of the self relative to these.   Having  been  internal-
ized,  such  characteristics  manifest themselves as general
laws governing the representation of objects and events when
the  relevant information is fully available (normal percep-
tion), when it is only partially available (perceptual  fil-
ling  in or perceptual interpretation of ambiguous stimuli),
and when it  is  entirely  absent  (imagery,  dreaming,  and
thought).    Phenomena  of  identification,  classification,
apparent motion, and imagined transformation illustrate  the
precision and generality of the internalized constraints.

*****  Followed by a lunchbag discussion with speaker  *****
***  in the IHL Library (Second Floor, Bldg. T-4) from 12:30-2  ***