philosophy and computer science - timothy r. colburn
TRANSCRIPT
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
1/256
Phi losophy
and
Compute r
Science
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
2/256
E X P L O R A T I O N S IN P H I L O S O P H Y
J am e s H.
Fetzer,
Ser ies
Editor
Aesthetics
James
W .
M a n n s
S o c i a l
Philosophy
Gera ld F. Gaus
Modern Political Philosophy
Richard
Hude lson
Philosophy
a n d Computer
S c ienc e
Timothy R. Colburn
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
3/256
\ \ - T C
E
X
P
L
0
R
A
T
I
0
N
s
i n p h i I o s o p h y
Phi losophy
and
Compute r Science
Timothy
R .
Colburn
James
H . Fetzer, Series Editor
9 \ / L E . Sharpe
Armonk, New
York
L o n d o n ,
Eng l and
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
4/256
Copyr igh t
2000
by M .
E. Sharpe,
Inc.
All rights
reserved.
N o part of this book m a y be reproduced in a ny form
wi thou t wr it ten
permission
from th e publisher,
M .
E. Sharpe, Inc. ,
8 0 B us in es s P a rk D r iv e , A rm o n k , N e w Y o rk 10504.
Library
of Congress
Ca ta loging-in -Pub l i ca t ion Da ta
Colburn, Timothy
R.,
1 9 5 2 -
Phi losophy and computer
science
/T imothy
R.
Colbtun.
p. cm .
- ~
(Explorat ions
in
ph i losophy)
Inc ludes bibl iographical references and index.
I S B N l -56324-990-1 (h e :
alk.
paper).
ISBN
1-56324-991-X (pbk. :
alk.
paper)
1 .
Computer
s c ie n c e P h i lo s o p h y .
2 . Art i f ic ia l intell igence.
I. Title. I I . Series
QA76. l67.C65
1999
0 0 4 D C 2 l
99-28936
CIP
Printed in
the
United
S t a t e s of
Amer ica
The
paper
used
in
this
publ icat ion
m eets the
min imum
requirements
of
Amer ican
Nat iona l Standard
fo r
Informat ion
Sciences
Pennanence of
Paper
fo r Printed Library Materia ls,
A N SI Z
39 .48- I984 .
BMc)098765432
BMp)1098765432
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
5/256
T o Car r i e , for
l ead ing
m e b ack to th e a c a d em y ,
a n d
to
J i m , fo r
lighting
th e
way.
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
6/256
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
7/256
Contents
Series Preface ix
Acknowledgments
xi
1 Introduction
3
Part I Philosophical Foundations
of
Artificial
Intelligence 1 1
2 The Definition and
S c o p e
of
AI
1 3
3 AI and the History of
Philosophy
1 9
4
AI
and the Rise of Contemporary Science and Philosophy 4 1
Part
II
The New Encounter of
S c i e n c e and
Philosophy 5 1
5 AI and Logic 53
6
Models of the
Mind
6 8
7
Models of Reasoning 87
8 The Naturalization of Epistemology 1 0 5
vii
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
8/256
Part
III The Philosophy
of Computer S c i e n c e
1 2 7
9 Computer Science and Mathematics
1 0 Two Views of Computer Science
1 1
Abstraction in
Computer
Science
1 2
Software,
Abstraction,
and Ontology
Notes
Bibliography
Index
1 2 9
1 5 3
1 7 4
1 9 s
2 1 1
2 2 1
233
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
9/256
Series
Preface
T h e
series, Exp lo ra t ions
in
Ph i l osophy, is
i n t e nde d
to p ro v id e u n d e rgra d -
uates
a n d
o the r readers
with quality introductions
n o t o n ly
to th e
princi-
p a l
areas
of
ph i l os ophy ,
including
traditional
t op ics o f
investigation~
epis temology,
eth ics,
a n d
so c ia l a nd political p h i l o s o p h y b u t also to
c o n t e m p o r a r y subjec ts of impor tance--such
as
c o m p u t e r science, f em-
inist s tu die s, a n d
th e
eth ics
o f re p ro d u c tio n .
In each case, th e ed i to rs
h a v e chosen au t h o rs w h o co u ld n o t
o n ly e xp la in
th e cen t r a l p rob l em s
encoun tered within their re s pe c tiv e d o m a in s but w h o cou ld
also
propose
p r omis i n g
so lu t ions
to those prob lems, i n c lu d ing n o v e l a pp r oa c hes a n d
original
ana lyses .
T h e
present
v o l um e, P h i l os ophy
a n d
C o m p u t e r
Science,
p ro v i de s
a
lu -
c id
a n d
stimulating introduction to s om e
of th e
m o s t
f a sc i n a t i n g i deas ex-
p lo red in
c o n t e m p o r a r y
p h i l o so p h y a n d science, w h i ch
co nce r n
th e scope
a n d
limits
of c om pu t e rs a n d their
powers .
T h e a u t h o r , Timothy C olb u rn ,
w h o
is both
a
ph i losopher
a n d
a
c o m p u t e r s c ie n t is t, supp l ies
a
f r a m e w o r k
that is historical
a n d
ana l y t i ca l ,
exp la i n i n g
th e p h ilo s o p h ic a l o rig in s of
m a n y p ro ble m s
that a re raised
by
m o d e r n
c o m p u t i n g
mach i ne s . A n y o n e
with a n i n te res t
in artificial
in t e llig en c e , rea s on in g
by
mach i ne s , a n d th e
t heo r e t i ca l
f o u n d a t i o n s of c o m p u t e r science will
find
it re w ard i ng .
J a m es
H .
Fetzer
ix
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
10/256
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
11/256
A c k n o w l e d g m e n t s
Th is b o o k w o u l d not
h a v e
been possible without th e support a n d en -
c ou ra g em en t of m y
brilliant
m e n t o r a n d
f r iend ,
Jim Fetzer. H e
p r ov i d ed
m e with th e
s em i n a l
i deas
that
first began to un i te
m y
distinct i n te l lec-
tual
l i ves in p h i l o so p h y a n d c o m p u t e r science, a n d he offered m e myriad
opportunities to
express
mysel f . I c a n n o t
thank
h im enough .
I w o u l d
also
l ike to thank Kluwer A c a d e m ic P u b lis h e rs
for
permiss ion
to use
pa r t s
of m y p ap e r s a p p e a r i n g in th e journal
M i n d s
a n d M a ch in es ,
a n d
fo r
permiss ion
to use
part
of
m y contribution
to
th e v o lu m e
Program
Ver ifica t ion :
F u n d a m e n t a l
Issues
in C o m p u t e r
Sc ience
(1993) .
T h a n k s
a re a lso d ue th e e d ito rs of The Monist: An
International
Qu ar te r l y of
G e ne r a l
P h i l o so p h i ca l
Inquiry,
for permiss ion to i nc lude part of
m y c on -
tribution
to th e J a n u a r y , 1 9 99 , is su e
o n
p h i l o so p h y of c o m p u t e r sc ience.
xi
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
12/256
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
13/256
Philosophy
and
Computer
Science
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
14/256
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
15/256
. _ . _ ]_ ..._
Introduction
P h i l o s o p h y
a n d
c o m p u t e r
science isnt
that
a rather
o dd c o m bin a-
tion? Such
is th e typical cocktail-party
response
w h e n
l ea r n i n g
of m y
acade mi c training
in
th e d is c ip lin e Socra tes
ca l led
the
l ove
of w i s d o m
a n d
m y
s ubs eq uen t
i m m e r s a l
in
th e wor ld of bytes,
p ro g ram s ,
sys tems
ana l ys i s ,
a n d g o v e r n m e n t cont rac ts . A n d such might be th e reac t ion to
th e title o f
this
book. B u t despite it s
cloistered
reputat ion and it s lit-
erary ,
as
oppos ed
to
techno log ica l ,
image,
th e
tradition
of
ph i l osoph i ca l
i n v es t ig a t i on ,
as
a ll of
us
w h o ha ve been sed uced
by it k n o w ,
ha s n o turf
limits. While fe w
but
th e truly p repa red ve n t u re into
ph i1osophy s
h a r d -
core inner ci rc le of ep is temo logy,
m et a phy s i c s , (meta)e th ics , a n d logic,
literally anything is fair ph i l osoph i ca l
g a m e
in
th e ou te r circ le
in
w h i c h
m o s t
of
us
exis t . And so
we
h a v e th e ph i l osophy ofs : ph i l osophy of
sci-
ence, p h i l o so p h y of
art,
of l anguage, ed uc a t i on . S om e of th e ph i l osophy
o fs
even
ha ve
n a m e s
befitting their integration into
vital
areas of m o d -
e rn
socie ty,
for
e xa m ple , m e d ica l e th ics a nd
e n v i r o n m e n t a l
eth ics ,
w h i ch
we ca n say a re sho r te r namesfor th e ph i l osoph ies of
e th i ca l dec is ions in
m e d ic in e a nd ecology. O n e o f th e a i ms of this b o o k is to m a k e a n ear ly
contribution to
a
n a s c en t p h i l o so p h y of c o m p u t e r science.
Which is
n o t
to say that there ha s
n o t been
a v a s t a m o u n t of work
d o n e
w h i ch
ca n
be descr ibed
as
th e cross-d isc ip l i na ry en c oun t e r of ph i l osophy
with c o m p u t e r science. Desp i te th e
typical
cocktail-party
reac t ion
to
this
c o m b i n a t i o n , th e
so lu t ions
to
m a n y
p rob l em s in c o m pu te r sc ience
ha ve benefi ted f rom
what
we might
ca l l appl ied
ph i l os ophy .
