/LMAR=0/XLINE=3/FONT#0=BAXL30/FONT#1=BAXM30/FONT#2=BASB30/FONT#3=SUB/FONT#4=SUP/FONT#5=BASL35/FONT#6=NGR25/FONT#7=MATH30/FONT#8=FIX25/FONT#9=GRKB30
␈↓ α∧␈↓␈↓ u1


␈↓ α∧␈↓α␈↓ βONOTES ON RESEARCH TOPICS, ESPECIALLY FOR THESES

␈↓ α∧␈↓␈↓ αTI␈α
am␈α
prepared␈α
to␈α
supervise␈α
research␈α
in␈α
the␈α
following␈α
topics,␈α
although␈α
I␈α
am␈α
also␈α
willing␈α
to
␈↓ α∧␈↓consider␈α∞other␈α∞topics␈α∞proposed␈α∞by␈α∞students.␈α∞ Deciding␈α
which␈α∞of␈α∞these␈α∞topics␈α∞would␈α∞lead␈α∞to␈α
PhD
␈↓ α∧␈↓theses␈α
will␈αrequire␈α
work␈αby␈α
the␈αstudent.␈α
 In␈αmany␈α
of␈αthem,␈α
the␈αstudent␈α
would␈αhave␈α
to␈αcontribute
␈↓ α∧␈↓substantially to the definition of the problem as well as to finding a solution.

␈↓ α∧␈↓SYSTEM PROGRAMMING

␈↓ α∧␈↓␈↓ αTIt␈αwould␈α
be␈αhard␈α
to␈αpersuade␈αme␈α
to␈αsupervise␈α
the␈αdesign␈αof␈α
a␈αnew␈α
programming␈αlanguage
␈↓ α∧␈↓or␈αdesign␈αan␈αoperating␈αsystem␈αas␈αa␈αthesis␈αproject.␈α There␈αwould␈αreally␈αhave␈αto␈αbe␈αa␈αnew␈αidea␈αand
␈↓ α∧␈↓not␈α∞merely␈α∂a␈α∞dissatisfaction␈α∞with␈α∂the␈α∞details␈α∂of␈α∞present␈α∞implementations.␈α∂ (Note␈α∞of␈α∂February,␈α∞16
␈↓ α∧␈↓1979:␈αI␈αhave␈αa␈αnew␈α
idea␈α-␈αsee␈αbelow).␈α Instead␈α
I␈αwould␈αprefer␈αto␈αsee␈α
research␈αin␈αnew␈αareas␈αsuch␈α
as
␈↓ α∧␈↓the following:

␈↓ α∧␈↓1.␈α∀Information␈α∃description.␈α∀ Many␈α∀files␈α∃exist␈α∀that␈α∃are␈α∀accessible␈α∀to␈α∃our␈α∀computer␈α∃over␈α∀the
␈↓ α∧␈↓ARPAnet␈α∞or␈α∞by␈α∞telephone.␈α∞ Develop␈α∞a␈α∞language␈α∞for␈α∞describing␈α∞a␈α∞class␈α∞of␈α∞such␈α∞files␈α∂that␈α∞would
␈↓ α∧␈↓permit␈α
a␈α
program␈α
to␈α
answer␈α∞a␈α
question␈α
by␈α
extracting␈α
the␈α∞information␈α
from␈α
such␈α
a␈α
file.␈α∞ Easy␈α
(I
␈↓ α∧␈↓hope)␈α⊃example:␈α⊃many␈α⊃ARPAnet␈α⊃sites␈α⊃have␈α⊃telephone␈α⊃directories␈α⊃in␈α⊃their␈α⊃computer␈α⊃as␈α⊃do␈α⊃we.
␈↓ α∧␈↓These␈α∞directories␈α∂all␈α∞have␈α∂approximately␈α∞the␈α∞same␈α∂information␈α∞but␈α∂have␈α∞different␈α∂file␈α∞formats.
␈↓ α∧␈↓Devise␈αa␈αway␈α
of␈αdescribing␈αsuch␈αfiles␈α
so␈αthat␈αa␈αsingle␈α
program␈αcan␈αfind␈αaddresses␈α
and␈αtelephone
␈↓ α∧␈↓numbers from any of the files using the description of that file.

