What Does a Concept Script Do?

ThePhilosophical
Vol.34 No. 136
Quarterly
ISSN 0031-8094$2.00
WHAT DOES A CONCEPT SCRIPT DO?
BY CORA DIAMOND
I
Two remarks
mademe wonderwhata conceptscriptdoes. The firstwas
ofFrege'sclaimsforhisconceptscript:
Hans Sluga's,in thissummary
statements
possessan objectiveconceptualcontent.
(1) Meaningful
in ordinary
langurepresented
(2) That contentis onlyinadequately
age.
(3) It is possibleto design a systemof notationin which the
canbe givenan adequateand
ofanystatement
content
conceptual
clearexpression.
methodology.
philosophical
Implicitin thisprogramis a threefold
is seen as thedetermination
The taskof philosophy
of theobjective
of their
a critique
of philosophically
content
statements,
interesting
intoan adequate
inordinary
andtheirtranslation
expression
language,
which the analytictraditionhas
language.It is a methodology
to carryout. It has done so by adoptingtheoutlineof
endeavored
butmodifying
thedetails.(Sluga,p. 67)
Frege'sprogram
of
as thenaturaldevelopment
Sluga thensees lateranalytical
philosophy
I thought,
whatFregewas at in hisconceptscript.That lineofdevelopment,
withthelinegoingtotheTractatus.
couldbe contrasted
Putting
things
crudely,
of"philosophically
on theTractatus
viewthereis no suchthingas a translation
intoan adequatelanguage;one pointof developing
statements"
interesting
sucha languagewouldbe toenableus toseethatthereareno suchtranslations,
If
to puttogether
suchstatements.
ourtendency
andthusto helpus overcome
of
of the analyticaltraditionis the naturaldevelopment
the methodology
oftheTractatus,
thenthemethodological
pronouncements
Frege'sphilosophy,
with
thematter
it wouldseem,arenot.Well,whichis?Or is theresomething
me. The otherremarkthat
thatquestion?That thenwas one thingtroubling
made me thinkthatI had some problemsaboutconceptscriptswas Peter
Geach's,in his veryilluminating
pieceon sayingand showingin Fregeand
The firstofthefourtheseshe defendsis:
Wittgenstein.
344
CORA DIAMOND
of logicwouldobligehimto
Fregealreadyheld,and his philosophy
there
that
are
whichwill clearly
hold,
logicalcategory-distinctions
ina well-constructed
showthemselves
butwhich
formalized
language,
cannotproperly
be assertedin language:thesentencesin whichwe
seek to conveythemin the vernacular
are logicallyimproperand
intowell-formed
admitofno translation
formulas
ofsymboliclogic.
All thesame,thereis a testforthesesentences
havingconveyedthe
intendeddistinctions-namely,
that by theiraid masteryof the
formalized
languageis attainable.'
of
WhatpuzzledmeherewasGeach'sclaimthatwe showourunderstanding
distincthelogically
ill-formed
sentences
whichseekto conveythesecategory
tionsinourmastery
ofa well-constructed
formalised
language.Is thismeantto
be consistent
withthe idea thata well-constructed
formalised
languagecan
itselfhelp us avoid the misunderstandings
which
those
logically
against
in thevernacular
That problemis closely
sentences
wereaddressed?
improper
relatedto thefirst,
The
theone thatarosein connection
withSluga'sremarks.
centralquestionfromwhichI startedwas really:whatis therelationbetween
theideaa philosopher
likeFregeorWittgenstein
hasofa conceptscriptandhis
In
in philosophy?
view(implicitor explicit)ofwhatmethodsare appropriate
of
there
is
a
connection
between
the
two.
The
tight
development a
Wittgenstein
of
as ithasbeen
is
connected
with
the
conceptscript
disappearance philosophy
Are
there
such
links
in
about
a concept
practised.
implicit Frege'sthought
any
of
doctrines
about
Geach's
the
roots
article,showing
script?
Wittgenstein's
and
in
the
roots
of
that
saying showing Frege'swork,suggests
Wittgenstein's
viewof philosophy
itselfmaybe lookedforin Frege,and thatwas whatI
proposedto do.
II
oftheroutehe had
The naturalplaceto startis withFrege'sownsummary
taken.
I started
outfrommathematics.
The mostpressing
need,itseemedto
to
was
this
with
foundation
...
science
a
better
me,
provide
The logicalimperfections
of languagestood in the way of such
I
to
tried
overcome
theseobstacleswithmyconceptinvestigations.
In
I
this
was
led
from
mathematics
to logic.
script.
way
Whatis distinctive
aboutmyconception
of logicis thatI beginby
of
to
the
content
the
word 'true',and then
of
givingpride place
P. T. Geach,"Sayingand Showingin Fregeand Wittgenstein",
in Essayson Wittgenstein
in
HonourofG. H. vonWright,
ed. JaakkoHintikka,
ActaPhilosophica
Fennica,28 (1976),54-70.
WHAT DOES A CONCEPT SCRIPT DO?
345
immediately
go on to introducea thoughtas that to whichthe
question'Is ittrue?'is in principle
applicable.So I do notbeginwith
I
and
to
or judgement;
them
concepts put
together forma thought
comebythepartsofa thought
the
byanalysing thought.(PW,
p. 253)
There is an awfullot in thosefewsentences.The important
thingis the
relationbetweenFrege'snotionof a conceptscriptand his use of theword
He recognised
thatpeoplemightobjectthathe was usingtheword
'thought'.
in
fromtheordinary
one.In fact,we shouldseehim
a
sense
different
'thought'
as layingout,inlargepartthrough
ofwhathe calleda concept
thedevelopment
to.
the
notion
of
that
he
refers
script,
thought
A conceptscriptis a modeofexpression
of thoughts
so thatwhatit is for
themto be trueis clearfromhowtheyarewritten,
at
and, thesametime,the
in
of
the
notation
relations
different
areclearin the
logical
expressed
thoughts
a
of
The
difference
between
those
conceptscriptand
expressions
thoughts.
is
that
in
the
latter
there
tells
us,
onlyan imperfect
ordinary
languageis,Frege
the
of
the
sentence
and the
between
structure
perceptible
correspondence
in the
structureof what is expressedby it. Similarityof construction
in
the
structure
of
what
is
total
hide
may
logicaldissimilarity
expressions
for
are
a
matter
left
of
constituents
the
relations
the
expressed; genuinelogical
one to guessat (PW, pp. 12-13).
ofthoughts,
butI now
I said thata conceptscriptis a modeofexpression
are whatthe
wantnotexactlyto go backon thatbut to qualifyit: thoughts
couldbe putin termsofa
conceptscriptshowsthemtobe. WhatI amdenying
arewhatcan be trueor
crudepicture,roughly
this:we are toldthatthoughts
false,and so we can thenlookaroundandfindwhichamongall thethingsthat
we can thinkof- sentences,propositions,
thoughts,what not- have the
ofbeingtrueorfalse.Havingfoundwhatwe arelooking
we areafter,
property
has
itsnature.Lo andbehold,itcanbe analysed,
for,we couldtheninvestigate
such-and-such
sortsof parts,or whatever.No. That withwhichlogic is
concerned
hasitsnaturemadeclear(or,ifyoulike,itsnature-as-far-as-logic-isof
concerned)througha kindof writingof sentencesin whicheverything
to logicis madeclear.Makingthatclearis goingto takesomedoing.
interest
whatFregeis aimingat.
whatitis tomissaltogether
ButI canstartbyshowing
withconcepts
I haveseena criticism
ofFregeforhavingidentified
properties
in hissense.Fregehad said "I calltheconceptsunderwhichan objectfallsits
of thecriticwas thatthatidentification
would
and thecomment
properties",
countforFregeas oneandthe
notdo. Sincehavinga heartandhavinga kidney
sameconcept,butarenotwhatwe shouldcountas thesameproperty
(because
to the
is accidental),
we shouldnot(according
theirhavingthesameextension
the idea of a
critic)acceptFrege'suse of his term'concept'in explicating
(Currie,p. 94). But whenFregesaysthathe callstheconceptsunder
property
346
CORA DIAMOND
whichan objectfallsitsproperties,
he doesnotmeantobe givingan explication
of theidea of a property.
The pointis nottheidea of a property
butwhata
is. The Fregeanidentification
property-as-far-as-logic-is-concerned
may or
maynotactuallydo whatFregewantsit to do, but whathe wantsto do is
thatbelongsto our ordinary
idea of a property
preciselyignoreeverything
of
what
is
interest
to
And
has
narrow
interests.
If you
except
logic.
logic
very
and I fallunderall thesameFregeanconcepts,
thanas faras logicis concerned
to you fora
you and I are one and the same thing.Substitutea reference
reference
to me in anycontext,
and truthvaluewillneverbe affected.
Why
shouldproperties
be any betteroffthanyou and me? Logicwill treattheir
as ittreatsours,ignoring
whatis notofinterest
toit.In theconcept,
we
identity
can see- this is Frege's claim- exactlywhat interestslogic in what we
thinkof as a property.
The critic'smistakewas to thinkthat
ordinarily
have certaincharacteristics
and a philosophy
of logicoughtto do
properties
But thatis notphilosophy
oflogicforFrege.
justiceto thosecharacteristics.
of logic,how exactlydoes a concept
Given,then,thespecialisedinterests
them?In Fregewe can see thebeginnings
ofsomething
whichis
scriptfurther
whatthought
is is madeclearnotso muchin
fullydevelopedin theTractatus:
sentences
aboutthought
but in theclearexpression
of thoughts
in a concept
script.
Beforeexplaining
difference
between
that,I needto mentionan important
so thatit can be keptseparatefroma different
Frege and Wittgenstein,
difference
betweenthem.Withinanysphereofscientific
as Frege
investigation,
seesit,we can distinguish
betweentwothings,
themakingclearofthesubject
matterofthatactivity
and theestablishing
of thelawsgoverning
thatsubject
matter.
In gravitational
forexample,we arenotconcerned
withthe
mechanics,
chemicalproperties
of bodies,and a notationformechanicsin whichwe did
betweenbodiesaccording
to theirchemicalproperties
wouldmark
distinguish
to the science;it would to that extentbe a
somethingtotallyirrelevant
notation.