Fo r
exa m p l e ,
h a r d w a re lo gic
gate design
wou ld n o t
be
possib le
without
b o o le a n a lg eb ra ,
deve loped
by
th e n in e teen th -cen tu r y
m a t h e m a t i c i a n
George
B oo l e ,
whose
w o r k
he lped
la y th e f o u n d a t i o n
fo r
m o d e r n
logic. Later work
in logic,
particularly
th e d e v e l o p m e n t of pred ica te
ca lcu lus by
Gottlob Fre ge , h as
been
d ra w n u p on e x te n s iv e ly by
researchers in
sof tware
eng ineer ing
w h o
desire
a formal
l anguage for c o m p ute r p ro gra m semant i cs . Pr ed i ca te
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
16/256
4
CHAPTER
1
calcu lus is
also
th e formal m o d e l used
by
m a n y of
those
w h o
i m p l e m e n t
a u t o m a t e d reason ing systems for
m e c h a n i c a l t heorem
p r ov i n g . These
t heorem
p r o v i n g techn iques ha ve even
fo rm ed th e
basis fo r
a sty le
of
genera l -purpose
c o m p u te r p ro g ra m m i n g cal led logic p rog ra m m i n g .
Fur the r mor e , th e app l i ca t i on of ph i l osoph i ca l m e t h o d s to c o m p u t e r
s cie nc e is not limited to
those in
logic. T h e s t u d y of eth ics , for
e x a m -
ple,
has fo un d bro ad
a pp l i ca t i on to compu te r - r e l a ted
issues
of
p r i vacy ,
secur i ty , a nd la w. While these issues a re
n o t
regarded a s ge r mane to
th e
science
of c o m p u t i n g
p e r
s e ,
t he y h a v e
ar isen
directly a s a
resu l t of
th e
dras t i c
changes
socie ty
has
u nde r go ne
d ue
to
th e
ubiquity
a nd po w er
of
c om pu t e rs .
In 1990,
a
major U.S.
sof tware v e n d o r
a t t e m p t e d to o p e n l y
m a r k e t a large mailing
list
c om p i l ed f rom public sources, but wa s forced
to withdraw it w h e n th e pub l ic outcry
o v e r
i n vas ion of privacy bec a m e
to o
great . While th e
sca l i ng
b ack of th e
U.S.
Strategic
Defense Initiative
in th e 19 8 0s cou ld
be seen
a s a
response
to t e ch n i ca l feasibility ques-
t i ons , a
major
u nde r l y i ng
m o r a l co nce r n i w as
w h e t h e r
a
nation
ought
to
en t r us t
its secur i ty , to such
a
large ex ten t , to m a c h in es . And n o w ,
with
th e p erv a d in g i nfluence of th e Wor ld Wide Web,
socie ty
ha s
been
forced
to c o n f r o n t issues re ga rd i ng free do m a nd
decency
in
th e digital
wor l d .
Within
th e
field of la w, m a n y s tic k y e th ic a l ques t i ons re la ted to
c o m -
puters h a v e
ar isen: Is
un a u tho r i zed use of a c o m p u t e r f rom th e p r i v a c y
of ones o w n h o m e , without d a m a g i n g a ny files o r
p ro g ram s (i.e.,
h a c k -
i ng ) ,
th e same a s b re ak i ng a n d enter ing? C a n a u tho r s of p ro g ram s that
a re exper t s
in m ed ic in e o r la w be sued
fo r
m al p rac t i ce ? S hou l d
c o m -
puter
p ro g ram s be
copyr igh tab le , o r s ho uld they
be
f ree, l ike a ir? S h ou ld
p r o g r a m m e d trading be a l lowed
o n
th e s tock
exchange? A nsw e r s
to th e
l as t tw o
ques t i ons ,
a nd o th ers
l ike
it,
w o u l d
ha ve
s ig n ifica n t
effects
o n
th e c on d uc t of o u r ec on om y .
N o n e
of
these
ques t i ons
cou ld
h a v e
been
pred ic ted
a
m ere fe w decades
ago. T o d a y , it
wou ld be
difficult
to find
a
college
curriculum that
d id
n o t
inc lude ,
in
e i the r th e
c o m p u t e r
sc ience o r
th e
p h i l o so p h y d e p a r t m e n t ,
a course
entitled
C o m p u t e r s
a n d S o c ie ty ,
Values a n d T ec hn o lo gy , o r
th e
l ike.
But o u r inquiry
in
this b o o k goes bey on d th e application of philo-
soph ica l m e t h o d to specific issues l ike those just
m e n t i o n e d .
Our
inquiry
a t t e m p t s to seek
a
n ew en c oun t e r between p h i l o so p h y
a n d
sc ience
by
e x a m i n i n g
th e
ways t hey ca n
cha nge one a n o t h e r
in th e
c o n t e x t of one
of sciences
newes t
d isc ip l ines. Th is t ype of
inquiry
is
in
addition to
th e traditional concern of p h i l o so p h y of science, w h i ch ,
in its
ana lyses of
concep ts
lik e e x p la n a tio n , t heory , a n d th e on to log ica l
s ta tus
of in fer red
ent i t ies, is typically unaf fec ted
by
th e c o n t e n t of particular sc ient ific d is-
cover ies .
This
in s u la r n a t u re
of p h i lo s o p h i c a l c o n t e n t a n d
m e t h o d
is
be ing
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
17/256
INTRODUCTION 5
c h a lle n ge d b y w o r k
in
th e area of c o m p u t e r science k n o w n as artificial in-
te l l igence (AI),
particularly
in
th e traditional
ph i l osoph i ca l areas of logic,
ph i l osophy of m i n d , a n d e p is te m o lo gy .
Although even th e definition of th e fie ld of AI is fraught with philo-
s oph i c a l debate ,
genuine
ph i l osoph i ca l ques t i ons
c o m e
to th e fore a s re -
searchers
attempt to
m od e l h u m a n
i n te l l igence
in c om p ute r
p ro g ram s :
What is
th e
structure of
h u m a n knowledge
(so that we m a y represen t
it
in c o m pu te r
m e m o r y ) ? What is th e
process of
h u m a n
thought
(so
that we
m a y
m o d e l reason ing, l ea rn ing, a n d
creativity
in c om p u te r
p r o -
grams)?
I n te res t ing ly ,
whi le
AI
researchers
m u s t
ask
th e
s a m e
sor ts
of
cogn i t i ve ques t i ons
as ph i l osophers
d o , th e y u s ua lly agree with th e
per-
vas i ve
a ssu m ptio n, s ta te d by
Hobbes in
th e seventeenth centu ry ,
that
cognition is
computation,
a point
of
v iew
certainly n o t shared
by
a ll
ph i l osophers .
O n e
of th e f asc ina t ing
aspec ts
of AI is
its
co nce r n for both
c o m p u t e r
s c ie n tis ts a n d p h ilo s o p he rs .
A s
a subfield
of c o m p u t e r science,
it is a y o u n g d isc ip l i ne , but th e ques t i ons it
ra ises h av e been
th e o bje cts
of ph i l osoph i ca l
in v e s t ig a tio n fo r
centur ies . Th e re
is n o d ea r th of writing
o n this co nfl u e nce of concerns f rom
seeming l y
d i s p a ra t e d i sc ip lin e s , but
Part
I of
this
b o o k , Ph i l osoph i ca l
F o u n d a t i o n s
of Artificial I n te l l igence ,
p rov i d es a f resh treatment of their re l a t i onsh ip
by returning
to histori-
ca l ph i l osoph i ca l prob lems a n d
l ook i n g
a t th e m
in
th e
light
of h o w they
w o u l d
set th e stage
for
a n age
w h e n peop le w o u l d begin
p r o n o u n c i n g
c e rta in c o m p u te r
p r o g r a m s intelligent.
Th is retro-treatment of h is to r i ca l p h i l o so p h y is a n i n te res t ing exercise,
because it a l l ows us to
imag ine
a n
e p o ch a l sweep of ph i l osoph i ca l musings
t h rou gh th e ages,
in
w h i ch
concepts
of m i n d a n d re a so n in g a re first roo ted
in
th e
formal
o r
th e
d i v i n e , t hen
be co m e p o w ers
of
h u m a n i t y s
o w n
indi-
viduality, a n d finally a re
m an i fe s t
in
h u m a n i t y i s o w n ar t i f ac ts .
H o w e v e r
one
feels
a b o u t th e
inexorability
of this
sweep,
one thing
is
c lear : T h e
cons t ru c t i on of mo de l s of
m i n d
a n d reason ing ha ve today forced m a n y
ph i l osophers out of th e c lo is tered confines of their a
priori
w o r l ds .
O n e re as on
fo r th is emergence
is
cons is ten t with th e traditional role of
p h i l o so p h y as
a
gu id ing beacon ,
a
give r rather than rece i ve r of w i sdo m
in
its e n c o u n t e r with science. AI u n d e r w e n t
a
resurgence in th e
19 8 0s that
was primarily
th e resu l t
of
its s w itc h in g focus
f rom
sys tems
for d o i n g
mere ly
a u t o m a t e d
reason ing
to so-cal led
knowledge-based sys tems.