␈↓ α∧␈↓␈↓ αTThe␈α
description␈α
should␈α
also␈α
permit␈α
putting␈αinformation␈α
into␈α
the␈α
file.␈α
 Preferably␈α
it␈αshould
␈↓ α∧␈↓be␈αpossible␈αto␈αdescribe␈αthe␈αfile␈αpartially,␈αso␈αthat␈αif␈αthe␈αM.I.T.␈αphone␈αfile␈αnormally␈αhas␈αsome␈αweird
␈↓ α∧␈↓entry␈αlike␈α
our␈αBAM␈αallocation,␈α
we␈αwouldn't␈αhave␈α
to␈αknow␈αabout␈α
it␈αin␈αorder␈α
to␈αread␈αor␈α
change␈αa
␈↓ α∧␈↓file or to create a new one.  In the last case, there would probably have to be a default given.

␈↓ α∧␈↓␈↓ αTIt␈α
is␈αimportant␈α
to␈αunderstand␈α
that␈αwe␈α
are␈αtalking␈α
about␈αdescribing␈α
existing␈αfiles␈α
that␈αother
␈↓ α∧␈↓people␈α∂have␈α∂created␈α∂according␈α∞to␈α∂their␈α∂own␈α∂tastes␈α∞and␈α∂not␈α∂a␈α∂shiny␈α∞new␈α∂file␈α∂system␈α∂that␈α∞would
␈↓ α∧␈↓conquer the world.

␈↓ α∧␈↓␈↓ αTMany␈αfiles␈αare␈αnot␈αdirectly␈αreadable␈αand␈αmodifiable.␈α Instead␈αthey␈αare␈αparts␈αof␈αinformation
␈↓ α∧␈↓systems␈αdesigned␈αfor␈αdirect␈αhuman␈αuse.␈α In␈αorder␈αthat␈αa␈αcomputer␈αcould␈αextract␈αinformation␈αfrom
␈↓ α∧␈↓or␈αput␈αinformation␈αinto␈αsuch␈αa␈αfile,␈αit␈αwould␈αneed␈αa␈αdescription␈αof␈αthe␈αsystem␈αof␈αinteraction.␈α One
␈↓ α∧␈↓such␈αfile,␈αaccessible␈α
over␈αthe␈αARPAnet␈α
is␈αthe␈αLawrence␈α
Berkeley␈αLab␈αfile␈α
of␈αcensus␈αdata.␈α
 A␈αfile
␈↓ α∧␈↓description␈α
system␈α
that␈α
would␈α
permit␈α
programs␈α
to␈α
answer␈α
questions␈α
about␈α
this␈α
file␈α
once␈α
it␈αhas␈α
been
␈↓ α∧␈↓described␈α∂would␈α∂be␈α∂a␈α⊂substantial␈α∂achievement.␈α∂ To␈α∂play␈α⊂fair,␈α∂the␈α∂description␈α∂system␈α⊂should␈α∂be
␈↓ α∧␈↓designed before the designer sees the manual for the LBL data base.