The samegeneralpointappliesto logicas Fregeseesit:it
misleading
is forhima sciencewitha realm,a subjectmatter,
ofitsown.Thereis on the
onehandthebusinessofmakingclearwhatis ofinterest
to it,thecharacter
of
its subjectmatter,and on the otherthereis the establishing
of the laws
thatsubjectmatter.
governing
(See, e.g.,FG, pp. 107-10.)If a conceptscript
doeswhatitis meanttodo,itshowsthecharacter
ofthought
bythesystematic
in thescriptofeverything
withwhichlogicis concerned.
The logical
marking
lawsthencontainthought
aboutthought,
and theircontentis something
that
in a conceptscript.(That
goesbeyondwhatis shownofthenatureofthoughts
is theviewFregetakesmostofthetime.In one verylatepassage,he suggests
thatlogic
that,ifwe had a logicallyperfect
language,therewouldbe nothing
thediscipline,
couldtellus thatwentbeyondwhatwasalreadycleartous
itself,
inthewaythingswerewritten
inthatlanguage.Butthislateviewwouldstillbe
WHAT DOES A CONCEPT SCRIPT DO?
347
thoughtsabout
compatiblewiththe idea thatlogicallaws werethemselves
thought.)
Wittgenstein
rejectstheanalogybetweenlogicand othersciencesand the
idea thatthereis on theone handthemakingclearin a conceptscriptofthat
withwhichlogicis concerned,
of thelaws
and on theothertheestablishing
aboutthesubjectmatter,
thelawsbeingthenthoughts
Thereis
aboutthought.
no 'on theotherhand' here,and logicis no science.
Whenyou have,in the
ofthought,
moreremains
tobe
madeclearthecharacter
conceptscript,
nothing
said-or so he thought.That, then is one point on which Frege and
differ.
It shouldbe distinguished
fromanother:
canwhatis made
Wittgenstein
ofthought
in thewaysentences
in a concept
clearofthecharacter
arewritten
in
the
of
also
be
sentences
about
character
putdirectly,
thought?
script
In discussingthat,I am goingto use, insteadof theword'thought',
the
for
it
to
which
should
be
as
an
abbreviation:
'that
which
regarded
expression
the question'Is it true?'is in principleapplicable'.A that-to-which-thatandwith
is expressedin a sentence.It getsexpressed
clearly,
question-applies
tologic,irrelevant
totheapplicability
toitofthequestion
allthefrills
irrelevant
oftruthorfalsity,
leftoff,ina conceptscript.Whenoneis struggling
tograspa
thought,one may come out firstwithhalf-sentences,
stammerings,
vague
to
Whileoneis struggling
jumbles,andonlyintheendgettoa propersentence.
ofthought,
ofthat-to-whichabouttheinternal
character
getstraight
something
one's stammerings
and gropings,
one's beginthe-question-of-truth-applies,
taketheformof sentencesin ordinary
ningsto getwhatglimmers,
language
betweenthought
aboutthought,
abouttherelation
anditsexpression,
aboutthe
elements
ofa thought,
andso on.Butjustas wemaysaythatthat-to-which-thein a sentence,
has its expression
thoughtaboutthe
question-of-truth-applies
natureof that-to-which-the-question-of-truth-applies
gets put clearlyin a
a wayof writing
sentences.
The word'thought'whichI have
concept
script,
can thatbe takentomeanthat-to-whichallowedto standin thelastsentence:
Can suchthought,
or 'thought',
also be put in
the-question-of-truth-applies?
sentences?These questionsget unambiguousanswersfromWittgenstein.
he does not hold thatin generalwhatcan be
Fregetreatsthemdifferently;
showncannotbe said.(I cannotdiscussin detailthisdifference
betweentheir
views;I touchon it briefly
againin ?11.)
Let megivea simpleexample(andthena morecomplicated
one)ofhowthe
character
of thoughtmaybe made clearin a conceptscript.We maystart
ofthetruthor falsity
aboutthedependence
ofa thought
on whatthe
thinking
is about,andwemayputtogether
ofphilosophical
thousands
sentences
thought
in suchgropings.
Fregeleadsus to a betterkindofgroping,
makinguse ofthe
notionsof argument,
function
and value,termsof not-so-ordinary
language,
with metaphorsof completenessand incompleteness.
But
supplemented
in termsof arguments
and so on is itselfa move,thecrucialmove,
thinking
348
CORA DIAMOND
inwhichthecharacter
ofthedependence
towards
a notation,
hisconceptscript,
thatwe see Frege'sconceptscriptand notations
is plain. I am suggesting
to putourthought
to thepreviousattempts
about
derivedfromit as standing
in something
thedependence
likethewaya clearlyputsentencestandsto the
thatprecededit.
half-formed
thoughts
is thedisputebetweenWittgenAnotherexampleworthlookingat briefly
to anything
thatcan
steinand Fregeaboutthekindof articulation
necessary
that
In
held
a
the
Tractatus,
Wittgenstein
anything
capableof
express thought.
must
articulated.
He took
has
be
the
sense
that
a
sentence
logically
having
in thiswithFrege,and I am surethathe was right.
himself
to be disagreeing
of
is that(a) theexpression
WhatI taketobe Frege'sview(inhislaterwritings)
in
of
is
kind
articulation
as
there
a
has
the
same
complex
thought, sentence,
hasa sense
but(b) a complexdesignation
like'Frege'sbirthplace',
designations
ofa sortwhichit is possibleforan unarticulated
propernameto have,which
that
of
sense
a
sentence
has does notrequire
that
the
kind
(c)
together
imply
The sensethata sentencehas couldbe
has thatsensebe articulated.
whatever
thesenseof an unarticulated
propername.(Since I wantto treatthisview
I
itto Frege.)
shall
not
discussthegroundsforascribing
as
an
example,
purely
about?
Whatis itthento takethatdisputeas an exampleofwhatI wastalking
of a
the expression
Viewedas I suggest,to disputeaboutwhatarticulation
fora conceptscript.
musthaveis to disputeaboutthespecifications
thought
in a concept
"viewaboutthoughts"
Wittgenstein's
getsitsproperexpression
can occurofanylogically
script,in theexclusionfromplaceswheresentences
in the exclusionfromplaces wheresimpleand
unarticulated
expressions,
oranysign
undefined
canoccurofanylogically
complexexpression
expressions
defined
via others,and so on. In a conceptscriptofthesortFregeusesin the
Grundgesetze,
anyargument
placeopento sentencesis opento anysimpleor
complexpropernames,and vice versa.This opennessof argument
placesis
whatitcomesto to say,as he did,thatsentences
arepropernames:logicdoes
notneeda kindofwriting
thatmarksa distinction
betweenthembyopening
argument
placesto one butnottheother.
Let me imagine,
at thispoint,twosortsofcomments.
too freewithexpressions
like 'whatlogic
(1) "You have been altogether
in'. Is it nottimeyouweremoreforthcoming
needs','whatlogicis interested
aboutwhattheyare supposedto mean?"
(2) "You havenotmadeveryclearwhatis involvedin thedisputebetween
- not,thatis,madeclearwhat
aboutlogicalarticulation
FregeandWittgenstein
kindofdisputeitis. For,as youhaveexplainedthedifference
betweenthem,it
fora conceptscript,and any
is a difference
betweentwosetsofspecifications
further
pointofviewon thedisputewouldgetexpressedin yetanothersetof
fora conceptscript.That means thatwe cannotcriticisea
specifications
WHAT DOES A CONCEPT SCRIPT DO?
349
thebusinessofthemakingofconceptscripts,
outside
conceptscriptbystepping
ofworking
outhowto writesentences
so thatrelations
ofinterest
to logicare
markedin a systematic
outside
and thoroughly
To
would
consistent
way. step
in a conceptscriptis thoughts.
be to say:Whatarewe expressing
have
a
They
and
once
we
find
out
shall
able
to
see
what
what
that
we
be
logicalstructure,
is,
tobuildintoa conceptscriptthebettertoreflect
thestructure
logicalstructure
We
should
and
have.
be
able
to
between
thoughts
judge
Frege's Wittgenstein's
whichhas a logicalstructure
to
conceptscriptsbydetermining
corresponding
thatof thoughts.But if we rejectthatview,whatexactlyis involvedin
or evaluatinga conceptscript?If it is truethatthoughts
are in
criticising
essencewhattheyareshowntobe ina conceptscript,
howis itnotthecasethat
Frege'sconceptscriptembodiesone notionof thoughtand Wittgenstein's
andthatno questionofwhichis a betterembodiment
ofwhatthought
another,
is can arise?If itis nota substantial
aboutsomething,
thatthey
thesis
thoughts,
in sentences
from
areexpressed
whicharelogically
speakingindistinguishable
thinkthatFrege is wrongin taking
propernames,how does Wittgenstein
sentences
to be propernames?"
To beginwith,thesecondcomment
is confused.It is notthecase,on the
in
thatwhatcountsas thought
is different
view I have been expounding,
differently
designedconceptscripts;nor is it the case thatthe onlywayto
criticise
a conceptscriptis to comeup withanotherone. For helpwithboth
we need to get back to the remarksof Frege'swhichI quoted
comments,
earlier,whichI said had so muchin them.Whattheyhave in themis his
oflogic,in which,he said,he gaveprideofplaceto thecontent
of
conception
theword'true'.
Whatmade it possibleforFregeto use the word'true'to "indicatethe
Ifweattachthepredicate
essenceoflogic"wasitspeculiarcharacter.
'is true'to
a sentence,
thesenseofthesentence
doesnotchange,butthepredicate
is noton
thataccountsenseless.It is forthisreason,Fregesays,"thattheword'true'
to the
seemsto make the impossiblepossible:it allowswhatcorresponds
to thethought"
assertoric
forceto assumetheformofa contribution
(PW, pp.