Prior
to this th e m o s t important
t heo r e t i ca l
tool for th e AI researcher
was
lo gic , a n d it wa s thought that
by
a u t o m a t i n g formal a n d wel l u n d e r s t o o d
pa t te r n s
of
i n ference, on e
cou ld
em u l a t e
h u m a n
intelligent b e h a v i o r
in
a c o m p u t e r p ro gra m . In so fa r as logic h a d been th e p ro v in c e of ph i loso-
phers
a n d m a t h e m a t i c i a n s , it wa s obv ious that p re v i o u s work
by
them
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
18/256
6
CHAPTER
1
h a d a bear ing
on
AI. However , m a n y AI
researchers began
to
bel ieve
that
th e role of reason ing in m a c h i n e i n te l l igence h a d
been
overemphas ized
a t
th e
expense
of know ledge . Early AI p rog ra m s
were
ve ry good at p r o v i n g
t h e o r e ms
in
first-order
pred ica te
logic,
but
such
p ro g ram s
p r o v e d
hugely
inef f ic ien t w h e n used to i m p l e m e n t nontrivial
sys tems fo r
r ea s on i n g in
specific
areas. It
became
obv i ous
that
m ore ef for t
s p e n t on acqu i r i ng
a n d
digitally represen t i ng know ledge in a specific
a rea ,
c om b i n ed with
even
a
minimal
reason ing m e c h a n i s m , wou ld p a y off with p ro g ram s m o r e accu-
rately
e m u l a t i n g h u m a n e xp e rtis e
in
that area, a n d th e
first
truly success-
ful
a p p lic a tio n s in
AI
became
k n o w n
as
exper t
systems. Such
p ro g ram s
were
sa id
to
be
know ledge-based because
m u c h
of
th e i n tense
ef for t
in
their d e v e l o p m e n t centered a ro u n d th e re p re s en ta tio n a n d
manipulation
of s p e c ific k n o w le d g e, a s oppos ed to th e
effic ient
m o d e l i n g of me ch an i sms
of pure reason .
T h u s AI
researchers
became
i n terested
in
th e c o n c e p t of k n ow l ed g e
as
wel l as logic, a n d it
seemed reasonab le
to
suppose
that t he y c ou ld le a rn
so m e t h i ng
f rom
p h ilo s o ph e rs , w h o
ha ve been
thinking
a b o u t
knowledge
for
a l ong
tim e . T h e
m o s t
important area of
AI directly
re la ted to epis-
t e m o l o gy became
k n o w n as
knowledge
represen ta t i on . But it
wa s c lear
that,
to truly emula te intelligent b e h av i o r , not o n l y mo de l s of knowledge
rep resenta t ion
but
also
mode ls of c om in g to kn ow were necessary . In o t h e r
w o r ds , AI
p ro g ram s
h a d to be
able
to l ea rn . So t he re a re severa l impor-
tant aspects of know ledge with w h i ch AI researchers a n d practitioners
m u s t be
c on c e r n ed .
This ques t ion of h o w
p h i l o so p h y
ca n
help us
d o
AI is
c ouc hed
in
th e
language of
interdisciplinary cooperation, in which o n e discipline per-
ha ps
serend ip i tous ly
benefits
a n o t h e r b y o ffe rin g re l e van t
work
a l r e a d y
d o n e o r a n insightful outlook
p r ev ious l y
unseen . That this re l a t i onsh ip
is ev en p o s s ib le b etw e e n AI a n d p h i l o so p h y is d ue to-the o ve r l ap of sub-
ject
matter: p h i l o so p h y is concerned
with
issues of
h u m a n knowledge
a n d
reason ing,
a n d AI is concerned with
mode l i ng
h u m a n k n ow le d ge a n d
reason ing .
But ph i l os ophy , in
its
genera l
role of
critically
ev a lua t i n g bel iefs, is
mo r e t h a n m ere l y
a
potential partner with AI. P e r h ap s th e
m o s t
v is ib le
role p h i l o so p h y
ha s
p layed
in
this re la t ionsh ip is that of
w at ch do g , in
w h i ch
it
del ineates
th e
limits,
a n d s om e tim e s
even
attempts to
des t roy
th e f o u nda t i o ns , of AI.
These cr i t iques
proceed by taking
to ta s k th e c la im
that
computation
is even
a n
appropr i a te m o d e l
fo r
h u m a n thought o r
consc iousness
in th e
first
p lace .
Their primary
focus
is
n o t
logic o r knowledge, but of m i n d .
Here th e
primary
ques t ion
is w h e th e r p h ilo s o p h y
ca n
tell AI
what
it
ca n
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
19/256
INTRODUCTION 7
d o. M a n y
ph i l osophers bel ieve th e
a n s w er to th is
ques t ion
is yes,
but
t he y
a re
la rgely
i gnored
by
AI
researchers
a n d
practitioners, because
th e
latters
focus
is
n o t
m i n d but lo gic a nd know ledge .
While
this ignor ing
is
d o n e
a t their per i l ,
it is
bec om i n g c lear
that a
s w i n g
in
th e
oppos i te
d i rec t ion ha s oc c u r red , to th e
point
where it is
c l a i m e d
that t echno log-
i ca l advances ,
espec ia l l y
those in c o m pu te r
science,
ca n shed
light
o n
traditional
ph i l osoph i ca l
prob lems.
To cla im this was
un th i n ka b le
within m a n y ph i l osoph i ca l
ci rc les
just
tw o
decades
ago,
and there
a re still those w h o will s tead fas t l y
resist
c o u n t e n a n c i n g th e possib i l i ty .
But
s ince
cer ta in
ques t i ons
traditionally
thought to be
p h i l o so p h i ca l su ch
as:
H o w
d o
we
c o m e
to k n o w
th ings?
What is th e structure of know ledge? What is th e
n a tu r e
of th e mind?-
a re n o w be ing asked
by
AI a n d cogn i t i ve science researchers
as
wel l , it is
i nev i tab le that
th e s e r es e a rc h e rs
will offer
answers
in
th e
techn i ca l t e rm s
with w h i ch
t hey a re familiar.
In shor t ,
th e
rapid
growth
in
th e
speed
a n d
comp le x i t y
of
c o m p u t i n g m a c h in es is tempting
peop le to put forth
mode ls of th e h u m a n m i n d
in
t e rm s of c o m p u t e r science. But w h y a re
c o m p u t e r science
mode ls so
t emp t i n g ? T o a n s w er th i s ,
it
he lps to discuss
so m e t h i ng a b o u t th e
relation
between
science
a n d
p h i l o so p h y
in
genera l .
Science
a n d
p h i l o so p h y a re
of ten
d is t i ngu ished
by pointing out that
sc ience seeks exp la n a t i on w h ile p h ilo s op h y seeks justification. T o a sk
what
causes th e t ides is a sc ient ific ques t ion ,
whi le
to
a sk what wou ld
cons t i tu te ade qu a te
grounds- fo r
be l iev ing
that
I s e e th e m o o n is
a
philo-
s oph i c a l
one .
What
rea so n d o yo u
ha ve fo r
b elie v in g X ?
is a typical
ques t ion
asked by
p h ilo s o p he rs , a n d th e
n a tu r e
of X d e t e rm i n es th e kind
of p h i l o so p h y
u nde r t ak e n .
For e x amp l e ,
What
rea son d o
yo u
h a v e for
be l iev ing
that
m e r c y
killing
is
w ro ng?
is a
ques t ion
for
n o r m a t i v e
eth ics ,
What reason
d o
yo u ha ve
fo r
be l iev ing in th e ex is tence of
d i se m b o d i e d
m i n d s ? is a
ques t ion for p h i l o so p h y of m i n d ,
a n d What
rea son d o
yo u
ha ve fo r b elie v in g that this
a r g u m e n t is
v a l i d ?
is a ques t ion
for
ph i l osoph i ca l logic. So p h i l o so p h y ha s been charac ter i zed as th e
critical
ev a lua t i on of
beliefs
t h rou gh th e
ana l ys i s
of concep ts
in a
given
area
of
i nqu i r y .
Of
course,
s cie nc e is also concerned with critically
ev a lua t i n g
bel iefs
a n d a n a l yz i n g
concepts .
But w h e n one l ooks a t th e k in d s of th ings X is
in
th e
ques t ions
a bo v e, o ne
not ices
that
n o n e
of
them
lend
themse l ves
to
e mp i r i ca l s tudy .
O n e
need n o t wi tness a c t u a l cases of
m e r c y
killing to
c o m e to
a conc lus ion
a b o u t
w h e t h e r it
o u g h t to
be
d o n e . By definition,
a
d i s em bod i ed m i n d is on e that c a n n o t
be subs ta n t i a ted
through p h ys i ca l
obs e r v a t i on . And th e
validity
o f a n
a r g u m e n t
form is n o t de t e rm i ne d
by
lo o kin g fo r
i ns tances of th e form
in
th e wor l d . So p h i l o so p h y ha s
also
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
20/256
8 CHAPTER 1
been charac ter i zed
as a
n on emp i r i ca l , o r a priori, d isc ip l i ne ,
in
distinct
c o n t r a s t
with science.
C o m p u t e r
science, be ing a sc ien ce,
wou ld
seem to
be d is t ingu ished
f rom
p h i l o so p h y just as a n y
o t h e r
science. But
c o m p u t e r sc ience
is u n i q u e
a m o n g th e sc iences
in
th e types of models it creates. In
seek ing
e x p l a n a -
t i ons, sc ience o fte n c o n stru c ts mo de l s to tes t hypo theses
fo r
e xp la i n i ng
p h e n o m e n a .
Fo r e xa m p le ,
it might
be
hypothes ized that th e p h e n o m e n o n
of th e northern
l ights
is caused by th e interaction of so la r
a t o m s
with th e
ear th s
m a g n etic fie ld .
To
test
this hypothes is ,
a
m o d e l of th e ear th
a n d
its
m agne t i c field cou ld
be
crea ted
in
a
l ab ora to ry ,
comp le te
with
a p p r o p r i a t e
magne ts
a n d
gaseous e lements .
T h e n ,
if,
u n d e r
th e
right
co nd i t i o ns , luminosity is
observed ,
th e hypothes is m a y
be
said to
be c on -
fir med .
This m o d e l , in th e
form
of
e x p e r i m e n t a l
a p p a r a t u s , is of
course
a p h ys i ca l
o b je c t, lik e
m a n y mode ls
built
a n d m a n i p u l a t e d in a n y of th e
natural
sc iences.
T h e mode ls
built
a n d
m a n ip u la te d in c om p u te r science,
however ,
a re n o t
p h ys i ca l
a t
al l .