␈↓ α∧␈↓2.␈α∂(February␈α∂1979).␈α∞ Design␈α∂a␈α∂"history␈α∞oriented␈α∂programming␈α∂language".␈α∞ The␈α∂object␈α∂of␈α∂a␈α∞high
␈↓ α∧␈↓level␈αprogramming␈α
language␈αis␈αto␈α
allow␈αthe␈α
programmer␈αto␈αexpress␈α
his␈αprocedure␈α
in␈αterms␈αclose␈α
to
␈↓ α∧␈↓those␈αin␈αwhich␈αit␈αis␈αconvenient␈αfor␈αhim␈αto␈αthink.␈α We␈αoften␈αthink␈αin␈αterms␈αof␈αthe␈αpast,␈αe.g.␈α␈↓↓Do␈αnot
␈↓ α∧␈↓↓generate␈αcastling␈αas␈αa␈αlegal␈αmove␈αif␈αthe␈αking␈α
or␈αrook␈αhas␈αmoved␈↓.␈α ␈↓↓If␈αthe␈αdriver␈αhas␈αhad␈α
5␈αspeeding
␈↓ α∧␈↓↓convictions␈αin␈α
the␈αpast␈α
2␈αyears,␈αstart␈α
proceedings␈αto␈α
revoke␈αhis␈αlicense␈↓.␈α
 Many␈αother␈α
examples␈αcan
␈↓ α∧␈↓(and␈α∂should)␈α∂be␈α∂given␈α∂(by␈α∂the␈α∂student)␈α∂in␈α∂which␈α∂the␈α∂idea␈α∂of␈α∂the␈α∂program␈α∂involves␈α∂the␈α⊂use␈α∂of
␈↓ α∧␈↓information␈αabout␈αpast␈αevents.␈α These␈αevents␈αcan␈αhave␈αoccurred␈αin␈αthe␈αoutside␈αworld␈αor␈αthey␈αmay
␈↓ α∧␈↓be events occurring in the operation of the program itself.
␈↓ α∧␈↓␈↓ u2


␈↓ α∧␈↓␈↓ αTIn␈α∞all␈α
present␈α∞programming␈α
languages,␈α∞writing␈α
a␈α∞program␈α
depending␈α∞on␈α
the␈α∞past␈α
requires
␈↓ α∧␈↓the␈αprogrammer␈αto␈αdecide␈αwhat␈αinformation␈αhas␈αto␈αbe␈αcollected␈αand␈αhow␈αit␈αis␈αto␈αbe␈αstored.␈α Much
␈↓ α∧␈↓labor␈α
would␈α
be␈α
saved␈α
if␈α∞the␈α
program␈α
could␈α
refer␈α
directly␈α∞to␈α
past␈α
events,␈α
␈↓↓and␈α
the␈α∞compiler␈α
would
␈↓ α∧␈↓↓decide what information had to be obtained and saved and how it should be represented␈↓.

␈↓ α∧␈↓␈↓ αTOne␈αapproach␈αto␈αthe␈αproblem␈αwould␈αbe␈αto␈αconsider␈αa␈αvirtual␈αprogram␈αthat␈αstored␈αa␈αrecord
␈↓ α∧␈↓of␈α⊃all␈α⊃events␈α⊃that␈α⊃occurred␈α⊂-␈α⊃all␈α⊃input␈α⊃and␈α⊃output␈α⊃and␈α⊂all␈α⊃assignments␈α⊃-␈α⊃in␈α⊃a␈α⊃history␈α⊂array.
␈↓ α∧␈↓Programs␈α⊂could␈α∂refer␈α⊂to␈α∂this␈α⊂array,␈α∂but␈α⊂couldn't␈α∂change␈α⊂it.␈α∂ With␈α⊂present␈α∂memory␈α⊂sizes,␈α∂direct
␈↓ α∧␈↓implementation␈αof␈α
this␈αis␈α
impossible,␈αso␈αthe␈α
compiler␈αwould␈α
have␈αto␈α
decide␈αwhich␈αinformation␈α
was
␈↓ α∧␈↓actually␈α∞referred␈α∞to␈α∞and␈α
store␈α∞only␈α∞that.␈α∞ (To␈α
stretch␈α∞the␈α∞imagination␈α∞a␈α
bit,␈α∞recall␈α∞that␈α∞there␈α
are
␈↓ α∧␈↓6␈↓πx␈↓10␈↓∧23␈↓␈αmolecules␈αin␈αa␈αmole,␈αand␈αa␈αcomplete␈αhistory␈αof␈αthe␈αactions␈αof␈αthe␈αKL-10␈αwould␈αstore␈αonly
␈↓ α∧␈↓of the order of 10␈↓∧6␈↓ words per second.