251-2).We cannot,though,bringoutwhatis peculiarto 'true'unlesswe can
whosesense,in
makeclearthedifference
betweenitandsomeotherexpressions
certain
it
into
the
sense
ofwhatthey
circumstances,
similarly
collapses,as were,
have been attachedto. What 'is true' does with sentences,
these other
will
do
with
some
other
of
function
expressions.
expressions
groups completed
I can explainthisin two ways,corresponding
to two uses of 'true'as a
predicate:
is true'= 'The thought-that-p
(1) 'p is true'= 'The thought-that-p
is a senseoftheTrue'.
350
CORA DIAMOND
(2) 'p is true'= 'p is one and thesameas theTrue' = 'The truth
valuep is one and thesameas theTrue'.
In thefirst
in thesecond
case,whatis meantby'true'is ascribedto a thought;
else
is
meant
'true'
and
is
to
that
ascribed
a
truth
value.But
case,something
by
thethought
that[thethought-that-p
is a senseoftheTrue] is thesameas the
thoughtthat[thetruthvaluep is one and thesameas theTrue], and that
uses of 'true'
thoughtis the sameas thethoughtthatp. The twodiffering
I
it
want
to
show
is
what
simplychop up differently.
peculiarto 'true'by
other
it
with
function
whose
sense
contrasting
expressions
collapsesin certain
I
and
can
do
either
that
with
the
first
circumstances;
bysticking
typeofuse of
'true'or by sticking
withthesecond.I can get a class of peculiarfunction
in eithercase- a classofexpressions
whosesensecollapsesin the
expressions
samewaythatof 'true'does. It wouldbe morecomplicated,
thoughin some
waysmoreFregean,to use thefirstkindofcase; butif I stickto thesecond,
caseforhimself.
anyonecanmakethemovesto themorecomplicated
Sticking
tothelesscomplicated
casewillenablemetomakeclearwhatI needto;so that
is whatI shalldo.
We canstartwithanygroupoffunctions
whosevalueis alwaysoneorother
of twoobjects,forexample,withfunctions
whosevalueis alwaysWismaror
Frankfurt.
Call theseB-functions
andexpressions
forthemB-expressions.
We
candefinea particular
thatthevalueof
B-expression,
say'W( )', byspecifying
the functionW( ) is Wismarwheneverthe argumentis Wismarand is
otherwise
Frankfurt.
in justtheway
'W( )' actswithcompleted
B-expressions
Whatever
thepeculiarity
is thatsuits'true'for
'( ) is true'actswithsentences.
theessenceoflogic,itis notmerely
thekindofcollapsibility
ofsense
indicating
itshareswith'W( )', for'W( )' cannotbe usedtoindicatetheessenceoflogic.
How thenarewe to getat thesignificant
difference
between'true'andsuch
as 'W( )', whichwemaycall"collapsing"expressions?
Let meput
expressions
thisina slightly
different
it,and'( ) is true',inso
way.'W( )', as I havedefined
far as that is what Frege would regardas a predicatewith a properly
determinate
initsargument
sense,caneachtakeanycomplete
expression
place.
The characteristic
collapseofsenseoccurs,foreach,onlyin certaincases;that
is one reasonI haveavoidedcallingthemidentity
But we maycall
operators.
the class of argument-expressions
forwhichthereis thecollapsethe "preferred"
kindofexpression.
We cannotbringoutthepeculiarcharacter
of'true'
unlesswe can makeclearthe difference
betweenit and collapsingfunction
kind of argumentis not sentences.
It is the
expressionswhose preferred
difference
between
a collapsing
which
is
and
others
which
a
predicate
expression
arenot.But thedifference
betweena predicate
and theseotherexpressions
is
thedifference
fora function
betweenan expression
whosevalueis a truthvalue
and expressions
forfunctions
whosevalueis not.The fundamental
difference
WHAT DOES A CONCEPT SCRIPT DO?
351
is thatthesenseofa completed
betweenthetwotypesofexpression
expression
whosevalueis a truthvalue)is something
ofthefirstsort(one fora function
is. Other
function
thesenseofno othersortofcompleted
expression
judgeable;
namethisor thator the otherdependingon
sortsof completedexpression
is fulfilled.
certainconditions,
buttheirsenseis neverthatanysuchcondition
Thus 'Frege'sbirthplace'
namesWismarifhe was bornthere,butitssenseis
or thattheFrankfurt
conditions
are fulfilled,
notthattheWismarconditions
of
not
lie in its
or
all.
The
character
does
at
that
'true',
are,
peculiar
anything
in
with
its
class's
that
combined
but
feature,
preferred
being
beingcollapsible,
to
Becauseit doesnotadd anything
and itsbeingitselfa predicate.
sentences,
it
to
do
it
has
ofitspreferred
what
seems
when
thesenseofan expression
sort,
initsargument
suchan expression
placeis itselftoexpressthe'theseconditions
of thoughtitself.A thoughtis something
are fulfilled',
the characteristic
in beingjudgeable;
from
other
senses-of-complete-expressions
distinguished
function
indicators
ofjudgeablesensenormally
butthegrammatical
also,orcan
warnsus to
force.Thatis whyFregerepeatedly
ofassertoric
do so,as indicators
forcefrompredication.
dissociateassertoric
'True', in seemingto haveas its
are
sense the generalformof thoughtish
sense,the 'the truth-conditions
thatall thoughtshave in common,can also seem to expressthe
fulfilled'
This thenis whyitcan seemthatwhatit
ofthetruthofa thought.
recognition
force.
as a predicateis actuallytheassertoric
to thethought
contributes
as true,and the usingof sentences
The businessof recognising
thoughts
- thatis thepeculiar
thisrecognition
withassertoric
forceinexpressing
uttered
of 'true'.Whathas it to do with
businessindicatedby thepeculiarcharacter
tobe
is recognised
The generalformofthispeculiarbusinessis: a thought
logic?
tobe true,
fromotherthoughts
inferred
truewithout
recognised
beinglogically
fromothers.Suppose,forexample,I recognisethe
or it is logicallyinferred
sentencetobe a
and I takeanother
truthofa thought
bya sentence,
expressed
I taketoexpressthesamethought
as the
conditional
whoseconsequent
material
I mustthenbe abletoinferthetruthofthe
onewhosetruthI haverecognised.
as a whole.Thatrulebelongsto
sentence
thought
expressed
bytheconditional
to be true;itis a "law ofthought",
thebusinessofrecognising
or,as
thoughts
Whatunderliesit is
Fregesaysit wouldbetterbe put,a "law of judgment".
of the
of thetruthconditions
butmerelytherelations
nothingpsychological
truthconditions
thatsuch-and-such
A thought
twothoughts.
beingessentially
it standsin to
fromthelogicalrelations
is inseparable
arefulfilled,
itsidentity
while
the thought
relationscannotchange
otherthoughts;its inferential
of
the rules
character
remainsunchanged.The whollynon-psychological
is.
is no
There
from
what
is thusinseparable
judgment
governing
judgment
of
the
how
of
sets
rules
different
roomforessentially
activity judging
governing
ofwhatis
ofchess,withvariants
shouldbe carriedon. Therecan be variants
to be
of
the
be
variants
in
that
sense
But
cannot
allowed.
there
prescriptions
352
CORA DIAMOND
in recognising
followed
to be true;therecannotbe whatusedto be a
thoughts
validinference
(wasa validinference
bythethenlawsoflogic)butis no longer
the tie
one. 'True' is suitedto indicatethe essenceof logic by indicating
and whatjudgment
itselfis.
betweenitscompleteunarbitrariness
Fregedid allowthattheremightbe beingscapableof graspingthoughts
But giventhatwe
of thought.
withouthavingto use perceptible
expressions
ofthoughts,
humanbeingsdo havetouselanguageintheexpression
something
it
oflogic.Or,rather,
followsfromthepeculiarcharacter
ofcentralimportance
the
of
it
is
from
the
character
follow
doesnot,properly
logic;
peculiar
speaking,
in thecharacter
of logicreflected
of language,considering
peculiarcharacter
whichthe
The principles
thoughts.
through
languageas capableofexpressing
forany
in
same
must
be
the
lawsoflogicbearon thought
expressed language
of
this
the
the
the
grammar
language.Fregeputs point way: logicalcomponent
of everylanguage(i.e., thatpartof the grammarthatdoes not reflectthe
and so on ofthespeakers)is thesamein all
interests
psychological
capacities,
It
languages(PW, pp. 6, 142). cannot,forexample,be thecase thatin one
thesamethingofa conceptas ofan objectandin
predicate
languageyoucannot
of languagemanyhuman
someotherlanguageyoucan. In thedevelopment
the
wereat work(PW,
instincts
anddispositions,
including logicaldisposition,
If
at
work
the
had
been
there
languagewould
logical
disposition,
only
269).
p.
of
and
we
should
be abletosee
to
the
be farbetter
thoughts,
expression
adapted
in
the
ofouractual
it
in
it
is
there
but
as
were
overlaid
what
grammar
clearly
Talk whichuses'true'
sharedwithall languages.
thelogicalfeatures
languages,
is one wayto getat something,
and takesadvantageof itspeculiarcharacter
"theessenceoflogic";andthesamethingis beinggotat ifwe talkaboutwhat
to thepeculiarcharacter
of 'true':thesinglelogicalin a sensecorresponds
can be
foundin everylanguagein so faras thoughts
structure
grammatical
is
ofgrammar
i.e., thelogicalcomponent
expressedin it at all. Corresponds:
to
in thestructure
oflanguageto thejudgeability
whatcorresponds
belonging
in thestockofhumanthoughts,
eachofwhich(orat anyrateits
everythought
logical core) is expressiblein everylanguage,accordingto Frege. The
in different
of different
langulogical-components-of-grammar
impossibility
the
of different
to theimpossibility
prescriptions
governing
agescorresponds
I saidthatwe have,in
ofrecognition
ofthetruthofthoughts.
peculiaractivity
in all
oflogicalgrammar
Frege'suse of'true'andin talklikehisoftheidentity
ifwe
Bothwouldbe unnecessary
at something.
twowaysofgetting
languages,
andhencethe
inwhichtheessenceofthought
hada logically
language,
perfect
essenceoflogiclayopento view.
in Frege'sview.Thoughtsmaybe communiI haveignoreda complication
on thelogical
thespeakers'actuallyrelying
catedin ordinary
languagewithout
ofthesentences
articulation
theyutterandhear.The hearersuse cuesto guess
thesentenceas a wholeexpresses.But Fregecombinedthat
at whatthought
WHAT DOES A CONCEPT SCRIPT DO?
