/LMAR=0/XLINE=3/FONT#0=BASL30/FONT#1=BASI30/FONT#2=BASB30/FONT#10=BAXM30/FONT#11=ZERO30/FONT#3=STA200/FONT#4=NGB25
␈↓ ↓H␈↓␈↓βS␈↓∧ Artificial Intelligence Laboratory, STANFORD UNIVERSITY, Stanford, California 94305

␈↓ ↓H␈↓∧Telephone 415 497-4430␈↓ 
⊃April 20, 1978 




␈↓ ↓H␈↓Dr. Robert Marston, President
␈↓ ↓H␈↓University of Florida
␈↓ ↓H␈↓Gainesville, Florida

␈↓ ↓H␈↓Dear Dr. Marston:

␈↓ ↓H␈↓        Many␈α∞thanks␈α∂for␈α∞your␈α∂hospitality␈α∞on␈α∂your␈α∞island␈α∂and␈α∞at␈α∂your␈α∞home␈α∂and␈α∞for␈α∂so␈α∞patiently
␈↓ ↓H␈↓letting␈α
Marvin␈α
and␈α
me␈α
harangue␈α
you.␈α
 I␈α
enclose␈α
the␈α
enclyclopedia␈α
article␈α
on␈α
AI␈α
that␈α∞I␈α
promised
␈↓ ↓H␈↓and␈α∂also,␈α⊂in␈α∂case␈α∂you␈α⊂may␈α∂≡nd␈α∂it␈α⊂interesting,␈α∂an␈α∂article␈α⊂that␈α∂covers␈α∂some␈α⊂of␈α∂the␈α⊂common␈α∂area
␈↓ ↓H␈↓between␈α⊃arti≡cial␈α⊂intelligence␈α⊃and␈α⊃philosophy.␈α⊂ I␈α⊃send␈α⊂it,␈α⊃because␈α⊃it␈α⊂elaborates␈α⊃my␈α⊃reasons␈α⊂for
␈↓ ↓H␈↓believing␈α∃that␈α∀arti≡cial␈α∃intelligence␈α∀and␈α∃philosophy␈α∀have␈α∃more␈α∀in␈α∃common␈α∀than␈α∃the␈α∀usual
␈↓ ↓H␈↓connection between philosophy and a science, the philosophy of which may be studied.

␈↓ ↓H␈↓        Here␈α∞are␈α∞some␈α∞comments␈α∂about␈α∞possible␈α∞activities␈α∞of␈α∂the␈α∞Sloan␈α∞Foundation␈α∞in␈α∂support␈α∞of
␈↓ ↓H␈↓interdisciplinary studies in cognition.

␈↓ ↓H␈↓1.␈αArti≡cial␈α
intelligence's␈αmost␈α
important␈αneed␈α
is␈αan␈α
increased␈αemphasis␈α
on␈αbasic␈αresearch.␈α
 Besides
␈↓ ↓H␈↓AI's␈αinternal␈αtendency␈αto␈αattempt␈αpremature␈αapplications,␈αthere␈αhas␈αbeen␈αa␈αGovernment,␈αespecially
␈↓ ↓H␈↓DoD, tendency to demand short range goals.

␈↓ ↓H␈↓2.␈α∂One␈α∞program␈α∂that␈α∂could␈α∞help␈α∂basic␈α∂research␈α∞immediately␈α∂would␈α∂be␈α∞a␈α∂substantial␈α∂number␈α∞of
␈↓ ↓H␈↓nationally␈α
competitive␈αpostdoctoral␈α
fellowship␈α
positions␈αin␈α
cognitive␈α
science.␈α Much␈α
of␈αthe␈α
research
␈↓ ↓H␈↓that␈αneeds␈αdoing␈αis␈αperhaps␈αbest␈αdone␈αby␈αyoung␈αscientists␈αwithout␈αother␈αcommitments␈αand␈αcan␈αbe
␈↓ ↓H␈↓done␈αon␈αan␈αindividual␈αbasis.␈α Experimental␈αresearch␈αin␈αAI␈αand␈αcomputational␈αlinguistics␈αrequires
␈↓ ↓H␈↓computer␈α
facilities,␈α
but␈α
these␈α
are␈α
available␈α
in␈α∞a␈α
number␈α
of␈α
places.␈α
 Such␈α
a␈α
program␈α∞is␈α
especially
␈↓ ↓H␈↓suited␈αfor␈αa␈αtemporary␈αe≥ort␈α
such␈αas␈αis␈αplanned␈αby␈α
the␈αSloan␈αFoundation.␈α Three␈αyear␈α
fellowships
␈↓ ↓H␈↓seem the most appropriate.

␈↓ ↓H␈↓3.␈α⊗It␈α⊗would␈α⊗be␈α⊗worthwhile␈α⊗to␈α∃equip␈α⊗more␈α⊗computer␈α⊗science␈α⊗and␈α⊗computational␈α∃linguistics
␈↓ ↓H␈↓departments␈α
with␈α
suitable␈α
computer␈α
facilities.␈α
 Ten␈α
years␈α
ago,␈α
facilities␈α
suitable␈α
for␈α
AI␈α
work␈αcost
␈↓ ↓H␈↓more␈α
than␈α
$1,000,000,␈α
but␈α
now␈α
$200,000␈α
will␈α
buy␈α
an␈α
excellent␈α
facility␈α
for␈α
a␈α
moderate␈αsized␈α
research
␈↓ ↓H␈↓group.  Some matching might be expected.