C o m p u t e r
s cie nc e is
a
science concerned
with
th e
s t u d y of c o m p u t a -
tional processes. A c o m p u t a t i o n a l process
is
d is t i ngu ished f r om, say, a
c hem i c a l
o r e le c tr ic a l process,
in
that
it
is s tud ied in w a y s that ignore
its
p h ys i ca l
nature.1 Fo r e x amp l e , th e process
by
wh ich
a
ca rd p la ye r
ar ranges
ca rds in
h er h a nd , a n d
th e
process
by
w h i c h
a
c o m p u t e r sor ts
n a m e s
in
a c us t om er list, though t he y
share
nothing
in
c o m m o n p h ys i -
c ally , m a y
never the less
e m b o d y th e s a m e c o m p u t a t i o n a l
process .
T h e y
m a y , for e x amp l e , both
proceed by s c a n n i n g th e
i t ems to be
a r range d
one by one ,
d e t e r m i n i n g
th e p r o p e r place of each s c a n n ed item re la t i ve
to th e i t ems a l rea d y
scanned ,
a n d i nse r t ing it into that p lace , perha ps
n ec es s i ta t in g the
m o v i n g
of
p r ev ious l y
scanned
i t ems
to
m a k e
r o o m .
This
process
( k n o w n as
a n
insertion sor t
in
c o m p u t e r sc ience t e r ms)
ca n be
prec ise ly descr ibed
in a
f o r m a l l anguage
without talking
a b o u t
p l a y i n g
ca rds o r
se m i co ndu c t i ng
e lements .
W h e n so
descr ibed, on e
has a
c om pu-
tational
m o d e l
of th e
process
in th e form of a c o m p ute r p ro gra m . This
m o d e l ca n be
tes ted,
in a w a y
ana logous
to h o w
a hypo thes is
is tes ted
in
th e natural
sc iences,
b y e xe cu tin g th e program a n d o b se rv i ng
it s
be-
h av i o r . It
ca n a lso be rea son ed a b o u t a b s tra c tly , so
that we m a y
a n s w er
ques t i ons a b o u t
it,
such as,
a re
there o the r
processes w h i ch will ha ve
th e
same
effect
but
d o
it
m ore
effic ient ly?
Building
computational
mo de l s
a n d
a n s wer i n g these k i nds of ques t i ons form
a
large part of
what c o m p u t e r
scient is ts d o .
T h e exp los i ve
growth in
th e
n u m b e r
of c o m p u te r a p p lic a tio n s
in th e
las t severa l deca des has
s how n
that th e k in ds of rea l wor ld processes
ame nab l e to m o de lin g by c om p ute r a re l im i t less. Not o n l y
h a v e
tradi-
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
21/256
INTRODUCTION 9
tional
activities,
l ike
record
keeping,
i nves t ing, pub l i sh ing ,
a n d
ba n k in g ,
been
s imp ly conver ted to
con t r o l
by
c o m p u t a t i o n a l
mo de l s , but
who l e
n ew k i nds of activity ha ve been c rea ted that wou ld not be possib le
with-
out
such
mo de l s . These a re th e by-now-familiar virtual activities we
descr ibe in th e l anguage of
cyberspace:
e -m a i l, c h a t r o o m s , Web surfing,
on- l i ne s hopp i n g , Internet g a m i n g , a n d so on . But l o ng before
co m p u t e rs
c a m e to d o m i n a t e ev e ry d a y
l i fe,
c o m p u t a t i o n a l mo de l s were
em p l oy ed
to
descr ibe processes of a spec ia l s o rt, w h ic h h a v e exis ted a s l o ng a s
m o d e r n
H o m o sap iens
has
ex i s ted .
These
a re th e
processes
assoc ia ted with h u-
m a n re aso n in g
a nd kn ow le dge
organ iza t ion ,
a n d
computational
mo de l s
of
them
a re
th e concern
of
AI.
T h e s t u d y of th e na tu re of h u m a n
r e aso n i ng
a n d kn o wle dge ,
in
th e
form of
logic a n d e p is te m o lo gy ,
has , of c ou rs e, be en
a
focus of weste rn
p h i l o so p h y
s ince
Plato a n d Aristotle. However , not
until
th e
latter
part
of th e twentieth cen tu r y a n d th e a d ve nt of th e
digital
c o m p u t e r did
it
be-
come
possible to actually bui ld mode ls of reason ing that co n t a i ne d al leged
rep resenta t ions
of h u m a n k n ow le d ge . B e fo re that t ime ,
if yo u w an te d
to
s t u d y
h u m a n
rea so n in g o r th e
structure of
h u m a n
kn o wle dge , y ou
re -
m a i n e d for th e
m o s t
part
in
th e a priori wor ld of
ph i losophy,
utilizing
perha ps
a
d a t u m o r
tw o
f rom psycho logy . With co m p u t e rs , however ,
it
became possible to test ones
ep is temo log ica l
theory
if th e
theory
was
real izable
in
a
c o m p u t e r mo de l .
It therefore became reasonab le
to at
l east ask: C a n AI, a s a n
emp i r i ca l
d isc ip l i ne concerned with building a n d
obs e r v i n g mo de l s of h u m a n c o gn itiv e b eh a v io r, help us d o p h i l o so p h y?
If we
t ake
ser ious ly th e
charac ter iza t ion
of p h i l o so p h y given a bove ,
name l y , that it is
by
definition a
n o n e m p i r i c a l
d isc ip l i ne , t hen we m a y
re ga rd th e
ask ing
of
this
ques t ion
a s
a
ca tegory
mis take.
R od er i ck
Chis -
ho lm
is a
primary p r o p o n e n t
of this
v i ew:
Contemporary interest
in
th e nature
of
k n o w l e d g e pertains
not
only to
that
b r a n c h
of
p h i l o s o p h y
ca l led theory of know ledge o r
ep is temo logy ,
but
also
to th e fields of
information
t heo r y ,
artificial i n te l l igence, a n d
c o gn i t i ve sc ience. T h e
latter
d isc ip l ines
a re
not
alternatives
to th e
traditional theory
of
k n o w l e d g e
because they a re branc h e s of empirical sc ience a n d not of
p h i l o s o p h y .
Fo r
th e
m o s t
part,
th e
f ac ts
with
which
they
a re
c o n c e r n e d a re not
r e l e v a n t
to th e traditional p h i l o s o p h i c a l
ques t i ons .
However , m a n y c o n t e m p o r a r y researchers, both p h ilo so p he rs a n d c o m -
puter sc ient is ts , wou ld
t ake
issue with
th is
v i ew , finding it a m e t h o d o l o g -
i ca l d o g m a that m a y
h a v e
seemed
reasonab le
in
a n ear l ie r time
w h e n it
wa s
n o t
possib le to build digital mo de l s of
h u m a n
re a so n in g a n d
k n o w l -
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
22/256
1 0
CHAPTER
1
edge,
but
w h i ch
is
q ue stio n able a n d n a iv e
in th e
Information
Age.
This
is
a
genuine d ispu te , a n d m a y
ex t en d
to
what
s om e wou ld regard as
th e
v e ry fo un d a tio n s of ph i l osophy . But it is th e
case, fo r better
o r worse,
that
c o m p u t e r
s cie nc e is b e g i nn i ng to
a ffe ct h ow
we
d o
ph i losophy, a n d
Part II of th is book , The N ew E n c o u n t e r of Science
a n d
Ph i l osophy ,
descr ibes this reverse effect
in
deeper
deta i l , suggest ing
a
con t i nu ing
n ew
symbios is
in
th e fu ture.
While it is a n ew reve la t ion
fo r
some ph i l osoph i ca l thinking to be a f-
fected
by
th e running of c o m p u t e r p rog ra m s , i n d e p e n d e n t
ph i losophers
a re
by
n o m e a n s out of business. AI
is but a
subfield of c o m p u t e r sc ience,
a n d th e
cons ide ra t ion
of m i n d
a n d
reason notwithstanding, c o m p u t e r sci-
ence is r ipe
fo r
th e good o ld - fash ioned
ana l ys i s
that ph i l osophy p rov i d es .
T h u s , Part
III of
th is
b o o k , The P h i l o so p h y of C o m p u t e r Science,
a t-
temp ts
to
place
c o m p u t e r science within th e bro ad s pe ctru m of inquiry
that const i tu tes
science.
T h e concern here
is
to
d e a l
with th e inevitable
identity crises that c ro p u p
in
th e sel f- image of a n y a d oles ce nt, w h i ch
c o m p u t e r
science
certainly
is .
We will ask ph i l osoph i ca l
ques t ions l ike:
What
is
th e
relation
between
m a t h e m a t i c s
a n d
c o m p u t e r
science? Is
there
a
sense
in
w h i ch c o m p u t e r science
is
exper imen ta l
science?
Is
a
c o m p u t e r
p r o g r a m m e r
m ere l y
a
d a t a
w i za rd , o r
ca n she also
engage in information
model i ng? What
is
th e n a tu r e of
a bs t r a c t i on
in
c o m p u t e r science? What
a re th e o n to lo gic a l imp l i ca t i ons of
c o m p u t e r sc ience concepts?
In shor t ,
th is
b o o k is a b o u t th e p h ilo s o p h ic a l fo u n d a t io n s of
artificial
i n te l l igence
in
particular a n d c o m p u t e r science
in
genera l , a n d it is
a b o u t
th e con t r ib u t i ons that p h i l o so p h y a n d c o m p u t e r science
ca n
m a k e to each
o ther . It asks
where
c o m p u t e r science fits
in as a
science, h o w ph i l osophy
can guide c o m p u t e r science, a n d
h o w w o rk in
c o m p u t e r sc ience
ca n
s ha pe
ones
ph i l osophy . I
d o a dv ise
th e
r e ade r
that chapters 7 a n d
8 on
mo de l s
of
re as on in g a n d
th e naturalization of e pis te m o lo gy m a y i mp o se m o r e
d e m a n d s
u p o n
readers t h a n d o th e o the r chap te rs .