353
ideawitha viewofthestructure
ofnaturallanguageon whichitis possiblefor
thesenseofa sentenceto be constructed
fromthatofitslogicalcomponents,
andnotjustguessedatas he thinks
ofconstructing
weoftendo. The possibility
thesensein thatwayis essentialforinference.
The grammar
of Englishdoes notallowus to put together
sentenceslike
'Five is greater
thanthreeis greater
thanthree',inwhicha wholesentence
has
of someothersortsmaygo. Fregecameto
beenput whereonlyexpressions
believethat,ina correct
'five'couldgo,'Fiveis greater
than
notation,
anywhere
three'couldequallygo. Now, givenhis viewthatthelogicalelementof the
ofalllanguages
is thesame,Englishandhisconceptscriptmustshare
grammar
So ifhis conceptscriptis actuallycorrectin
thatuniversallogicalgrammar.
in
it
what
allowing
corresponds to 'Five is greaterthanthreeis greaterthan
the
exclusion
of suchsentencesfromEnglishis notdoneby universal
three',
ofEnglishgrammar.
but
logicalgrammar by the"psychological"
component
thatanythought
canbe
earlier,
(We canseethattheFregeanclaimI mentioned
in
thin
sense
of'can'.
It
the
in
is
true
a
i.e.,
can,
expressed everylanguage,
only
it
but
the
of
the
allows
to
be
non-logical
expressed;
component
grammar
logical
in anysentencethatit
ofthegrammar
maypreventitsexpression
component
ofEnglishgrammar
willallow.)Justas thenon-logical
mayexclude
component
fromEnglishwhatlogicalgrammarwould not exclude,it may also allow
ofwhatit
wouldexclude,liketheputting
combinations
whichlogicalgrammar
cannotrecognisea genuine
countsas a propernamewherelogicalgrammar
propername.
We are now in a positionto get back to the disputebetweenFregeand
ofa conceptscript.In so faras a languageis
aboutthecharacter
Wittgenstein,
at all, it must(hereFregeand Wittgenstein
thoughts
capableof expressing
ofitsgrammar
whateverylanguagehas.
agree)haveas thelogicalcomponent
The pointaboutthelogicalgrammar
appliesnotonlyto anynaturallanguage
but also to a conceptscriptdesignedwiththe intentionthatthe logical
of the thoughtsexpressedin its sentencesshouldbe shown
characteristics
of thosesentences.A conceptscriptis
structure
clearlyin the perceptible
whenit treatsin thesamewaywhatis logicallydifferent
(what
unsatisfactory
or treatsin different
thelogicalgrammartreatsas different)
wayswhatis
similar.
Let us stickwithourexampleandaskwhatitwouldmeanfor
logically
to be rightin excludingsentencesfromargument
placeswhere
Wittgenstein
is right,andifwe assumeas
fornumberscan go. If Wittgenstein
expressions
both held about language,what
correctthe view Frege and Wittgenstein
inFrege'sconceptscriptbetweenthelogical
follows
is thatthereis a divergence
whichcombinations
of
andtherulesfixing
itshareswithall languages
grammar
sentences
oftheconceptscript.The upshotofthe
arewell-formed
expressions
to Frege'sintentions
and to hisbeliefsaboutwhat
is that,contrary
divergence
will be allowed.The
he had succeededin doing,nonsensicalcombinations
354
CORA DIAMOND
is analogoustoonesthatariseinordinary
situation
as in theexample
language,
I mentioned,
in whichthenon-logical
of
allowswhatit
component grammar
as
a
name
where
the
refuses
to
one.
regards proper
logicalcomponent
recognise
we
assume
and
about
the
articulaIf,instead,
Fregeright Wittgenstein
wrong
tionofsentences,
therewillagainbe a divergence
betweenwhattherulesofa
allowand whatits
conceptscript(one meetingWittgenstein's
specifications)
ownlogicalgrammar,
allows.In thiscase,theupshotof
thelogicalgrammar,
thedivergence
wouldbe theexclusionfromtheconceptscriptofcombinations
whichwouldbe allowedby the logicalcomponent
of grammar;
again,the
situation
is analogousto one whichcan obtainin ordinary
language.
Whatthendoesall thiscometoas a replyto thesecondcomment?
Thereis,
on theviewI havebeenexpounding,
a rightand wrongin disputeslikethat
betweenWittgenstein
and Frege.Whenthedesignof a conceptscriptgoes
not
ofthescriptfailstomatchsomething
it
is
that
the
structure
external
wrong,
to it,butthattheconceptscripthas structural
whichdivergefromits
features
innerstructure,
whichis thusnotrevealedclearly.
own,as it were,inevitable
Later on, Wittgenstein
came to say that,when we thinkthatthereis
ofillusion.Butat
wearethevictims
something
queeroruniqueaboutthought,
tomakeclearwhatitis wemaythusseemtosee;foritis
thispointI amtrying
notall illusion.Takinga viewaboutthenatureof thoughts
is notquitelike
takinga view about the natureof whales,and we misreadFrege and
are concernedwiththe
both,ifwe forgetthattheythemselves
Wittgenstein
characterof that difference,
and thus concernedto "place" what they
themselves
say.I mean'place' in thesensein whichwe can see Fregehimself
or about
aboutconceptsand theirincompleteness,
"placing"hisownremarks
to theessenceoflogic.In ?III, I shallshowtherelation
'true'and itsrelation
of what a conceptscriptis
betweenthis 'placing'and an understanding
supposedto do.
III
In ?II I did notdrawattention
totheissueofrealism.It wasthere,all right;
and in PartIII I shallbe concerned
withit. I shalldescribea kindof realist
viewinordertoaskwhatweshouldmakeofitifwe tookFregeseriously.
Here
is theview.
Amongthe kindsof thingsthereare, are conceptsand objects. That
- say,thenumber4- is an object,is whyitis appropriate
fora term
something
of a propername,i.e. whatFregecalls
forit to have the logicalcharacter
or saturatedness;
that somethingis a conceptis why it is
completeness
ofa predicate,
fora termforittohavethelogicalcharacter
i.e.tobe
appropriate
'incomplete',to have one or more gaps for argumentexpressionsof the
sort,thesortthatstandforthingsof thekindthatcompletethe
appropriate
WHAT DOES A CONCEPT SCRIPT DO?
355
Our linguistic
thusproperly
expressions
conceptforwhichit is an expression.
ofthe
whichmatchestheindependently
fixedlogicalcharacter
havea character
for.
The
of
those
is
and
stand
character
things prior, belongs
thingsthey
logical
to themon theirown;and we can in theuse we fixforoursignsgetitrightor
wrong.
Well,whatis tobe madeofthat,ifwetakeFregeseriously?
did,
Wittgenstein
intheremarks
where
anditshows,forexample,
5.47intheTractatus,
following
he is discussing
therelation
sharesimply
betweenlogicandwhatall sentences
in beingsentences.
he
everylogical
Languageitself, says(at 5.4731)prevents
mistake.
Becauseitdoesso,we cannotgivea signthewrongsense.We cannot,
of
he means,givea signa sensewhichis inappropriate
to thelogicalcharacter
aboutwhatthelogicalcharacter
whatitstandsfor,cannotmakea mistake
is of
thethingwe wantthesignto standfor,and thengivethesigna kindofuse
thuswrongly
conceived.Thereis no such
to thelogicalcharacter
appropriate
to
in
a
combination
withothersigns,
a
as
wrong
thing allowing sign figure
into
the
which
the
combinations
unsuitable
thingmeantcan
given
possible
enter.
I need now to showhow Frege
But whyis thattakingFrege seriously?
butnotby
enablesus to see whatis wrongwiththekindofrealismI sketched,
view.
as an alternative
somekindofanti-realist
providing
The heartof therealistviewthatI havesketchedis thatlogicalcategories
the
ofthings.Whatwe can getholdofthrough
providea kindofclassification
to
tieddirectly
conceptscriptFregedevelopedis a notionof'logicalcategory',
oftheconceptscript,a first
kindof
kindsofargument
place.(In hisexplanation
ofthearguments,
and
argument
placeis characterised
bythe"completeness"
thatthey
othersbythenumberandkindsofargument
placesin thearguments
arethemselves
opento.I shallspellitoutinmoredetaillater.)Ifwe graspthat
of things?'will
notion,the question'Are logicalcategoriesa classification
simplynotbe askable;it willdropawayas meremuddle.
In thekindofrealismI sketched
theremightbe somequestionwhether
one
was correctin puttingSocratesin thelogicalplace foran object.One would
havecommitted
somekindoflogicalgaffeifone had gothislogicalcharacter
thathe wasan object,andhadgothimin thewrongkindof
wrongin thinking
But,forFrege,whatis inthelogicalplaceof
place,givenhislogicalcharacter.
in 'On
an objectis an object.Take, forexample,his well-known
discussion,
Conceptand Object'abouttheconcepthorse(TWF, pp. 45-8). It is truethat
on quitedistinct
whathe saystherecan be criticised
groundsand was nothis
forme at thispoint,though,
as an
itis important
ownlastwordon thematter;
illustration
ofhismethod.He arguesthatwhenwe saythattheconcepthorseis
'theconcepthorse'standsforan object.It is,logically
a concepteasilyattained,
to is thusin thelogicalplaceforan object
thesubject;whatitrefers
speaking,
is
not
is an object.He
andtherefore
goingonthefactthat'theconcepthorse'is a
356
CORA DIAMOND
grammatical
subject;thegrammatical
subjectof'A horseis whatyouprobably
werein yourlast incarnation'
wouldnot have been takenby Frege to be a
'On Conceptand
logicalsubject.As PeterGeachhaspointedoutin discussing
in
" with"special
not
it
had
an
Object',Frege
adopted 'ontology'
"deliberately
classesof objectthatweresurrogates
forconceptsand functions
in case of
need!" (Geach,op. cit.,p. 56). There is no possibility
of logicaldisasterin
to theconcepthorsein the
puttingSocratesin thewrongplaceor in referring
absenceofan obligingontological
of
category conceptsurrogates.