␈↓ α∧␈↓3.␈αSuper␈αSETL.␈α SETL␈αis␈αJack␈αSchwartz's␈αset␈αoriented␈αprogramming␈αlanguage.␈α I␈αlike␈αthe␈αgoals␈α
of
␈↓ α∧␈↓SETL,␈αbecause␈αallowing␈αthe␈αprogrammer␈αto␈αwork␈αdirectly␈αwith␈αset␈αtheoretic␈αoperations␈αis␈αindeed␈α
a
␈↓ α∧␈↓powerful␈α
and␈αhigh␈α
level␈αtool.␈α
 However,␈αI␈α
think␈α
the␈αSETL␈α
project␈αis␈α
too␈αlimited␈α
in␈α
two␈αrespects.
␈↓ α∧␈↓First␈α
it␈αhas␈α
a␈α
standard␈αway␈α
of␈α
representing␈αsets,␈α
but␈αthe␈α
way␈α
sets␈αshould␈α
be␈α
represented␈αdepends
␈↓ α∧␈↓strongly␈α∂on␈α∞how␈α∂they␈α∂are␈α∞to␈α∂be␈α∂used.␈α∞ Sets␈α∂can␈α∂be␈α∞represented␈α∂as␈α∂bit␈α∞vectors,␈α∂lists,␈α∂hash␈α∞tables,
␈↓ α∧␈↓balanced␈α∞trees,␈α∞programs␈α∞for␈α∞testing␈α∂elements␈α∞of␈α∞larger␈α∞sets,␈α∞program␈α∞for␈α∂successively␈α∞generating
␈↓ α∧␈↓elements␈αof␈α
the␈αgiven␈αset,␈α
and␈αprograms␈αfor␈α
getting␈αelements␈αof␈α
the␈αset␈αby␈α
operations␈αon␈αother␈α
sets.
␈↓ α∧␈↓A␈α
high␈α
level␈α
system␈α
would␈α
provide␈α
for␈α
all␈α
these␈α
ways␈α
of␈α
representing␈α
sets,␈α
allow␈α
the␈α
program␈α
to
␈↓ α∧␈↓choose how particular sets should be represented or make the decision itself if asked.

␈↓ α∧␈↓␈↓ αTThe␈αsecond␈αissue␈αis␈αa␈αbit␈αvaguer.␈α Often␈αwhat␈αwe␈αneed␈αto␈αrepresent␈αis␈αinformation␈αabout␈αas
␈↓ α∧␈↓set␈α
rather␈αthan␈α
a␈α
full␈αdescription␈α
of␈α
the␈αset␈α
itself.␈α
 Deciding␈αwhat␈α
information␈α
and␈αhow␈α
it␈αmight␈α
be
␈↓ α∧␈↓represented will require the exercise of some inventiveness.

␈↓ α∧␈↓4. Implement core war. See COREWA[S77,JMC].

␈↓ α∧␈↓5. Business communication language.  See CBCL[f75,jmc].

␈↓ α∧␈↓THEORY OF ARTIFICIAL INTELLIGENCE

␈↓ α∧␈↓6. Minimal models of a collection of sentences

␈↓ α∧␈↓␈↓ αTSee me for an oral spiel, but first read minima[f75,jmc].

␈↓ α∧␈↓7.␈α∞Finitization.␈α∞ A␈α∂possible␈α∞scheme␈α∞for␈α∞showing␈α∂certain␈α∞sentences␈α∞satisfiable␈α∞by␈α∂translating␈α∞them
␈↓ α∧␈↓into␈α_theories␈α↔in␈α_which␈α_satisfiability␈α↔means␈α_satisfiability␈α↔in␈α_models␈α_of␈α↔a␈α_fixed␈α_size.␈α↔ See
␈↓ α∧␈↓finiti[f75,jmc].

␈↓ α∧␈↓8.␈α
Formalization␈αof␈α
knowledge.␈α
 Read␈α(McCarthy␈α
1959),␈α
(McCarthy␈αand␈α
Hayes␈α
1969),␈α(McCarthy
␈↓ α∧␈↓1977a),␈α∞(McCarthy␈α∞1977b)␈α
and␈α∞(McCarthy␈α∞1979)␈α∞and␈α
see␈α∞me.␈α∞ Also␈α∞see␈α
KNOW.ART[F75,JMC]
␈↓ α∧␈↓and KNOW[E78,JMC].