Still,
by th e conc lus ion
of
th is
b o o k , I hope that readers will be
suf f ic ien t ly
en l igh tened to begin
to a n s w er these ques t i ons fo r themse lves .
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
23/256
Part
I
hilosophical
Foundations
of
Artificial
Intelligence
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
24/256
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
25/256
_ . _ . 2 _
The Definition
and
Scope
of
AI
Since th e
first
tw o p a rts
of this b o o k
use
AI a s their
focus fo r e xa m in in g
th e relation between p h i l o so p h y a n d c o m p u t e r sc ience, I will o p e n this
part with a descr ip t ion of
w h a t
A1 is, a n d a
s h o r t
survey of th e k i n d s of
th ings that
AI p ro g ram s
do .
T h e
lure
a n d p rom i s e of c o m p u t e r science has a lways
been
th e abi l -
ity to m o d e l
digitally th e o bje c ts a n d processes w h i ch
o c c u r
in e v e r y d a y
l i fe.
W h e t h e r
it
be
b a n k
acco u n ts ,
baske tba l l
games,
o r s h op p in g m a l ls ,
th e digital m o d e l i n g of a n y entity is possible d ue to ever m o r e soph i s -
t i ca ted m e d i u m s of descr ip t ion ava i l ab le to
p r o g r a m m e r s
in th e form of
p r o g r a m m i n g languages a n d s o ftw a re d e v elo p m e n t e n v i r o n m e n t s . A s I
will
show
in ch ap t e r
1 1 , p r og ra m m e rs m o d e lin g
p o w e r
increases as their
ability to rep resent rea l -wor ld ent i t ies
in
so f tware
is facilitated,
a n d this
is
a c c om p l i s hed with ab s t rac t i on too ls .
T h e
less t he y
need
to
be
co nce r ne d
with
c o m p u te r-s p ec ific d e ta ils lik e m e m o ry m a n a g e m e n t
a n d
m a c h i n e
a d -
dress
manipulation,
th e
m o r e
t he y
ca n
co nce n t ra t e
on
rep resent ing
a n d
m a n a g i n g
rea l -wor ld
information a n d
objec ts .
At th e d a w n of th e digital age, a group of v i sio n a r y th in k e r s specu la ted
that
th e
r ea l world objects capab le of be in g m o d ele d in c lu d e th e intel-
l igent
r ea s on i n g processes
exh ib i ted
by
h u m a n s .
At
a
s m a l l confe rence
a t Dartmouth
in 1 9 5 6
these
th inkers
gathered
to
discuss th e
following
t h e me :
Eve ry aspect of l ea rn ing or a n y o th e r fe a tu re of i n te l l igence
can
in principle
b e
so
prec ise ly descr ibed t h a t a m a c h i n e can b e
m a d e
to
s i m -
ulate it.1 J o h n M c C a r t h y , one of th e at tendees,
c a m e
u p with th e term
artificial
i n te l l igence
to
descr ibe
th e field
c o m m i t t e d
to this
th em e, a n d
this
l a be l has s tuck . O n e w a y to
prec ise ly
desc r ibe i n te l l igence
is
by
c a r e fu lly c h a r a c te riz in g
it as a kind of s y m b o l p ro c es s in g go v e rn e d
by
strict ru les of s y m b o l manipulation. McCarthy real ized
that
in o r d e r to
tes t this
s ym bo l s ys te m hypothes is
he
needed
a p r o g r a m m i n g tool that
w o u l d a l low him to think
in
t e rms of
s ym b ols a n d
l ists of s y m bo l s ,
rather
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
26/256
1 4 CHAPTER
2
t h a n in t e rm s of bytes, r eg is te rs , a n d m e m o r y addresses . So he c rea ted
th e
list
p ro c e s sin g la n g u a ge
ca l led
Lisp
that con t i nues
as
th e
d o m i n a n t
AI research l anguage t od a y .
Marvin
M i n s k y ,
a n o t h e r of th e Dartmouth
con ference a t te nde e s , has
de fi ne d
AI as the sc ience of
m a k i n g m a ch in es d o th ings
that w o u l d re -
qu i re in t e llige nce if
d o n e
by men. 2
This definition
does not
presuppose
a
t he ory of h u m a n i n te l l igence such a s th e s ym bo l sys tem
hypo thes is ,
but
charac ter izes AI as be ing concerned
on l y
with m a c h i n e
behav io r
i n so fa r
as it m i m i c s h u m a n intelligent be ha v io r. T hu s we
ca n descr ibe tw o
dif-
fe ren t
a pp r oa c hes
to
AI,
one
that
a t t e m p t s
to
m a ke m a ch in es
think
l ike
h u m a n s ,
a n d
on e
that a t t e m p t s to
m a ke m a ch in es
a c t
l ike hum a n s . 3 T h e
f o r m e r
a p p r o a c h is t o d a y p laced within th e
interdisciplinary
field of cog-
nitive science,
in
w hic h th e o bje c ts of s t u d y a re th e
cogn i t i ve
processes
of a c t u a l h u m a n s
a n d
a n im a ls . T o tes t theor ies a b o u t these processes,
it
is of ten
e x p e d i e n t
to write p ro g ram s that
m o d e l
h u m a n
o r
a n i m a l
cog-
nitive arch i tec ture , a n d t hen to observe w h e t h e r th e p ro gra m s beha ve
(cogn i t i ve ly )
in th e
w a y
that
h u m a n s
a n d an i m a l s
do . T h e
latter
a p -
p r o a c h a ls o o bs erv e s th e cogn i t i ve b e h a v i o r of
p ro g ram s ,
but
ignores
th e
issue
of
w h e t h e r
AI faithfully
emula tes
th e
a c t u a l
structure
a n d
process
of
h u m a n reason ing, m ea s ur i n g
its
success
solely aga i ns t
th e b eh a v io r
of
a c o m p u t i n g artifact. This a p p r o a c h
ha s
th e
a d v a n t a g e
of
a v o i d i n g
both
con t rovers ies
o v e r
h o w h um an s think a n d
cr i t i c i sms
that mach i ne s will
ne v e r
be able
to
think l ike
h u m a n s .
But
there is a n o t h e r d im e n s io n a lo n g w h i ch
a pp r oa c hes
to AI
ca n be
d is t i ngu ished.
W h e t h e r
it is
thought o r b e h a v i o r that
is e m u l a t e d ,
s om e
a pp r oa c hes
to AI
d o
not use
h u m a n s
as their mo de l s , opting i n s t e ad for
th e
notion
of
a n
i d ea l
rational
agent .
Real iz ing
that
h u m a n
thought
a n d
b e h a v i o r a re both
myste r i ous l y
c o m ple x a nd
a t
t imes
unre l iab le , these
ap p r o ach e s m o d e l i n te l l igence
o n
a no t i on of i d ea l rationality.
T h e
re -
sulting sys tems
a re
j udged to be n o t h u m a n l i k e
in
their
b e h av i o r , but
agent l ike,
where
a n a g en t
is nothing m ore than a sys tem
that perce ives
a n d acts . O n e t ype of age n t that ha s
received
large a m o u n t s of
research
attention
in th e 1990s
is
th e sof tware agent, o r softbot.
Motivated by
th e
need to filter usefu l
knowledge
f rom e no r mo u s a m o u n t s of
information
ava i l ab le by
e lec t ron ic
mail a n d th e Wor ld Wide
Web, so f tware
agents
ha ve
s in ce been
app l ied in m a n y o the r
areas
including
e lec t ron ic
c o m -
merce
a n d
spacecra f t con t ro l .
W h a t e v e r
their a pp roa c h , AI practitioners a re taking a d v a n t a g e of
four decades
of research to p rod uc e sys tems within
a
wide
variety
of
a pp l i ca t i on s . These a pp l i ca t i on s
ha ve f o u nd their
w a y
into all
corners
of soc iety that a re
d e p e n d e n t u p o n t echno logy ,
including th e
military,
U
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
27/256
THE
DEFINITION
AND S C O P E
O F
AI
1 5
m ed i c i n e ,
science,
m a n u f a c t u r i n g ,
co mme r ce ,
a n d
fi nance .
T he k in d s
of
th ings that
AI
s ys te m s d o
mirror th e k in d s
o f ac t ions that
h u m a n s t ake in
their ca pa c i t y a s reasoners a n d d ec is io n m a ke rs
in
specific areas requiring
intellection. These areas
i nc lude
p l a n n i n g ,
schedu l i ng ,
r ecog n i t i on , c las-
s ifica t ion ,
d iagnos i s , pre d i c t i on ,
a n d
m a i n t en a n c e . T h e r ea s on i n g e m u -
l a ted
by
a
m a c h i n e
ca n
be
e ith e r in
con ju nc t i on
with h u m a n s ,
in
w h i c h
th e m a c h i n e acts
as
a n ass i s tan t , o r a lone,
in
w hic h th e m a c h i n e acts as
an a u t o n o m o u s intelligent agent .
T h e AI areas just m e nt i o ne d a re v ery bro ad , a n d t hus t echn iques d e-
v o t ed
to
t hem
a re
app l i cab le
to
multiple
ac t i v i t ies .
Fo r
exa m p l e ,
ef f ic ient
p l a n n i n g a n d s ch e du lin g a re
vital
to th e productivity a n d
profitability
of
large
m a n u f a c t u r i n g
co mp an i e s ,
so
th e
des ign,
cons t ru c t i on ,
a n d d ep lo y-
m e n t of c o m p u t e r p ro g ram s to
per fo rm
a u to m a t ed jo b -s h o p schedu l i ng
is a
focus of m a n y co mp an i e s , i n c lu d ing a u t o m o b ile m a n u fa c tu re rs .