Here is how the realistpictureworks.We thinkthatthereare thingsin
letus suppose,sinceourrealistis readingFrege,objectsand
reality,
including,
the
concepts, conceptsthemselves
beingofvarioussorts.And,now,ofanyof
thesevariousthings,we thinkthatwe can think:it goes intothisone of the
or someother.Whatwe have got in mind,though,is the
logicalcategories
to itemsin diferent
intelligible
applicationof the categorytermsthemselves
We think,thatis, thatwe can thinkthatSocratesis an objectand
categories.
havegotsomething
oftheconceptwe thus
right.Buthowthenarewe thinking
takehimto fallunder?Whatis absolutely
essentialtotherealistframeofmind
is that'it is an object'canbe falselysaid ofsomethings.Otherwise,
we could
notthinkthatwe coulduse itinclassifying
in
The
use
of
things reality.
'object'
as a classifying
termrequiresthattherebe thingswhicharenotobjects.Now
in thinking
that'object'is a respectable
Fregeis unlikeWittgenstein
predicate
andobject
a respectable
twosortsofcomplication.
concept.(HereI amignoring
First,Fregedid, in his early"Dialogue withPiinjer",takeviewscloserto
aboutsuchconcepts;I am ignoring
thatpiecehere.Secondly,I
Wittgenstein's
ofthegeneral'concepthorseis nota
am,whereI can,ignoring
complications
thatwhenI callsomething
concept'kind,inordernottohavetosayrepeatedly
a concept,or,e.g.,askwhether
is a respectable
such-and-such
concept,I am
ofa first-order
unfortunately
usinga termwiththelogicalcharacter
predicate
whenI needtospeakabouta second-order
concept,thatI amusinga termas a
itis forFregea
propernamewhenI cannotwantthat,andso on.)But,although
respectable
concept,itis nottheconcepttherealistwantstouse (anditis going
toturnoutthatthereis no suchthingas whathewants).Since0 isan objectis a
first
orderconcept,
whatcanoccuras itsargument
isan object.
Thatis,itis true,
ofanything
thatyousayis an object,thatitis an object.You couldnothavesaid
or thoughtthatit was and have been wrong.'Everything
is an object'and
'Thereis nothing
thatis notan object'aretrue.Andso theconcept0 isan object
is uselessforclassifying
(whichis whyFrege had earlierdeniedthatsuch
conceptshavecontent).
What aboutsomething's
beinga concept?Can a corresponding
pointbe
madeaboutthat?Ifwecomplete
withan appropriate
theconceptthat
argument
we trytorefertoby'() is a concept',we do notalwaysgetthetruthvaluetrue.
The secondorderconceptwe wantcanbe used in classifying,
butonly(given
WHAT DOES A CONCEPT SCRIPT DO?
357
first
orderfunctions
withoneargument.
Frege'sviews)inclassifying
Concepts
in thesensein whichI havebeenusingthat
arenot,in fact,a logicalcategory
in discussing
aredefined
orderfunctions
expression
Frege.Conceptsand first
forthesamearguments
and can occurthemselves
ofanysecond
as arguments
orderfunctions.
thatthe
WhatI havejustsaidappearsto clashwithFrege'sownstatement
is
not
but
an
For
I
the
horse
a
said
that
concept
concept
object.
onlythingyou
withtheconceptwe trytorefer
couldclassify
toas '( is a concept'is first-order
functions
withoneargument;
andifthatis correct,
itseemsthattheonlything
we shallbe able to sayis nota conceptis sucha function.
was
Fregehimself
clearenoughthatthepredicateof thesentence'The concepthorseis not a
tothesecondorderconceptwewant(anymorethanthe
concept'doesnotrefer
fivewordsI justusedreferto it,anymorethanthepronounI havejustused
does).Let us finda wayto referto it.
'(x) (x is a dog is a truthvalue)'
'(x) (x is a treeis a truthvalue)'
is a truthvalue)'
'(x) (x's birthplace
ofthosethreesentences
is an expression
forthesecond
The commonpattern
orderconceptwe want:a conceptthatall and onlyfirstorderconceptsfall
do whatis forthemanalogousto fallingunder(whatFrege
under,or,rather,
calls"fallinginto").The secondorderconceptwe tryto referto by'( is nota
concept'we can referto by usingthe same patternwitha negationat the
togeta newsentence
Thereis no suchthingas completing
beginning,
pattern.
withthesubjectof 'The concepthorseis easilyattained',
either
which
pattern
has noargument
places.And thatis howwe shouldunderstand
Frege'spoint
that'the concepthorseis not a concept'.The sentencelooksas if it saysof
thatitis nota concept,
butifwe thinkaboutwhatwe trytosaywith
something
orthinking,
'( is nota concept',wecanseethatthereis no suchthingas saying,
whatthesentencelooksas ifit says.
Let me turnto thesecondorderconceptwe tryto referto by '( is a first
withone argument'.
orderfunction
Frege givesus one expressionforthis
''
whatwe tryto say
is
So
3
a
(E + 4)'
concept:'Eb(e)'.
wayofsayingproperly
is
that
for
3
is
a
function'
4
anyexpression
(PMC p. 136).Frege'spoint
by': +
thatits
thesecondorderconceptwe wantwillhaveto containan indication
withone argument
arefirst
orderfunctions
placeoftheirown.One
arguments
ofthepair
can makethesamepointby usingthecommonpattern
'(x) (x is a dog is an object)'
is an object)'
'(x) (x's birthplace
358
CORA DIAMOND
or (givena Fregeandefinition
of theconditional
forall propernames)of the
pair
'(x) (x is a dog : x is a dog)'
'(x) (x's birthplace: x's birthplace)'.
In thecase ofeachpair,thecommonpattern
servesto standforthesecond
orderconceptwe want.Each patternis completedintoa sentenceonlyby
forfirst
withone argument.
orderfunctions
expressions
We cannowgettothecentral
is an
thatsomething
pointhere.Everythought
is
the
to
as
we
misrefer
object true;every
'(
thought
gotbycompleting concept
is a firstorderfunction
is true,and correspondingly
for
withone argument'
ofarguments.
The whole
functions
ofhigher
ordersandwithdifferent
numbers
idea of classifying
notions
since it is
these
breaks
down,
anythingusing
tothinktrulyofanything
thatitdoesnotbelongtoa givencategory.
impossible
- in a sense- to makea mistakeaboutthelogicalcategory
of
It is impossible
thattheconcept
sincewhatwe call'makinga mistake'(e.g.,thinking
anything,
is a concept)willturnoutnotto
ofwhichwe saythatitis easilyattained,
horse,
itwas.Therewillbe no suchthingas
ofthethought
wethought
be thethinking
that.A goodconceptscriptwillmakeclearto us thatwe werein a muddlein
inmind,thatwewereexpressing
a thought
thatwehadgotsomething
thinking
at all.
werepigeonholes
We use theEnglishword'category'
as iflogicalcategories
fora kindof classification,
a logicalclassification,
of things.As iftherewere
in suchsomekindofclassthatlogicalcategories
all were,and as ifbelonging
and-sucha category
weresomeonething,andsomeitemsbelongedinoneand
othersin anotherand othersin yetothers;as if,thatis, to be a memberof a
weresomething
thatwentacrossthewholelot of categories.
logicalcategory
the
misleadsus intothinking
The languageof putting
thingsintocategories
thata plantgoesintothisortheothergenus;itis the
businessis likerecognising
intoclassesdepending
on thefirst
ofputtingobjects
language,grammatically,
thatsomething
orderconceptstheyfallunder.We speakofgrasping
belongsin
such-and-such
a category
or thatan itemdoes; thatis, we use whatordinary
whichcallfor
as variables
indicating
objects,inconstructions
languageprovides
in
I
now
am
suchvariables.And
what
saying,speaking(inevitably)
precisely
no
is
not
to
it.
There
is
is
to
into
a
thatway, that putsomething
classify
category
kindofclassthatlogicalcategories
all are.
Thereis an important
kindofobjectionthatpeoplemake.The thingsI have
thattheyareno kindofclass,andthatthat
beensayingaboutlogicalcategories,
is
of
the
itself
thatdoes not work-theseare
something
way putting point
of
on
one
feature
Frege'sconceptscript(duplicatedin the
entirely
dependent
in
is thatevery
for
the
That feature
a
Tractatus).
specifications conceptscript
of
in
takes
a
or
other
function
expression
arguments
place
predicate
argument
WHAT DOES A CONCEPT SCRIPT DO?
359
only one logicalcategory.If we have a conceptscriptwhichmeetsthat
andiftheconceptscriptallowsthepredication
ofcategory
terms
specification,
at all, we shallthenbe able to sayonlyof thethingsthatactuallyare in the
in questionthattheyare.Whenever
is notin a
we saythatsomething
category
is
in
we
shall
be
because
of
a
shall
what
wrong,
particular
category
category
only
will
andcategory
webe abletosaythatitis notin thatcategory,
predicates not
whether
we accept
havea genuineclassifying
use.That willbe theconclusion,
is
of
that
it
an
view
that
we
can
an
object or
say
object (e.g.)