␈↓ ↓H␈↓4.␈α⊂Projects␈α⊂intermediate␈α⊂in␈α⊂size␈α⊂between␈α⊂the␈α⊂individual␈α⊂grants␈α⊂favored␈α⊂by␈α⊂NSF␈α⊂and␈α⊃the␈α⊂large
␈↓ ↓H␈↓ARPA supported laboratories are viable and likely to be very productive.

␈↓ ↓H␈↓5. Endowed professorships, while not very exciting to a foundation, put a solid basis under a ≡eld.

␈↓ ↓H␈↓        Concerning the controversy with the Chomsky school, I have only the following comments:
␈↓ ↓H␈↓αDr. Robert Marston␈↓ ¬nApril 20, 1978␈↓ 
nPage 2␈↓ 


␈↓ ↓H␈↓1.␈α∞Their␈α
doctrinal␈α∞insistence␈α
on␈α∞the␈α∞primacy␈α
of␈α∞syntax␈α
has␈α∞always␈α
been␈α∞harmful,␈α∞although␈α
their
␈↓ ↓H␈↓speci≡c␈αcontributions␈α
to␈αsyntax␈α
have␈αbeen␈α
useful.␈α Barbara␈α
Partee,␈αwhose␈α
knowledge␈αI␈αrespect,␈α
tells
␈↓ ↓H␈↓me␈α⊃that␈α⊂this␈α⊃doctrine␈α⊂has␈α⊃been␈α⊂modi≡ed␈α⊃recently,␈α⊂but␈α⊃I␈α⊂gather␈α⊃it␈α⊂hasn't␈α⊃reached␈α⊂the␈α⊃stage␈α⊂of
␈↓ ↓H␈↓admitting␈α∃that␈α⊗either␈α∃the␈α⊗AI␈α∃approach␈α⊗or␈α∃Montague's␈α⊗philosophical␈α∃approach␈α⊗have␈α∃made
␈↓ ↓H␈↓contributions to the study of language.

␈↓ ↓H␈↓2.␈α∪From␈α∪my␈α∪own␈α∪point␈α∪of␈α∪view,␈α∪and␈α∩from␈α∪the␈α∪point␈α∪of␈α∪view␈α∪of␈α∪everyone␈α∪else␈α∪in␈α∩arti≡cial
␈↓ ↓H␈↓intelligence,␈α
the␈αmost␈α
important␈αquestions␈α
about␈α
language␈αconcern␈α
its␈αsemantics␈α
-␈αwhat␈α
information
␈↓ ↓H␈↓is␈α∞conveyed␈α∂by␈α∞discourse.␈α∞ I␈α∂think␈α∞that␈α∞language␈α∂is␈α∞secondary␈α∞to␈α∂thought;␈α∞a␈α∂linguistic␈α∞utterance,
␈↓ ↓H␈↓especially␈α∞a␈α
complicated␈α∞one,␈α
is␈α∞the␈α
result␈α∞of␈α∞thinking␈α
about␈α∞how␈α
to␈α∞express␈α
one's␈α∞thoughts␈α∞in␈α
a
␈↓ ↓H␈↓publically␈αunderstandable␈αway.␈α AI␈αwork␈αcon≡rms␈αthis␈αview,␈αbecause␈αthis␈αis␈αthe␈αproblem␈αfaced␈αby
␈↓ ↓H␈↓a␈α⊃would-be␈α⊃robot␈α⊂designer,␈α⊃and␈α⊃the␈α⊂experiments␈α⊃on␈α⊃teaching␈α⊂apes␈α⊃to␈α⊃communicate␈α⊃con≡rm␈α⊂it,
␈↓ ↓H␈↓because apes, who do not ordinarily communicate facts to one another, can be taught to do so.

␈↓ ↓H␈↓3.␈α∞I␈α∞have␈α∞an␈α∂aversion␈α∞to␈α∞involving␈α∞funding␈α∞agencies␈α∂in␈α∞scienti≡c␈α∞controversies.␈α∞ It␈α∞heats␈α∂up␈α∞the
␈↓ ↓H␈↓arguments,␈αwastes␈αtime,␈αand␈αthere␈αis␈αno␈αguarantee␈αthat␈αthe␈αbest␈αwork␈αwill␈αwin.␈α Therefore,␈αin␈αthe
␈↓ ↓H␈↓face␈αof␈αcontroversy,␈αsolutions␈αthat␈αallow␈αyoung␈αscientists␈αto␈αvote␈αwith␈αtheir␈αfeet,␈αsuch␈αas␈αfellowship
␈↓ ↓H␈↓programs,␈α∞have␈α∞much␈α∞to␈α∞recommend␈α∞them.␈α∞ However,␈α∞I␈α∞doubt␈α∞that␈α∞this␈α∞controversy␈α∞is␈α∞either␈α
so
␈↓ ↓H␈↓≡erce or so pervasive as to require con≡ning support to such programs.

␈↓ ↓H␈↓        The␈α⊂above␈α⊃should␈α⊂be␈α⊂taken␈α⊃as␈α⊂a␈α⊂somewhat␈α⊃casual␈α⊂statement␈α⊂of␈α⊃opinion␈α⊂and␈α⊂not␈α⊃as␈α⊂an
␈↓ ↓H␈↓attempt at a de≡nitive presentation of the "AI point of view".


␈↓ ↓H␈↓Sincerely,



␈↓ ↓H␈↓John McCarthy
␈↓ ↓H␈↓Director
␈↓ ↓H␈↓Professor of Computer Science