But
p l a n n i n g a nd sc he du lin g a re
also
espec ia l l y important
within military
e n v i r o n m e n t s ,
in
wh i ch th e
ability
to generate l og is t i ca l p l a n s for , say,
th e rapid
mobilization
of c o m b a t forces
a t a n air
base
in
th e
face
of u n -
pred ic tab le ,
ambiguous ,
a n d
rapidly
c ha n g i n g
e n v i r o n m e n t s ,
is
critical
to national secur i ty . In f ac t , th e Defense A d v a n c e d Research Pro jec ts
A g e n c y (DARPA) work s with th e U.S. Air Force to f un d
AI
pro jec ts
in
force p l a n n i n g a n d aircraft d e p l o y m e n t , a m o n g others .
O n e of th e l a rgest areas of AI research
in
recogn i t ion
is
natural l an -
guage
process ing, of w h i ch there a re
m a n y
subareas.
A m o n g th e e a rly
hopes
for AI wa s th e ability to
per f o r m
natural
l anguage
translation,
say
f rom Russ ian to Engl ish . Th is p rov ed
no to r i ou s l y
difficult, a n d
po in ted
out
th e need
for knowledge of
c o n t e x t
w h e n dea l i ng with m a ch in e re p-
resen ta t i ons
of natural
l anguage.
(AI
fo lk lore
tel ls th e
story
of
a n
ear ly
sys tem
that a t t e m p t e d
to t r a n s l a te s imp le
sentences
f rom Engl ish to Rus-
s ian
a n d
then b ack to
En glis h. Wh en
th e sys tem was
given
The
spirit
is
willing but
th e fiesh is weak ,
it
r e sp o nde d with The
v o d k a is
good
but
th e m e a t
is rotten.)
But successes
ha ve
emerged ,
to
th e
point n o w
where t he re
a re u sa ble
front ends to da tab ase
sys tems supporting
natu-
ral l anguage quer ies .
Th e re is
also
n o w limited
v o c a b u l a r y
c o m m e r c i a l
dictation s o ftw a r e a v a ila b le ,
in
w hic h th e spoken
w o r d
is
c on v e r t ed
to
text.
In
th e area
of
d ia g n o s is , p e rh a p s
what first comes to
m i n d
is m ed i c i n e .
I n d eed , s om e of th e
m o s t
dramatically
successful
AI p ro gra m s ta k e a s in -
put
descr ip t ions
of th e
s y m p t o m s
a n d b ack gro u nd of a n af f l ic ted patient,
sa y
on e
suf fer ing
from lymph n o d e p ro ble m s , a n d
p ro du ce
as
output
a
co r re c t
d iagnos i s ,
a l on g with a de ta i led e x p l a n a t i o n
of th e fa c to rs used
in th e
p r og r a ms reason ing.5
But th e
m e t h o d s used
to create such
a
p r o -
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
28/256
16
. CHAPTER2
gram
a re e qu a l l y
app l i cab le
to o t h e r
areas in
w h i ch
co r re c t
d iagnos i s
is
requ i red , both within med ic ine
a n d
in areas
n o t even r emo te l y s im i la r .
Th i s is because AI p ro g ram m e rs h a v e l ea rned that th e
bes t
w a y to de-
sign p ro g ram s is
by distinctly
se p a ra t i ng
th e p ro gra m s knowledge f rom
its contro l . In th e s a m e
w a y
that th e c o n t r o l m ec ha n i s m of a c o m p a c t
d isk p la ye r ca n a ccep t a nd
read
m a n y d i f fe ren t
d isks,
a wel l -des igned
AI p ro g ram s c o n t r o l me ch an i sm
ca n
acce p t a n d
process
knowledge f rom
m a n y d i f fe ren t areas. T h u s th e lymph n o d e programs c o n t r o l m ec ha n i s m
might be app l ied to know ledge o f, say, card io log ica l d isorders , to
p ro du ce
a
n ew
p r og r a m fo r
d iagnos ing
diseases
of
th e h ea rt.
But
bey on d
that,
th e
p ro g ram s c o n t r o l me ch an i sm might
just a s wel l be app l ied
to
n o n m e d i -
ca l d o m a i n s ,
f rom
th e t r ou b le shoo t i ng
of c o m p l e x
m a c h i n e r y
to even
th e
debugging
( u nco ve r i ng
p r o g r a m m i n g
mis takes ) of c o m p u t e r p ro g ram s .
P r o g r a m s that a re
able
to
per f o r m d iagnoses
o r o therw ise ana lyze
in
a n a rea normally
regarded as
re qu i r i ng h u m a n e xp e rtis e a re examp les of
exper t systems m e n t io n e d e a rlie r. T h e y first became
widespread
in
th e
19 8 0s a n d
n o w
there a re t hous a n d s
in
e v e r y d a y
use. T o giv e
just a
tas te
of th e
types
of exper t sys tem a pp l i ca t i on s b e in g c re a te d t o d a y , h ere a re
s om e re c e n tly re p o rte d
in
a n international journal o n
e x p e r t sys tems:
0 Evaluating a n d monitoring th e l eve l of
air
pollution.
0
Aiding probation off icers in
sentence
r e c o m m e n d a t i o n .
o
Predicting p a y m e n t
p e r fo rm ance
on c o n s u m e r l oans .
0 Auditing
tasks in a
large public acco u n t i ng firm.
0 D i agno s i ng equ ine
cough.
0 Ass is t ing
in
acco u n t i ng
ed uc a t i on .
0
Process s ele ctio n a n d cos t i ng of th e h o le -m a k in g ope ra t i on
in m a n u fa c tu rin g .
T h e focus
in e xp ert sys tems is on e m u l a t i n g
h u m a n e x pe rtis e
by
repre-
sen t i ng
e x p e r t
knowledge
in
a
l anguage
that
both
mach i ne s
a n d
h u m a n s
ca n
u n d e r s t a n d . Th is l anguage
is of ten
symbo l i c logic,
but it
ca n
also
be one of
a va r ie ty
of
o t h e r
w e ll-s p e cifie d a n d
u n a m b i g u o u s languages
fo r
rep resent ing
th e fa c ts
a n d
ru les that
ca n
be
el ic i ted
f rom
h u m a n
sub-
jec ts . Currently, there is ev en
a n
ef for t a t S ta n fo rd University to
create
a knowledge i n te r l ingua, o r u n i v e r s a l l anguage for r ep r es e n tin g fa c ts a n d
ru les.
T h e use
of
these languages
is
cen t r a l
to
what ha s c om e
to
be
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
29/256
THE
DEFINITION
AND
S C O P E O F
AI
1 7
regarded as
th e c lass ica l
a pp roa c h to AI,
in w hic h
th e
symbo l i c
repre-
senta t ion of fa cts a nd ru le s, a lo n g with their manipulation via logic o r
logic- l ike
c o n t r o l me ch an i sms , is p a r a m o u n t .
N o w h e r e
in th e c lass i ca l a p p r o a c h
is
there
a n y
concern with th e w a y
in
w h ic h th e brain
phys io logica l ly goes
a b o u t th e
process
of
thinking
a n d
reason ing .
Th e re wa s a
tim e , in
th e
i n f a n cy of
m o d e r n
AI,
w h e n brain
ne u ro p h ys i o l o gy wa s cons ide red a s a m o d e l of computational thought,
but
fo r
va r ious
reasons
it wa s la rgely a b a n d o n e d
in
f a v o r of th e c lass i ca l
approach .7 In th e 1990s,
how eve r ,
a
resurgence
of th e
brain m o d e l
for
AI
o ccu r re d ,
d ue
la rgely
to
th e
success
of
artificial
n e u r a l
ne tworks ,
o r
p ro g ram s
that opera t e on th e principle of pa ra l l e l distributed process-
in g PDP . In P D P , c o m p u t a t i o n is
accompl i shed
through th e
work
of
m a n y
s m a l l
process ing e lements
taking
input f rom
a n d emitting
output
to
on e
a n o t h e r in m u c h th e
same
w a y
that
ne u r o ns
c o m m u n i c a t e
with
on e a n o t h e r
in a
bio log ica l
b r a i n .
Because of th e m a s s i ve connectivity
exh ib i ted
a m o n g
ne u r o ns in
th e
b ra in , th is
a p p r o a c h
to
AI is
somet imes
cal led
connec t ion i sm.
It turns out that
artificial n e u ra l n e tw o rk s
( A N N s )
a re g ood a t on e
thing that bra ins d o ve ry
w ell, a nd
that is
co m p l e x pattern
r ecog n i t i on ,
w h e t h e r th e
pattern
be a n image of
a
face, t yped charac ters ,
a
vo ice print,
o r quantitative
proper t ies of
c he m ic a l c om p o un d s. ANNs
ca n be m a d e
to recognize these pa t te r n s t h rou gh training techn iques that a re partially
b o r ro w e d f rom
w h a t
we k n o w
a b o u t
brain phys io logy . T h u s A N N s cons t i -
tute on e f ace t within
a large
area of AI
concerned
with m a c h in e le a rn in g .
Although
A N N s a re n o t th e on l y m o d e l fo r
m a c h i n e
l ea rn ing,
t he y
h a v e
p rov ed
successful enough a t l ow- leve l
tasks
to warrant a n exp los ion of
cu r r en t
research.
A m o n g th e
m a n y
c u r re n t a p p lic a tio n s
of this
research
are:
0
Class ifica t ion of bio log ica l part ic les f rom e l e c t ro n - m i c ro sco p y
images .
0
Class ifica t ion of seismic events .
0
Lea r n i n g
to fa c to r p o ly n o m ia l equa t ions .
0
Detec t i n g
c re d i t card
f r a ud .