Frege's
In
is
is
nonsense.
case
thereany
view
that
that
neither
actually
Wittgenstein's
So
the
of
classification
objecpredicates.
usingcategory
possibility genuine
tion runs-what we need to do is call into questionthe Fregean and
to havingnoargument
commitment
placesopentomorethan
Wittgensteinian
one category.Frege himself,the objectormightpointout, was willingto
aboutRussell'sparadox.He
whenhe was worrying
questionthecommitment
of
some
with
the
idea
places
opentopropernamesof
allowing
argument
toyed
and a thirdsortopento both
onelogicalkindofobject,someopento another,
(TWF, pp. 235-6). And,theobjectormightcontinue,evenif we couldnot
to considerthemove,therearegoodreasons
pointto Frege'sownwillingness
forit. We mustallow thatthereare genuinequestionsof the sortFrege's
originalviewrulesout. Take the questionwe maywishto ask about 'The
whether
theconcepthorsethatwe arespeaking
concepthorseis easilyattained',
whichitis possibleforBucephalusandotherthingsto
aboutthereis something
we aresayingifwe shouldwanttosaythatit
fallunder.Is therenotsomething
is not?Or indeedifwe shouldwantto saythatit is?Can we notbe rightor
oftwopossiblethoughts
here,
wrongaboutthis?Is therenota genuinethinking
whichcontradict
eachother?
thoughts
Let me lookfurther
at theobjection.The idea ofargument
placesopento
morethanone kindofthingis absolutely
(theobjectormayclaim),
necessary
thatcanbe saidtrulyofitemsofdifferent
since(he says)therearethings
logical
to theideathatthere
is committed
kinds.It mayevenseemthatFregehimself
After
aresomethingswecansayofitemsindifferent
all,itmightbe
categories.
orderrelations
orderconceptsand first
asked,doeshe notthinkthatbothfirst
thattheyare incomplete,
arefirstorder,i.e., takethesamekindofargument?;
thoughnotin thesameway?;thattheyhaveas theirvaluesonlythetwotruth
values?;thatneitherof themcan fallundera firstorderconcept?But the
appearanceof his sayingthe same thingof conceptsand relationsis itself
inordinary
createdbywhathewouldtaketobe an imprecision
language.There
fallunder
orderrelations
orderconceptsand first
is no conceptthatbothfirst
in
an
that
fall
and
none
adequateconceptscriptthereis
(properly,
they into),
for
even
looks
like
an
that
expression sucha concept.
nothing
on thesideofsayingthatsome
to
we
Verywell; stoptrying getFregehimself
Is it notnevertheless
in
of
items
different
can
be
said
categories.
things
truly
360
CORA DIAMOND
true?An examplethatwasoffered
me in discussion
wasself-identity:
surely(it
was said),it is correctto saybothof anyobjectand of anyconceptthatit is
identicalwithitself.
in a Fregeanframeofmind.How closewe can
Supposewe lookat thisfirst
what
to
of
it
is self-identical
canbe seenifweconsider
a
that
get saying concept
itis tosayofa conceptpickedoutonewaythatitis thesameas a conceptpicked
out anotherway.GivenFrege'sviewofwhatwe maylooselycall conceptual
tobe fis thesameas whatitis forsomething
whatitis forsomething
identity,
to be g ifandonlyifwhatever
is fis g, and conversely
(PW, p. 120).Applying
toa concept,
the
or,rather,
Frege'sview,then,whatitis toascribeself-identity
fallsunderit falls
nearestwe can cometo doingthat,is to saythatwhatever
underit,'(x) (x is bald - x is bald)' expresses
whatsomeoneis trying
properly
to say if he saysthatbald,theconcept,is self-identical.
(Since forFregea
sentence
toa truth
refers
value,we canalsowriteitthisway:'(x) (x is bald = x
is bald)'.)
If we followFrege,then,we shallsaythatordinary
languagegivesus a false
thatwearesayingthesamethingoftheconceptbaldandofSocrates
impression
inwhich
Ifwehavea notation
whenwe sayofthemthattheyareself-identical.
we sayofSocratesthathe is self-identical
thisway:'Socrates= Socrates',and
ifwe saywhatwe weretrying
to sayabouttheconceptbaldthisway:'(x) (x is
is
bald = x is bald)', it wouldnotso muchas occurto us thatself-identity
ascribableto objectsand conceptsboth.The objectionthatit is
something
bothtoconceptsand
ascribable
possiblefortheretobe thingslikeself-identity
toobjects,andthatwe therefore
withargument
needexpressions
placesopento
ofdifferent
stated)on
expressions
logicalkindsrests(at leastas itwasoriginally
shows.
whatwasatissue.Thatis whatthecaseofself-identity
simply
assuming
We need,though,to see moreclearlywherethe difference
betweenthe
viewandFrege'slies.I saidthatwhatFrege'sconceptscriptgivesus
objector's
is a notionofcategory
tiedtokindsofargument
place.Butwhatcountsas a kind
on
needs
of
ofargument
placedepends the
logic,giventhewaywe analysea
sentencecontainingsuch-and-suchtermsthat can be taken as logically
The difference
betweenwhatidentity
comestoforobjectsandwhat
significant.
itcomestoforconcepts(centralinourexample)is a reflection
oftheneedlogic
has to allowfordifferent
kindsofsubstitution
salvaveritate.
for
To analysea sentenceis to takeitas containing
oneormoreexpressions
in
of
one
to
take
the
which
occur
more
than
and
(each
place)
arguments
may
restofthesentence
as havinga different
sortoflogicalrole.For thatpartofthe
sentence(whichI shallcalltheleftover
part)to havea determinate
meaningis
forthetruthvalueofthesentenceas a wholeto be fixedforeachmeaningful
inour
thatcanbe substituted
fortheargument
argument
expression
expression
sentence(if we are construing
our sentenceto have only one argument
our
foreachpair of argument
(ifwe are construing
expression),
expressions
WHAT DOES A CONCEPT SCRIPT DO?
361
sentence
tohavetwoargument
andso on.Thereis a fundamental
expressions)
kindofanalysis,
availableonlyin thecase ofsomesentences:
we canconstrue
thesesentences
tohave,as theonlyargument
a proper
theycontain,
expression
name. Frege uses metaphorto explainwhathis terms'propername' and
or 'saturation'.
We can see
'object' mean,the metaphorof 'completeness'
ifwe recall
these
of
what
is
at
beinggot by
expressions
something
metaphorical
of logic: how the truthvalue of the
thatall analysisreflects
the interests
in termsoffunction
sentenceas a wholemaybe construed
and argument.
We
insome
as leftover
havetolookat thewaylogictreatstheuseofoneexpression
contextsand as argumentin others,and at the way it tells,in certain
whether
we havedefinedone leftover
or two.
circumstances,
expression
If whatwe viewas theleftover
in
one
sentence
can
takentobe
be
expression
in
in
the
an argument
another
same
otherwise
one,
(or
construed)expression
will insistthatthe
if,thatis, logic will allow such an identification-logic
now viewedas argument,
now viewedas leftover,
makethesame
expression
to thetruthor falsity
ofwhatis said.Logic willthusinsistthatit
contribution
in
salva veritate
carrywithit in thetwocases thesamerulesforsubstitution
substituted
for
it
what
can
be
salva
sentences.
Whenitis a leftover
expression,
is anyotherexpression
whichfixesin thesamewayas oursthetruthor
veritate
ofsentences
as a wholeforeachmeaningful
expression
( orset
argument
falsity
Ifa leftover
ofargument
expression
expressions)
bywhichitcanbe completed.
itmustcarrywithit
is tobe viewedas occurring
alsoas an argument
expression,
inthelatteruse exactlythesamerulesforsubstitution
as ithadas
salvaveritate
totruthvalueas inits
oritwouldnotbe makingthesamecontribution
leftover,
in whichwe
forit in contexts
use as leftover.
That is,whatcan be substituted
can be specified
areviewingit as an argument
onlyin termsofthe
expression
foreverymeaningful
truthor falsityforsentencesas a wholedetermined
argumentexpression(or set of argumentexpressions)that our argument
Hencewewriteitin
canitselfbe completed
expression
byin itsuse as leftover.
or as
a conceptscriptwithargument
places,whetherit occursas leftover
toitsidentity
as faras logic
sincetheseplacesareessential
argument
expression,
is concerned,its identity(that is) as an expressionwith such-and-such
in
to thereference
ofwholesentences
reference,
makingthesamecontribution
whichit occurs.(It thus belongsto its identitythatit can onlybe used
predicatively.)
Let us now trysupposingthatwe had a leftoverexpressionwithone
forsentences
as a wholewasfixed
argument
place,andthatthetruthor falsity
withno argument
forit, forall argumentexpressions
places of theirown
withoneargument
expressions
placeoftheir
(propernames)andforargument
two
sort.Whatlogicwouldsee is thatwe haddefined
ownofsomedeterminate
kindsofcontribuwhichmadeentirely
different
leftover
expressions
equiform
inwhichtheywereleftovers.
The expression
would
tiontothewholesentences
362
CORA DIAMOND
sometimes
standforonekindoffunction
andsometimes
foranother;
in itstwo
different
referents.
usesitwouldhavetwoentirely
One ofFrege'scriticisms
of
Hilbertis thathe defines'point'so thatit has twomeaningsrelatedto each
otherin thatway(PMC, pp. 93-4). We do notmaketheissueshereclearifwe
no argument
saythatforFregeand Wittgenstein
placecan be opento more
thanone kindofargument
If
it
were
expression.
opento morethanone,logic
wouldnotethattheleftover
was
to
contexts
expression beingusedin different
maketwodifferent
kindsofcontribution
totruthvalue;logicwouldthuscount
it as twoleftoverexpressions
withdifferent
kindsof reference.
Given an
a
kind
makes
of
contribution
determinate
which
to
truth
value
expression
whereverit occurs,thereis no suchthingas puttingtwo different
kindsof
intoanyofitsargument
argument
expression
places.A conceptscriptdoesnot,
ruleoutdoingthat,forthereis no suchthing.Whatitdoesruleoutis
therefore,
theuse ofonesignin twologically
different
ways.Anditcando thisbya mode
ofwriting
leftover
so thatthekindofargument
foreach
expressions
expression
argument
placeis clearlyindicated.