␈↓ α∧␈↓9. Formalization of concepts.  Read same references as above and see me.

␈↓ α∧␈↓10.␈α
Formalization␈α
of␈α
blocks␈α
world.␈α
 Progress␈α
in␈α
problem␈α
solving␈α
requires␈α
programs␈α
that␈α
distinguish
␈↓ α∧␈↓between␈αthe␈αexistence␈α(in␈αa␈αmathematical␈αsense)␈αof␈αa␈αconfiguration␈αmeeting␈αthe␈αspecifications␈αand
␈↓ α∧␈↓␈↓ u3


␈↓ α∧␈↓its␈α
constructibility.␈α
 Thus␈α
an␈α
engineer␈α
normally␈α
designs␈α
a␈α
bridge␈α
and␈α
then␈α
figures␈α
out␈α
how␈αto␈α
build
␈↓ α∧␈↓it,␈α
e.g.␈α
decides␈α
where␈α
to␈α
build␈α
scaffolding␈α∞first.␈α
 The␈α
problem␈α
solving␈α
systems␈α
that␈α
I␈α∞know␈α
about
␈↓ α∧␈↓design␈α
by␈α
plannning␈α
to␈α
build,␈α
and␈α
this␈α
is␈α
too␈α
limited.␈α
 There␈α
is␈α
an␈α
informal␈α
writeup␈α
of␈α∞some␈α
of
␈↓ α∧␈↓these considerations, but see me first.

␈↓ α∧␈↓11.␈α∞Formalization␈α
of␈α∞concurrent␈α
and␈α∞continuous␈α
action.␈α∞ A␈α
key␈α∞barrier␈α
to␈α∞progress␈α
in␈α∞AI␈α∞is␈α
that
␈↓ α∧␈↓current␈αsystems␈αdeal␈αmainly␈αwith␈α
discrete␈αevents␈αthat␈αresult␈αin␈α
a␈αdefinite␈αnew␈αsituation␈αrather␈α
than
␈↓ α∧␈↓with␈αevents␈αoccurring␈αconcurrently␈αand␈αcontinuously.␈α There␈αis␈αsome␈αrelation␈αwith␈αthe␈αformalisms
␈↓ α∧␈↓for␈α⊃parallel␈α⊃programming,␈α⊂but␈α⊃AI␈α⊃must␈α⊂emphasize␈α⊃the␈α⊃handling␈α⊂of␈α⊃partial␈α⊃information␈α⊂about
␈↓ α∧␈↓systems.  There is a chess problem in Berliner's thesis that provides an entry.

␈↓ α∧␈↓MATHEMATICAL THEORY OF COMPUTATION

␈↓ α∧␈↓12. Proving intensional properties of programs

␈↓ α∧␈↓␈↓ αTA␈αproperty␈αof␈αa␈αprogram␈αis␈αextensional␈αif␈αit␈αis␈αa␈αproperty␈αof␈αthe␈αfunction␈αcomputed␈αby␈αthe
␈↓ α∧␈↓program,␈αi.e.␈αa␈αproperty␈αof␈αthe␈αinput-output␈αrelation.␈α Intensional␈αproperties␈αinclude␈αthe␈αtime␈αand
␈↓ α∧␈↓space␈αused,␈αthe␈αnumber␈αof␈αrecursive␈αcalls,␈αetc.␈α Our␈αapproach␈αto␈αintensional␈αproperties␈αis␈αto␈αmake
␈↓ α∧␈↓them␈αextensional␈αproperties␈αof␈αrelated␈αor␈α"derived"␈αprograms.␈α Ben␈αMoszkowski␈αhas␈αdone␈αsome␈αof
␈↓ α∧␈↓this,␈αand␈αsome␈αtechniques␈α
are␈αgiven␈αin␈αMcCarthy␈α
and␈αTalcott␈αLISP:␈αProgramming␈α
and␈αProving.
␈↓ α∧␈↓We␈α∩have␈α∩concentrated␈α∩on␈α⊃functional␈α∩programs.␈α∩ When␈α∩the␈α∩program␈α⊃is␈α∩the␈α∩fixed␈α∩point␈α∩of␈α⊃a
␈↓ α∧␈↓computable␈αfunctional,␈αthen␈αsome␈αintensional␈αproperties␈αof␈αthe␈αprogram␈αare␈αextensional␈αproperties
␈↓ α∧␈↓of␈α
the␈α
functional.␈α
 Which␈α
properties␈α
are␈α
extensional␈α
properties␈α
of␈α
the␈α
functional␈α
remain␈α
to␈α
be␈α
seen.
␈↓ α∧␈↓Conjecture:␈α⊃If␈α∩the␈α⊃function␈α⊃is␈α∩strict,␈α⊃then␈α⊃the␈α∩number␈α⊃of␈α⊃recursions␈α∩required␈α⊃to␈α∩evaluate␈α⊃the
␈↓ α∧␈↓function␈α∩for␈α⊃given␈α∩arguments␈α∩is␈α⊃an␈α∩extensional␈α∩property␈α⊃of␈α∩the␈α∩program.␈α⊃ Indeed␈α∩the␈α∩set␈α⊃of
␈↓ α∧␈↓evaluations␈αshould␈αbe␈αan␈αextensional␈αproperty␈αas␈αshould␈αa␈αcertain␈αevaluation␈αtree.␈α The␈αsequence
␈↓ α∧␈↓of evaluations will not be an extensional property of the functional.