T h e
success
of
A N N s
a n d o t h e r m a c h i n e
l ea r n i n g
t echn iques
in
th e
process ing
a n d
a ss im i l a t i n g of large a m o u n t s of d a t a
in
th e
training
p r o -
cess ha s
spur red
c u r r e n t i n te res t in a n a rea of
AI
k n o w n
as
knowledge
d i scovery
in
databases
(KDD;
also cal led
data mining). KDD
t r ies to
a u t o m a t e th e
process
of extracting f rom th e large a m o u n t s of d a t a ,
typ-
ically p r o d u c e d by sc ient ific a n d
bus iness
en d ea vo r s , u se fu l k no w le d ge
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
30/256
1 8
CHAPTER 2
that
wa s
p r ev ious l y
u n k n o w n .
Potential a pp l i ca t i on s i n c l u de a n a l yz i n g
purchases to de t e rm i ne c u sto m e r p ro file s , d is c o v e rin g
pa t te r n s
in
hea l th
care, o r even d iscover ing galax ies.
Th e re a re m a n y o t h e r areas of AI that I
h a v e
not m e n t i o n e d
here,
a m o n g
t hem
c o m p ute r v is io n a n d robot i cs , but I h o p e this sect ion has
descr ibed th e field a d eq ua t e l y enough
so
that
we m a y
n o w return to o u r
fo cus o n
th e relation of p h i l o so p h y to c o m p u t e r
science.
While
c o m p u t e r
sc ience
is a y o u n g
field
even within th e m o de rn
era,
it
turns
out that
through
AI
p h i l o so p h y
ha s
fo r t hous a n d s of years
been
p o n d e r i n g ques-
t i ons
that
m a n y c o m p u t e r
scient is ts
a re
concerned
with
n o w .
In
th e n e x t
ch ap t e r we will
lo o k a t
AI
t h rou gh
th e
ages in
th e history
of
ph i losophy .
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
31/256
_. 3 _ _ .
AI and the History of Philosophy
W h e t h e r
AI
is
concerned
with c re a tin g c o m p u t e r sys tems that think
a n d
a ct like h u m a n s ,
o r that think
a n d a c t
rationally, it
m u s t
address th e
m o d e l i n g of
reason ing
in
c o m p u t e r p rog ra m s . T h e p r o g r a m m i n g part of
this e n de a vo r, h ow ev er, is
se co nda r y
to th e p ro ble m of
con cep tua l i z i n g
a m o d e l of reason ing
in
th e first
p lace.
T h e availability of digital c o m -
pu te rs in th e twentieth cen tu r y
has
o n l y
a l l owed
th e
test ing
of
h u m a n
reason ing
mo de l s ;
th e c on c ep tu a l
creat ion
of
such
mo de l s
has been
th e
topic of ph i l osoph i ca l
invest iga t ion
for centur ies ,
in
th e form of logic.
In a
r ea l sense,
t hen , l og ic ians
h av e sin ce
antiquity
been
laying s om e of
th e
f o u n d a t i o n s
of AI,
a l t houg h
without a n y fo reknowledge of elect ron ics
they n e v e r
cou ld
h a v e
i m ag i ne d
th e
purposes to w h i c h
their
work wou ld
ultimately
be put
in
m a c h in es .
T h e
historical
f oun d a t i on s
of AI
la y
not o n l y
in
th e work of those
i n terested
in
formalizing r e aso n i ng through logic. J u s t as
important
to
AI
a re
tw o
u n d e r l y i n g
ph i l osoph i ca l
theor ies c on c e rn i n g
th e
mind
a n d
its thought
processes,
name l y ,
that t he y
a re
u n d e r s t a n d a b l e
within both
a
phy s i c a l
a n d
a c o m p u t a t i o n a l
m od e l .
T h e
first theory
i n vo l ves
taking
a
s t a n d o n th e traditional p h ilo s o p h ic a l p r o ble m of th e re l a t i onsh ip be-
tween m i n d a n d bo d y . Fo r
if ,
as
in
th e t he ory of mind/body d u a l i s m ,
th e m e n ta l events m a k i n g up consc ious thought
a n d
rea s on i n g a re n o t
necessar i l y d e p e n d e n t u p o n
cer ta in
p h ys i ca l
events
fo r
both
their
ex is-
tence
a n d co n t e n t , t hen to pursue th e
artificial
crea t ion of thought a n d
r ea s on i n g processes
in
a phy s i c a l m e d iu m (lik e
a
c o m p u t e r )
m a y
not be
easy
to
justify.
It
shou ld
be
n o t ed ,
however ,
that
a success fu l defense
of
a
phys ica l i s t i c
v iew
of th e
h u m a n mind
does
not
necessar i l y
imply
that
it is possib le to artificially p rod uc e thought a n d re as on in g processes
in a .
c o m p u t e r . It is possib le to be
a n
antidualist regard ing
th e
mind/body
re l a t i onsh ip , y e t still d e n y th e possibility of
so m e t h i ng
l ike thought o r
consc iousness
in
such nonb io log ica l
matter
as what m a k es u p
co m p u t e rs .
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
32/256
20
CHAPTER
3
T h e
second
t he ory
i nvo l ves
a
c o m m i t m e n t
to
th e
m a n n e r in w h i ch
m e n t a l even ts
respons ib le
fo r thought
a n d re a so n in g
t ake p lace , w h e t h e r
their m e d i u m is
p h ys i ca l
o r n o t . If
these
ev en ts a re not computational
in
n a tu r e , with their process ing ach ieved through
ru le l i ke ,
algorithmic
con t r o l ,
then to
attempt
to real ize th e n a tu r e of
h u m a n
r ea s on i n g
in a
c o m p u t a t i o n a l e n v i r o n m e n t ( l ike
a
c o m p ute r p ro gra m )
m a y
be
similarly
misgu ided . A full u n d e r s t a n d i n g of th e ph i l osoph i ca l
f o u n d a t i o n s
of AI
requi res
s om e
knowledge
of th e c o n te x t p ro v id e d
by
both logic
a n d
phi-
l o s ophy of m i n d .
It
is
i n te res t ing
that
theor ies
o n
logic
a n d
ph i l osophy
of
mind
hos -
p i tab le to th e m o d ern AI project
ca n
be
seen a s flourishing
in Greek
antiquity, th en e ith er re legated to se co nda r y
s ta tus
o r re v ile d through
m e d i e v a l t imes, a n d finally res tored to respec tab i l i t y
in
m o d ern tim e s in
pre pa ra t i on
fo r
th e grea t sc ient ific a n d techno log ica l a d v a n c e m e n t s of th e
twentieth
cen tury .
Ancient
Views
of the Mind
With
respec t
to
th e
re la t ionsh ip
of
m i n d
to matter,
it
is
accep ted
that
th e
first
major figure to
fra m e th e
prob lem
ph i l osoph ica l l y
was Plato
(c . 427347 b.c.). That a person cou ld
be
regarded s i m p l y as a
b o d y
was
r e p u g n a n t
to P l a t o n i s m ,
as illustrated
in th e f o l l ow ing
d ia logue
between
Socrates
a n d Alc ib iades:
Soc: [T]he user a n d th e
thing
he uses a re
d i f ferent ,
a re they
not?
A lc : H o w d o yo u
m e a n ?
Soc: Fo r i ns tance, I
suppose
a sh o e mak e r
uses
a
r o u n d
tool,
a n d a
square one ,
a nd o th ers , w hen he
cu ts .
Alc: Y e s .
Soc: A n d th e
cutter
a n d user is
q u ite d iffe re n t
from
what
he
uses
in cu t t ing?
A lc : Of
course.
Soc: And
in th e s a m e
w a y what th e h a r p e r uses
in
harping
will be
d i f fe ren t f rom th e h a r p e r
h imsel f?
A lc :
Yes.
Soc: Well
t h e n ,
that
is
what
I
was
ask ing just
n ow w het he r
th e user a n d
what
he
uses
a re a lways ,
in y o u r o p in io n , tw o
d i f fe ren t
t h ings .
Alc : T h e y are.
Soc:
A n d
m a n
uses
h is whole
b o d y
too?
A lc : T o be
sure.
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
33/256
AI AND THE HISTORY O F PHILOSOPHY
21
Soc: A n d
we said that
th e user a n d w h a t
he
uses a re di f ferent?
Alc: Ye s .
Soc: So m a n
is
d iffe re n t fr o m h is o w n bo dy ?
A lc :
It
seems so.
Soc:
T he n w h ate ve r is
m a n ?
Alc : I c a n n o t say.
Soc: O h , but yo u c an th at h e
is
th e user of th e bod y .
Alc :
Yes.
Soc: And th e user of
it m u s t
be
th e
soul?
Alc :
It
m us t .
S o c : And ruler?
Alc :
Yes.1
By th e
us e r a nd
ruler of th e b o dy ,
we
ca n
t ake
S o cra te s s ou l to be
th e reason ing
m i n d
with w h i ch we
a re c on ce rn e d.
T h e strict separa t ion
of h u m a n s
essences
f rom their bodies is cons is ten t with
th e
celebra ted
Platonic T h e o r y of Forms, w h i ch st ra t ifies
reality into
th e m u t a b l e
objects
of
ordinary
exper ience a n d th e
i m m u t a b l e ,
tim e le ss , a n d t r a n s c e n d a n t
rea lm of
th e
Forms, l ike jus t ice, equa l i ty , o r un i ty , with w hic h th e mind is
c on c ern e d w h en it reasons o r ph i losoph izes.