For thereto be logicalanalysisof any sentence,i.e., forthereto be the
ofitas argument
construal
andleftover
expressions
part,logicwillneedrulesof
at leasttwo different
sortsforsubstitution
of the expressions
salva veritate
one sortfortheleftover
pickedoutbytheanalysisas logically
significant:
part
and at least one othersortforthe argumentexpressions.(The rules for
substitution
in factfixthekindofrule
theargument
characterising
expressions
theleftover
part.)I arguedearlierthat,ifthereis anypossibility
characterising
of construing
in one sentenceas argument
whatwas leftover
expressionin
wemustusethesamerulesforsubstitution
inbothcases.
salvaveritate
another,
These tworequirements
ofkindsofruleforsubstitution
a structure
generate
In a conceptscript,
salvaveritate
ofexpressions.
willbe clear;and
thestructure
whichcan occuras leftover
willbe written
withan
anyargument
expression
of itsownargument
indication
ruleto
placesand of thekindof substitution
in themare subject.In a conceptscriptin whichthe
whichtheexpressions
ofargument
indication
therewillbe
placesis thuscarriedout systematically,
withnoargument
written
them
argument
expressions
properly
places.Writing
thatwayindicatesthekindofsubstitutability
salvaveritate
theyhave:thatof
names.The substitutability
rulesforpropernamesarewhatcorrespond
proper
on thelinguistic
in theproperlimitedsense.(Anyotherkindof
sideto identity
forexpressions
notto identity
salvaveritate
butto
substitutability
corresponds
a relation
analogousto it,on Frege'sview.)An object,Fregesays,is anything
theexpression
forwhichdoesnotcontainanyemptyplace(TWF p. 32). Such
an expression
canneverbe identifiable
as thesameexpression
as one whichin
someanalysisof somesentenceis a leftover
It
is
fora
expression. impossible
in
name
to
a
the
of
no
be
there
such
is,
proper
predicate;
eyes logic,
thingas
that.The verycloseconnection
betweenbeingan objectand beingcapableof
WHAT DOES A CONCEPT SCRIPT DO?
363
intherelation
ofidentity
is extremely
for
standing
properly
speaking
important
the
When
he
considered
Russell's
paradoxbysplitting logical
Frege.
avoiding
ofobjectsintoproperandimproper
thathe would
objects,he thought
category
have to have one relationof identity
whichcould take as its
nevertheless
bothproperand improper
objects.That coursemakesthemess
arguments
the
solution
worse
thanit wouldotherwise
be; and
generatedby
proposed
in
solution.
is
is
fact
that
What
the
rejected
interesting
Frege
way his
of
how
the
illustrates
he
saw
one
description
proposal
logicalcategories: thing
he thoughthe had to leave in place in any acceptablesolutionwas the
connection
betweenbeingan objectand beingable to standin therelationof
formthatit is inconceivable
"a
givento us in sucha specific
identity, relation
thatvariouskindsofit shouldoccur"(TWF, p. 235). Theremaybe relations
whichis nota function
thereis,as faras logicis
analogoustoit,butforanything
it.
concerned,
only
tothesamethingtwice
The upshotofall thisis thatyoucannotbe referring
firstthisand thenthatto it) iftheexpression
(cannotbe ascribing
youuse to
to thetruthor
different
kindsof contribution
referto it makestwoentirely
tothesamethingtwice
inwhichitoccurs.You canrefer
ofthesentences
falsity
withdifferent
twiceover
senses,butnotbythesameexpression
byexpressions
thatitcarrieswithitaredifferent
salvaveritate
whentherulesforsubstitution
inthetwocases.It wouldthenrefertotwodifferent
Whatkindofthing
things.
about,is notseparablefromthe
to,talkingabout,thinking
youare referring
forit. If we say
to truthor falsity
madebytheexpression
kindofcontribution
thatSocratesfallsundertheconceptbald and thattheconceptbald is selfidentical,we can takeourselvesto be speakingabout the same thing,the
tobe sayingoftheconcept
conceptbald,twice,onlyifwe do nottakeourselves
baldwhatwe sayofSocrateswhenwe saythatheis self-identical.
inthedisputebetween
I wentintoall thistoshowwhatis involved
Fregeand
itselfinoneway,butitgoes
theobjector.
The disagreement
expresses
naturally
witha muchmorefundamental
The relatively
disagreement.
misleading
wayof
is: it is aboutwhether
thereare concepts,likeselfputtingthedisagreement
underwhichthingsin different
logicaltypescan fall.
identity,
To showwhatI
is logicalanalysis.
Whatis at theheartofthedisagreement
candidatethat
tosetalongsideself-identity
another
meanbythat,I needbriefly
comeup withwhentheylookforconceptsapplicabletoanything
philosophers
kindof thingthe
The idea is thatwhatever
at all: beingan objectofthought.
conceptbaldis,we can at anyratethinkaboutit,so itand Socratessharethis:
ofthought.
theyare objects
thatI quotedearlier:"I do notbegin
Againstthis,we haveFrege'sremark
I comeby
orjudgment;
toforma thought
withconcepts
andputthemtogether
If
the
thethought". youtake judgment
itself
thepartsofa thought
byanalysing
as primary,
as Fregedoes,whatyoufindis thatyoucanindeedtakebald- the
364
CORA DIAMOND
ofthought
atall.
onlybyitsnotbeingan object
concept- as anobjectofthought
in
That is,in thesenseinwhichyouthinkaboutan objectbyhavinga thought
as object(expressedin a sentencein whicha propername
whichit figures
inwhich
to itoccurs),youthinkabouta conceptbyhavinga thought
referring
it occursas a concept,i.e. a thoughtexpressedby a sentencecontaining
an
There
is
to
it.
no
used
referring
category-unamexpression
predicatively,
andyouonlythinkthatthereis ifyou
biguousthing,
beingan objectofthought,
thinkyoufindtheconceptbaldon itsown,whenyoustartlookingforobjectsof
thoughton theirown. The conceptbald (if we reallymean the concept)is
you
something
youcan thinkabout,all right-onlybecauseit is notsomething
we thinkabout',
can thinkabout.Unlesswe areverycarefulwith'something
aboutsomething
is belongsto whatit is,
we go wildlywrong.Whatthinking
of
whatkindofthing,and thatis shownin theuse ofa termin theexpression
thoughts.
in thinking
thatone can
fromFrege fundamentally
My objectordiffers
of
use
terms
from
the
kind
havein the
it
is
one
is
about
what
thinking
separate
one
can
be
about
is
he
believes
that
That
of
thinking
why
expression thoughts.
one and thesamething,theconceptbald,whenone saysthatSocratesfalls
underit and thatit is also whatSocrateshimself
is, namely,self-identical.
Thereis a sensein whichthedisputants
pass each otherby.The objector
thesisthatFregehas(the
ascribestoFregea viewofa certain
kind,a substantial
objectorthinks)builtinto his conceptscript,but whichcan be discussed
of
can be thought
The supposedFregeanthesisis thatnothing
independently.
thingsofmorethanone logicalkind,and theobjectorbelievesthatithas been
whatcan go intoargument
builtintotheconceptscriptin therulesgoverning
conceptscriptwouldallow Socratesand the
places.He thinksthata better
of self(e.g., ascriptions
conceptbald intothesameplace in somethoughts
andwouldallowSocratesbutnottheconceptbaldintoa placeinother
identity)
baldness.One coulddisagreewithFregeabout
thoughts,
e.g.,thoseascribing
inthefundamental
whata goodconceptscriptwouldallowwithout
disagreeing
wayourobjectordoes.He doesnotso muchrejecta viewofFrege'sas notsee
ofletting
theuse teach
whatkindofviewit is. He doesnotsee thepossibility
about,wherelogicalanalysisis whatshowsyouhowa
youwhatyouaretalking
it makes (what kind of
termis being used, what kind of contribution
in whichit occurs.
contribution
logicsees it making)to thesentences
Some yearsago Hide Ishigurosaid thatwe shouldgiveup themisunderofmeaningin
ofWittgenstein,
thathe wentfroma 'Naming'theory
standing
ofmeaninginhislaterwork.2She arguesthatin
toa 'use' theory
theTractatus
as Russell
He neverthought,
sensehe alwayshelda use theory.
an important
2
ed. P.
ofNames",in StudiesinthePhilosophy
ofWittgenstein,
Ishiguro,"Use andReference
Winch(London,1969).
WHAT DOES A CONCEPT SCRIPT DO?
365
of its use in
did, thatthemeaningof a namecan be settledindependently
of
the
in
the
to
the
Tractatus
he
held
that
identity
objectreferred
propositions;
in
is
the
of
name
She
traces
this
a
name
settled
use
the
propositions.
by
onlyby
viewback to Fregeand extendsit to theuse of termsforthingsin logical
to makeclearis
otherthanthatofobjects.WhatI havebeentrying
categories
thatat the rootof the disagreement
betweenmy objectorand Frege is a
between
thewayweusetermsandwhatitis we
abouttherelation
disagreement
whichMiss IshiguroascribestoFregeis a
aretalking
about.The "use" theory
alientomyobjector.In hisposition
modeofthinking
aboutthatrelation
utterly
thatrecursin laterphilosophy.
The
vis-a-visFrege'swe can see something
in
looks
the
it
from
direction
objector(so seems,
resolutely
wrong
Frege'sside)
is Miss Ishiguro's,
fromher
to tellwhathisowntermsreferto.The metaphor
form
of
between
of
the
the
(in
disagreement
Wittgenstein hislater
description
know
from
own
oftheirmental
to
their
awareness
and
those
who
claim
works)
or talking
about.
processeswhattheyare thinking
ofwhatit
The disputebetweenFregeand theobjectorshowsus something
of our
meansto speak of philosophyas a battleagainstthe bewitchment
therecan
ofwhatkindofdifficulty
(PI, ?109),showsus something
intelligence
of the intelligence
That battleagainstthe bewitchment
be in philosophy.
changeditsformas therolein itoflogicalanalysischanged.Butwhatremains
thatturnsus awayfromthe
is the idea of it as a battleagainstsomething
inwhichwhatneedstobe seenis opentoview,andalsotheideathat
direction
whatwe needto see- ifonlywe couldsee thatthatis whatwe needto see!- is
howourtermsthemselves
areused.