␈↓ α∧␈↓13. Blobs

␈↓ α∧␈↓␈↓ αTA␈αblob␈αis␈αa␈αpiece␈αof␈αflowchart␈αwith␈αmany␈αentries␈αand␈αexits.␈α They␈αcan␈αbe␈αglued␈αtogether␈αto
␈↓ α∧␈↓make␈αlarger␈αblobs␈αand␈α
are␈αcharacterized␈αby␈αcertain␈αfunctions.␈α
 It␈αwould␈αbe␈αinteresting␈αto␈α
know␈αif
␈↓ α∧␈↓the␈α∂Yanov␈α⊂decision␈α∂procedures␈α∂are␈α⊂consequences␈α∂of␈α∂functional␈α⊂manipulation␈α∂of␈α⊂blob␈α∂functions.
␈↓ α∧␈↓See␈αthe␈αfile␈αBLOB[W76,JMC]␈α19-Feb-77␈αTHE␈αBLOB␈αFUNCTIONS␈αOF␈αFLOW␈αCHARTS␈α
for
␈↓ α∧␈↓more information.

␈↓ α∧␈↓14. First order theory of "cons should not evaluate its arguments"

␈↓ α∧␈↓␈↓ αTIn␈α∩a␈α∩well-known␈α∩report␈α⊃Daniel␈α∩Friedman␈α∩and␈α∩David␈α⊃Wise␈α∩argue␈α∩that␈α∩cons␈α∩should␈α⊃be
␈↓ α∧␈↓programmed␈α⊂as␈α⊂an␈α⊂FSUBR␈α⊂that␈α⊂merely␈α⊂conses␈α⊂its␈α⊂formal␈α⊂arguments␈α⊂without␈α⊂evaluating␈α∂them.
␈↓ α∧␈↓This␈α⊃enlarges␈α⊃the␈α⊃LISP␈α⊃data␈α⊃structures␈α⊃to␈α⊂include␈α⊃objects␈α⊃like␈α⊃the␈α⊃list␈α⊃of␈α⊃all␈α⊃integers.␈α⊂ Their
␈↓ α∧␈↓evaluator␈αwon't␈αevaluate␈α
it␈αwhen␈αit␈αoccurs␈α
in␈αa␈αcomputation␈αbut␈α
merely␈αexpands␈αas␈αrequired␈α
when
␈↓ α∧␈↓tests␈α
or␈α
printing␈αmake␈α
it␈α
necessary.␈α
 Friedman␈αand␈α
Wise␈α
suggested␈α
to␈αme␈α
that␈α
part␈α
of␈αthe␈α
extension
␈↓ α∧␈↓required␈αto␈αtreat␈αthese␈αobjects␈αlogically␈αis␈αto␈αmake␈α␈↓↓cons(␈↓πT␈↓↓,␈↓πT␈↓↓)␈α≠␈α␈↓πT␈↓↓,␈αcons(␈↓A␈↓↓,␈↓πT␈↓↓)␈α≠␈α␈↓πT␈↓↓␈↓.␈α This␈αisn't␈αall,
␈↓ α∧␈↓by␈α∂any␈α∞means.␈α∂ Obviously,␈α∂the␈α∞S-expressions␈α∂have␈α∞to␈α∂be␈α∂extended␈α∞to␈α∂include␈α∂some␈α∞computable
␈↓ α∧␈↓infinite␈α
S-expressions.␈α
 Can␈α
this␈αstill␈α
be␈α
done␈α
in␈αfirst␈α
order␈α
logic,␈α
and␈α
what␈αkind␈α
of␈α
a␈α
theory␈αdo␈α
we
␈↓ α∧␈↓get?  Is it good for anything or just a curiosity?  SKIP 2
␈↓ α∧␈↓␈↓ u4