Cer ta i n l y ,
th e mind o r sou l
has
knowledge
of ordinary
objects
l ike p e o p le , tre e s ,
a n d
stones, through
th e de l i ve rances of
sight,
hear ing ,
t ouch ,
a n d
so
on . But n e i t he r th e m i n d
n o r th e concepts
it
reasons a b o u t a re th em s elv e s part of this
co rp o re a l
wor l d . It wou ld h a v e
been
imposs ib le
fo r
Plato to
conce ive
of
a n
artifically
created
m i n d ,
fo r
anything
so
f abr ica ted cou ld
ex is t
o n l y
in
th e
imper f ec t
material
wor l d ,
whi le r e aso n i ng
occurs in
th e re a lm of th e
F o r ms .
That
th e w o rk in g s
of
th e m i n d
a re
n o t
to
be
couched
in phys ica l ,
cor-
p o re a l te rm s is
also
b orn out by
Pla to s
v iew that l ea rn ing, a primary
activity of th e m i n d , is rea l ly th e reco l lect ion of
knowledge acqu i re d
out-
s ide th e
p h ys i ca l
l ife of h u m a n beings
a n d within th e re a lm
of
th e Forms.
A s Socra tes r em a rk s to M e n o :
T h u s th e sou l , s ince it is immortal a n d ha s
been
b o rn
m a n y
t im es , a n d ha s seen
a ll
th ings
both
here a n d in
the
other world,
ha s
l e a r n e d everything that is. So we
need n o t be su r p r i sed if
it
ca n
reca l l th e
k n o w l e d g e
o f
virtue
o r
anything
else
w h i c h ,
as
we
see, it o nc e
possessed.
All nature is
a k i n ,
a n d th e s o u l ha s
l ea r ned
everything, so that w h e n a m a n h a s reca l led a single piece
of
k n o w l e d g e l e a r n e d it, in ordinary l anguaget her e
is n o reason
why he s h o u l d
not find out
all th e rest , if he keeps a stout heart
and d o e s
not grow weary of the
search, for seeking
and learning
a re
nothing
but recollection?
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
34/256
22 CHAPTER
3
Intimations of
r e i n ca r n a t i on i sm
in
P l a t o n i s m
notwithstanding,
th e
c o m -
m o na l i t i e s between P la ton i sm a n d l a te r Christian doc t r i ne regard ing
th e
re la t ionsh ip of m i n d a n d b o d y a re c lear ,
in
that mate r i a l i sm, o r a n y asso-
c ia te d v ie w that
al lows thought o r r e aso n i ng to be p re d i ca t e d of material
th ings,
is to
be re jected.
But whi le Christian th e o lo g y w o u ld d o m i n a t e m e d i e v a l ph i l osophy for
centur ies , it is important to point out
that antiquity
h a d its materialists
a n d s ke ptic s with
regard
to th e m i n d . H i p p o c ra t e s (c . 4 6 0 - 3 5 7 b.c.) was
b orn before
Plato a n d is
bes t kn ow n fo r
th e m e d ica l o a th .
H e wa s fasc i -
n a t ed
by
epi lepsy,
w h i ch a t
th e
time
wa s
given
th e
n a m e
sacred
d isease.
Hippoc r a tes
u nde rs t o o d that
th e disease
wa s gran ted a
divine origin d ue
both to its
st range m an i fe s t a t i o ns a n d
to genera l
public
i g n o ra n c e c on -
c e rn i n g its cause, but he a lso u n d ers to o d th e primary ro le p la y ed
in
th e
disease
by th e bra in . M o s t
importantly, he ex t en d ed
th e b ra in s role
f rom a
mere causer
of cer ta in d isorders to a cen t r a l f ac to r in all aspects
of m e n t a l l i fe:
M e n
o u g h t to k n o w
that from the brain, a n d from th e brain on l y ,
ar ise
o u r
pleasures,
j oys ,
l au gh te r
a n d je s ts ,
as
w el l
as
o u r
so r rows , p a i n s , griefs
a n d
tears .
Through
it,
in particular,
w e
think,
see,
hear , a n d distinguish th e ugly from th e beautiful, the
ba d from th e g o o d , th e p l e a s a n t from th e u n p le a sa n t . . . 3
B e y o n d
u n d e r s t a n d i n g th e c en tra l role
p layed
by th e b ra in , Hippoc r a tes
even
ve n t u re d phys ica l , causa l e x p l ana t i o ns of
m a d n e s s
a n d i n te l l igence:
M a d n e s s c o m e s from its mois tness . When th e brain
is abnormally
m o i s t ,
of
necess i ty it m o v e s , a n d w he n it m o v e s neither
s igh t
nor
hea r i ng
a re
still, but
we
s e e
o r h e a r n o w
on e
thing
a n d n o w
a n o t h e r , a n d th e t ongue speaks
in
ac c o rdanc e with th e things
seen a nd h ea rd o n a n y occas ion . But all
th e time th e
brain
is
still
a m a n is in te l l igent /1
Although
v iews of
th e
m in d o r
sou l a s
i n co r po r ea l
in
n a t u r e pers is ted
t h rou gh
th e c lass ica l
ages in th e
Platonist a n d Sto ic schools , t he re were
also
wel l deve loped sys tems of materialist p h i l o so p h y p o s t - P l a t o , partic-
ularly in
th e Ep ic ure a n schoo l . Epicurus ( 3 4 1 2 7 O
b.c.) ,
a n a d v o c a t e of
a t o m i s m ,
env i s ioned
th e
sou l
a s
be ing
c om pos ed
of
th e
s a m e
a t o m s
that
m a k e u p air, w i n d ,
a nd h ea t, as
wel l
as a
fourth
kind
of a t o m m a d e sen-
sua l l y
unde tec tab le
d ue to its m i n u t e size. O n e s s o ul, w h i c h
is
th e cente r
of th e
self,
is therefore
a n e m p i ric a l o b je c t be ca us e o n es
b o d y is . T h e
sou l is
intertwined with b o d y a n d
c a n n ot e xis t without it.
Although
m o s t
of Epicurus writings
h a v e
been
los t , tenets
of
th e
Ep i cu re an
schoo l
su rv i ve
in th e
writings of
l a te r devotees
a n d fo llo w e rs .
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn
35/256
AI AND THE
HISTORY
O F PHILOSOPHY 23
O n e of t h e m ,
bor n
o v e r one h u n d r e d a n d
seven ty
years
a f te r
E picu ru s
d ea t h ,
was
th e R o m a n p o e t Lucre t ius
(c .
99-55
b.c.) ,
whose didactic
p o e m O n
th e N a t u r e of Things
wa s
a ce lebra t ion of
Ep i cu re an i sm . It
is
descr ibed
as
the fu l lest surviving expos i t i on of th e
m o s t c ohe ren t
a n d influential sys tem of
materialistic
p h ilo s o p hy p ro d u ce d in c lass i ca l
antiquity.5
It
is materialist
in
that
it
a t t e m p t s to lo c a te th e mind within
a p h ys i ca l
s u b s t r a t u m :
N o w I asser t
that
th e mind
a n d
the s o u l
a re
k e p t to g eth e r in
close
union a n d m a ke
u p
a
single n a t u r e , but
that
th e
directing
principle
w h ic h
we
ca l l
mind
a n d
u n d e r s t a n d i n g ,
is
th e
h e ad
so
to
s p e ak a n d re ign s paramount in the
w ho le
b o d y . It ha s a fixed
sea t
in th e
middle region
of
th e breas t :
here throb
fe a r a n d
a p p r e h e n s i o n , about
these
s p o ts d w e ll s o o th in g jo y s ; th e re fo r e
here is th e
understanding o r
mind.
While m o d e r n mate r i a l i sm w o u l d find th e
m i n d
m i s p l a c ed in th e breast ,
th e message is c lear that th e
facu l t ies
of u n d e r s t a n d i n g
a re
not to be
d i s em bod i ed , but to
fin d
their
seat
in matter. W he n c on fro n te d
with
e xp la i n i ng
th e
re la t ionsh ip
of
th e
mind
a n d
b o d y
as
t he y
interact
to
w a k e a m a n f rom sleep, for i ns tance,
Lucre t ius
refuses to c oun t en a n c e th e
possibility of
a
co nne c t i o n
between th ings
co rp o re a l a n d n o n c o r p o r e a l in
n a tu r e , ch o o s i ng
i ns tead a
materialist s tance:
[T]he nature of
th e
mind
a nd s o u l is bodily; for w h e n it
is seen to
p u s h
th e
l im bs , rouse th e body from
sleep,
a n d alter
th e
c o un te na nc e a nd gu id e a nd turn about
th e
w ho le m a n, a n d
w h e n
we
s e e
that
n o n e of
th e s e e ffe c ts ca n
t ake p lace
without
touch n o r
touch
without
b o d y ,
m u s t
we
not
admit that
th e
mind
a n d
the
sou l a re of a bodily nature?
So here
is a
repudiation of th e
strict
Platonic du a l i sm that wou ld
c o m e
to
d o m i n a t e both
p h i l o so p h y a n d t heo logy
in
th e
m i d d l e ages
a n d
l a te r
find its fu l lest m e t a p h y sic a l fl ow e r in g in C a rte sia n d u a lis m . It serves to
r e m i n d
us that AIs materialist intuition,
fa r
f rom be ing a product of th e
m o d e r n analytical a p p r o a c h to p h i l o so p h y
b uttre s s ed b y n e u ro p h y sio lo g-
i ca l
s cie n ce , is f o u n d
in
diverse ages a n d intellectual e n v i r o n m e n t s .
The Age
of Modern Philosophy
T h e fo u n d a tio n s of th e
m o d e r n
a p p r o a c h to AI were laid
b y p h ilo s o p he rs
w h o were concerned with
genera l
charac te r iza t ions of h u m a n u n d e r s t a n d -
ing. These ph i l osophers
regarded
as f u n d a m e n t a l th e ques t ion of
w h a t
is
th e n a tu r e of consc iousness,
including h o w
h u m a n s think
a n d
perce ive,
-
7/23/2019 Philosophy and Computer Science - Timothy R. Colburn