Whatit was our objectorwas originally
objectingto mayhavesunkfrom
ofwhatthere
a kindofrealismaboutthelogicalcharacter
view.I haddescribed
of
is, a realismthatgoes withtheidea thatwe can do a logicalclassification
in whichtheybelong,wherelogical
things,
putthemintothelogicalcategories
categoriesare thoughtof as themselvessome kind of class, and we can
rightor wrong.I arguedthat if we
supposedlyget such a classification
thatkindofrealismdrops
whatis showninFrege'sconceptscript,
understand
away,sentencesexpressingit being recognisedas so much muddle.The
of
claimedinthefirst
placethatthereisroomfora logicalclassification
objector
is
no
in
there
claimed
also
that
good
general
thingsand,to backthatup, he
of a conceptto itemsin different
reasonto proscribeall ascriptions
logical
I wantnow to turnbackto a questionaboutrealismand whata
categories.
directions.
conceptscriptdoes,whichI shallgetto fromtwodifferent
I
have
said is all very
that
what
someone
complaining
First,imagine
the difference
that
For
it
is
not
but
Frege actuallysays
Frege.
interesting
butis founded
is notmadearbitrarily
and secondlevelfunctions
betweenfirst
thesameabout
deepinthenatureofthings(TWF, p. 41). He wouldsayexactly
andconceptand
andobject,betweenrelation
betweenfunction
thedistinctions
366
CORA DIAMOND
between
so on.ButhaveI notbeendenying
thathebelievesinanysuchrelation
I
that
not
been
and
the
of
Have
structures
nature
things?
suggesting
logical
it
do
likes
much
the
nature
can
what
and
logic
ignore
ofthings?
pretty
Secondly,imaginesomeoneascribingto Frege the view thatthereare
whichcannotbe expressedin language,but whichare nevertheless
thoughts
from
different
aretotally
Thus,ifwe saythatobjectsand functions
thoughts.
is
in
such
a
use
of
and
there
eachother,thatis a logically
language,
improper
of
what
we
ridofthelinguistic
and
caseno wayofgetting
impropriety,
saying
tosayinlogically
decentlanguage.On theviewweareimagining
aretrying
(we
it ascribedto Frege),althoughwe cannotsaythethingwe are
are imagining
to say,whatwe arestruggling
withis nevertheless
a truebutinexpresstrying
iblethought.
We shallcomeroundto thefirst
pointifwe startwiththesecond.I findit
ofsuchnonsensetoFrege(orabout
tobe tolerant
abouttheascription
difficult
inthecaseofWittgenstein).
Whenthereis nowayofsaying
thesameascription
whatwe aretrying
to say,whatwe comeoutwithis in facta kindof
properly
to no ineffable
truth.WhatFregethinksis that
nonsense,and corresponds
initwhich
ofordinary
wecanformsentences
an inadequacy
language,
through
of any
are acceptableaccordingto itsrulesbutwhichare nottheexpression
ifweareledto
It is possibletobecomeclearaboutwhathashappened,
thought.
see how thoughts
are expressedin a languagemorenearlyadequateby the
oflogic.In grasping
ofthedistinctions
embodiedin
thesignificance
standards
A truthis a truthabout
truths.
thatlanguage,we do notgraspanyineffable
a truethought
(thatis) is aboutwhatever
logicmayconstrueit as
something;
embodiedin theconceptscriptarenotwhat
beingabout.But thedistinctions
can be about.
anythought
The reasonshavealreadybeentouchedon. A conceptscript,I said in ?11,
doesnotgetsomething
external
to itselfright;ifit is notan adequateconcept
itis notthrough
external
toitself
wrong.On theother
script,
something
getting
a conceptscript.
hand,itis notthecasethatjustanything
goesin constructing
butare
As Fregehimself
embodiedin itarenotarbitrary
says,thedistinctions
foundeddeep in thenatureof things.We shouldnot ignorethatremarkof
he attachedto suchpoints- butwe
Frege's;we shouldgiveit theimportance
it.We needto ask:what
haveto be particularly
carefulnotto misunderstand
I
whichtheconcept
the
distinctions
natureof things?
as
have
For,
justsaid,
it.
to
must
in
it
not
fixed
external
have
are
script
Nothingexternal
byanything
Wherethenis thereality
to itfixesitslogicalstructure;
butit is notarbitrary.
liesinit.There
thatfixeswhatdistinctions
mustbe embodiedinit?Thatreality
is an order,a logicalorder,inthought
and inlanguage.Thought'sbeingabout
itshavinglogicalorder,itsbeingthataboutwhichthequestionoftruth
things,
- language,
in language
whichitselfhasa single
canarise,itsbeingexpressible
- all thesecome
in
as
is
of
so
far
it
logicalgrammar
thoughts
capable expressing
WHAT DOES A CONCEPT SCRIPT DO?
367
tothesame.Thoughtis aboutthings,
butthelogicalorderwhichis partofwhat
it is forthereto be anyaboutness
is notitselfone ofthethingsthought
can be
about. The distinctions
betweenfunctions
and objects,betweenfirstand
secondorderfunctions
and so on- thosedistinctions
areindeedfoundeddeep
in thenatureof things.But to understand
someonewho saysthatis to have
thekindof 'placing'sucha sentencerequires.To takeit to be an
understood
truthis a trulyperversemisreading,
'ineffable'
an attemptto represent
to
oneselfastateofaffairs
whilepretending
thatno representing
tooneself
is going
on. Puttingthe matteranotherway: the distinctions
in questioncome out
in
a
which
marks
them
clearly language
systematically,
byhavingexpressions
whichmakeplainthelogicalcharacter
of whattheystandfor.But in sucha
whichrefers
toanyoftheselogicaldistinctions.
languagethereis no expression
enables
us
form
to
like'the
language
Ordinary
referring
apparently
expressions,
between
distinction
first
andsecondorderfunctions'.
Butnowthehardthingis
tolearn,fromthefactthatthereareno corresponding
ina logically
expressions
that
the
distinctions
are
not
of
And
adequatelanguage,
objects anythoughts.
is correct
in sayingthatthesedistinctions
whatever
are foundeddeep in the
natureofthings,
itscorrectness
willnotlie in theimagery
thataccompanies
it,
offunctions
withtheirfunctionhood,
with
their
objects
objecthood,
lyingthere
withtheirdistinct
naturesin thenatureofthings.
Now oncemoreaboutrealism.Wittgenstein's
laterworkhas a significance
forquestionsaboutrealismandanti-realism
whichcanbe seenonlythrough
its
relation
to thebodyofideasI havebeentalking
about.Andwhatcharacterises
thatbodyofideasis itsstyleofanswertoquestionsofthegeneraltype:Where
is
thereality
whichmustguideus here?Whereis thereality
towhichourmodeof
mustbe responsible?
thought
I described
a kindofrealismin ?III, a realismwhichdependson theideaof
an external
relation
betweenthelogicalcharacteristics
ofthingsandthelogical
features
builtintoour modesof expression.
The kindof realistI imagined
thatthelogicalcharacteristics
ofthings
andthelogicalstructures
ofour
thought
modeofexpression
tomatchbutthattheymightnotdo so. It is actually
ought
to it whichcan be conceivedby
partofsucha viewthattheonlyalternative
someonewho holds it is: in our modes of expressionanything
goes. The
thatis,areconceivedthisway:either
thereis something
external
to
alternatives,
themodeof expression
fixedlogicalcharacter
of things)
(the independently
whichis a measureofthelogicaladequacyofthemodeofexpression,
orthereis
no measureofadequacywhichis in a hardlogicalsenseunarbitrary.
Thereis
also an anti-realist
viewcharacterised
thesame
bytheacceptanceofprecisely
Bothviewsthenconceivethesituation
thisway:thereis a
pairofalternatives.
of and externalto our modes of thoughtand
logicalrealityindependent
or thereis not and the logicalstructures
of our thoughtand
expression,
are
expression fundamentally
arbitrary.
368
CORA DIAMOND
of a conceptscriptas a tool forphilosophical
Frege,in thedevelopment
us
allows
to
clear
of
thetwoalternatives,
toleavethembehind.He
thought,
get
allowsus to thinkabout language,any languagein whichthoughts
can be
as
in
or
which
not
be
there
for
expressed, havingsomething it,
may may
clearly
us to see. A languageis morenearlyadequateby the whollyunarbitrary
oflogicin so faras itmakessystematically
standards
clearwhatis therein it,
whatit shareswithall languages.The standardof logicaladequacyis no
in themeasure.He lets
external
one,butthatdoesnotimplyanyarbitrariness
us see thatthereality
which
the
of
a
by
adequacy conceptscriptis measuredis
notexternalto it. In thatwayhe showsus a generalpossibility
of answering
in
where
the
which
is
to
such-and-such
we are
reality
questions
doing
us
a
see
somewhere
else
than
where
we
were
responsible.
By letting
reality
he
teaches
us
a
new
to
I
understand
such
The
realist
questions.
looking,
way
and
the
who
takes
the
to
a
view
share
his,
imagined,
person
opposite
eitherthereis something
or
wherewearebothlookingfora reality,
conception:
thereis not,and we are withoutobjectivestandards;thatis whattheyboth
think.
later
Look somewhereelse: thatis whatwe can hear in Wittgenstein's
lookwhereyoudo notthinktherecan be anyreasonforlooking.
philosophy;
Thatis theretobe heardintheTractatus,
anditsimply
makesclearera message
alreadyto be heardin Frege'swork.For him,theconceptscriptwas a toolof
A goodconceptscriptwouldlaybeforeus clearlywhat
intellectual
liberation.
we need to look at, and thushelp breakthe domination
of wordsoverthe
humanmind;itwouldhelpfreeourthought
fromthetrammels
placeduponit
to thepromptings
of logic(Bs, pp. xii-xiii;cf.TLP,
by languageinattentive
3.325).
totieFrege'suseoftheconceptscripttoocloselytotheuse,in
It is a mistake
theanalytic
oftechniques
oftranslation
ofphilosophically
tradition,
interesting
claimsintoan adequatelanguage.One pointoflookingat theexample'logical
area kindofclass'wasto see thesignificance
ofthefactthatthereis
categories
notranslation
ofit intoan adequatelanguage.We needstillto getclearabout
whatis radicalin Frege'svision;we can do thatonlyby gettingclearthe
relationbetweenthe droppingaway of some philosophical
viewsas mere
muddleandtheideaofthetruth
oflogicas notsomething
setoveragainstwhat
itme suresbutinternal
to it.
of Virginia
University