/LMAR=0/XLINE=3/FONT#0=BASL30/FONT#1=BASI30/FONT#2=BASB30/FONT#5=MICR25/FONT#6=NGB30
␈↓ ↓H␈↓¬␈↓ α≡OFFICE MEMORANDUM  *  STANFORD UNIVERSITY  *  OFFICE MEMORANDUM

␈↓ ↓H␈↓ε␈↓ βBSTANFORD ARTIFICIAL INTELLIGENCE LABORATORY

␈↓ ↓H␈↓α␈↓ ¬hMarch 17, 1977




␈↓ ↓H␈↓αTo:␈↓ αH␈↓Patrick Suppes, IMSSS

␈↓ ↓H␈↓␈↓αFrom:␈↓ αH␈↓John McCarthy, Director, Arti≡cial Intelligence Lab.

␈↓ ↓H␈↓αSubject:␈↓ αH␈↓Comments on ␈↓↓From Behaviorism to Neobehaviorism␈↓

␈↓ ↓H␈↓Dear Pat:

␈↓ ↓H␈↓        I␈α∪enjoyed␈α∀your␈α∪paper␈α∪␈↓↓From␈α∀Behaviorism␈α∪to␈α∪Neobehaviorism␈↓,␈α∀and␈α∪I␈α∪have␈α∀the␈α∪following
␈↓ ↓H␈↓comments.

␈↓ ↓H␈↓p.282.␈α∩ I␈α∩can't␈α∩agree␈α∩that␈α∩the␈α∩appropriate␈α⊃internal␈α∩realization␈α∩of␈α∩learning␈α∩a␈α∩new␈α∩problem␈α⊃or
␈↓ ↓H␈↓classi≡cation␈α⊂is␈α⊂the␈α⊃addition␈α⊂of␈α⊂a␈α⊂subroutine␈α⊃to␈α⊂a␈α⊂program.␈α⊂ If␈α⊃this␈α⊂were␈α⊂so,␈α⊂people␈α⊃would␈α⊂be
␈↓ ↓H␈↓inclined␈α∞to␈α∞speak␈α∂in␈α∞programs.␈α∞ I␈α∞think␈α∂it's␈α∞closer␈α∞to␈α∞adding␈α∂new␈α∞sentences␈α∞than␈α∞to␈α∂adding␈α∞new
␈↓ ↓H␈↓programs, but what changes correspond to learning something is a deep problem as yet unsolved.

␈↓ ↓H␈↓p.␈α283.␈α I␈αagree␈α
that␈αpsychology␈αis␈αnot␈α
reducible␈αto␈αphysics,␈αbut␈α
because␈αthe␈αphysical␈αhardware␈α
will
␈↓ ↓H␈↓support␈α∂almost␈α∂any␈α∞psychology.␈α∂ Therefore,␈α∂it␈α∞is␈α∂irrelevant␈α∂whether␈α∞dissection␈α∂of␈α∂the␈α∂brain␈α∞will
␈↓ ↓H␈↓reveal␈αthe␈αlanguage␈αspoken.␈α If␈αwe␈αstop␈αa␈αcomputer,␈αthen␈αphysical␈αexamination␈αof␈αits␈αcore␈α
memory
␈↓ ↓H␈↓can␈αreveal␈αthe␈αprogramming␈αlanguage,␈αbut␈αsince␈αthe␈αvery␈αsame␈αmemory␈αcould␈αsupport␈αa␈αdi≥erent
␈↓ ↓H␈↓programming␈α
language,␈αthe␈α
theory␈αof␈α
programming␈α
languages␈αis␈α
still␈αnot␈α
reducible␈α
to␈αelectronics.
␈↓ ↓H␈↓It␈αseems␈αto␈αme␈αthat␈αthere␈αare␈αtwo␈αpossible␈αsenses␈αof␈α"reducible".␈α The␈αcurrently␈αuninteresting␈αsense
␈↓ ↓H␈↓is␈α
whether␈α
the␈α
brain␈α
is␈α∞a␈α
physical␈α
system␈α
whose␈α
reactions␈α∞follow␈α
from␈α
the␈α
laws␈α
of␈α∞physics.␈α
 The
␈↓ ↓H␈↓second␈α∂is␈α∂the␈α∂sense␈α∂mentioned␈α∂above.␈α⊂ We␈α∂could␈α∂have␈α∂the␈α∂same␈α∂psychology␈α∂based␈α⊂on␈α∂di≥erent
␈↓ ↓H␈↓physics and a di≥erent psychology based on the same physics.

␈↓ ↓H␈↓        Finally,␈αI'm␈αpuzzled␈αabout␈α
how␈αmuch␈αdoctrine␈αis␈αleft␈α
in␈αbehaviorism␈αwhen␈αit␈α
becomes␈αneo-
␈↓ ↓H␈↓behaviorism.␈α~ For␈α~example,␈α~is␈α≠it␈α~compatible␈α~with␈α~doing␈α~metaphilosophy␈α≠analogously␈α~to
␈↓ ↓H␈↓metamathematics?␈α∩ The␈α∩idea␈α∩is␈α∩to␈α∩consider␈α∩a␈α∩system␈α∩that␈α∩includes␈α∩a␈α∩person␈α∩automaton␈α∩as␈α∩a
␈↓ ↓H␈↓subautomaton␈α
of␈α
a␈αworld␈α
automaton.␈α
 We␈αthen␈α
study␈α
what␈αvarious␈α
philosopher␈α
programs␈α
in␈αthe
␈↓ ↓H␈↓person␈α
will␈α
learn␈α
about␈α
the␈α
world␈α
as␈α
it␈α
depends␈α
on␈α
the␈α
program␈α
and␈α
the␈α
world.␈α
 This␈α
should␈α
be␈α
an
␈↓ ↓H␈↓entirely␈α⊂technical␈α⊂subject.␈α⊂ Controversy␈α⊂would␈α⊂arise␈α⊃only␈α⊂when␈α⊂one␈α⊂tried␈α⊂to␈α⊂map␈α⊃the␈α⊂technical
␈↓ ↓H␈↓results onto ourselves and the world.


␈↓ ↓H␈↓Best regards,



␈↓ ↓H␈↓John McCarthy, Director, Arti≡cial Intelligence Lab.