␈↓ α∧␈↓αREFERENCES

␈↓ α∧␈↓␈↓αMcCarthy,␈α∂John␈α∂(1959)␈↓:␈α∂"Programs␈α∂with␈α∂Common␈α∂Sense".␈α∂␈↓↓Mechanisation␈α∂of␈α∂Thought␈α∂Processes,
␈↓ α∧␈↓↓Volume I␈↓.  London:HMSO.

␈↓ α∧␈↓␈↓αMcCarthy,␈α∩J.␈α∩and␈α∩Hayes,␈α⊃P.J.␈α∩(1969)␈↓:␈α∩"Some␈α∩Philosophical␈α⊃Problems␈α∩from␈α∩the␈α∩Standpoint␈α⊃of
␈↓ α∧␈↓Artificial␈α⊃Intelligence".␈α∩␈↓↓Machine␈α⊃Intelligence␈α∩4␈↓,␈α⊃pp.␈α⊃463-502␈α∩(eds␈α⊃Meltzer,␈α∩B.␈α⊃and␈α∩Michie,␈α⊃D.).
␈↓ α∧␈↓Edinburgh: Edinburgh University Press.

␈↓ α∧␈↓␈↓αMcCarthy,␈α∞John␈α∞(1977a)␈↓:␈α∞"Epistemological␈α∞Problems␈α∞of␈α∞Artificial␈α∞Intelligence",␈α∞in␈α∂Proceedings␈α∞of
␈↓ α∧␈↓Fifth␈αInternational␈αConference␈αon␈αArtificial␈αIntelligence,␈αavailable␈αfrom␈αDepartment␈αof␈αComputer
␈↓ α∧␈↓Science, Carnegie-Mellon University, Pittsburgh, PA 15213, USA.

␈↓ α∧␈↓␈↓αMcCarthy,␈αJohn␈↓␈α(1977b),␈α␈↓↓Circumscription␈α
-␈αA␈αWay␈αof␈α
Jumping␈αto␈αConclusions␈↓,␈αStanford␈α
Artificial
␈↓ α∧␈↓Intelligence Laboratory, (to be published).

␈↓ α∧␈↓␈↓αMcCarthy,␈α∪John␈α∩(1979)␈↓:␈α∪"First␈α∩Order␈α∪Theories␈α∩of␈α∪Individual␈α∩Concepts␈α∪and␈α∪Propositions",␈α∩in
␈↓ α∧␈↓Michie,␈α≥Donald␈α≡(ed.)␈α≥␈↓↓Machine␈α≥Intelligence␈α≡10␈↓,␈α≥also␈α≥available␈α≡on␈α≥disk␈α≥at␈α≡SU-AI␈α≥as
␈↓ α∧␈↓CONCEP[E76,JMC].

␈↓ α∧␈↓␈↓αMoore, Robert (1979)␈↓: his M.I.T. PhD thesis.

␈↓ α∧␈↓This file is THESIS[F75,JMC] at SU-AI.

␈↓ α∧␈↓John McCarthy␈↓ 
:March 11, 1979