Professional Documents
Culture Documents
David Kaplan Demonstratives
David Kaplan Demonstratives
945. t / :",{':il
L7
l This paper rtas prepared for and lead (with omissions) at a sympoeium on Demon-
straliv€ at the March 19?7 meetiugs of the Pacific Division of the American
Philosophical Association. The conunentators rvere Paul Benacerraf and Char.les
Chastain. Much of the lnater.ial, includ.ing the formal system of section XvIII, was
originally presented in a series of lectures st the fabled 1971 Sulntner Institute in
the PhiloEophy of Language held at the Udversiiy of California, Irvine. @ 1977
try David l(aplan.
481
Demonstratives 483
Table of Contents
Preface
Preface 483
L lntroduction 489 In about 1966 I rvrote a paper about quantiflcation into epistemological
II. Demonstratives, Indexicals, and Pure Indexicals 489 contexts. There are very difficult metaphysical, logical, and epistemo-
logical problems involved in providiag a treatm€nt of such idioms which
III. Two Obvious Principles 492
does not distort our intuilions about their proper use and rvhich is up
IV. Remarks on Rigid Designators 492
to contemporary logical standards. I did not then, and do not now,
V. Argument for Principle 2: Pure Indexicals 498
regard the treatment I plovided as fully adequate, And I became more
VI. Terminological Remarks 500
and more intrigued rvith problems centering on what I would like to ca.ll
VI. (i) Content and Circumstance 500
lhe sentanlics of direct reference. By this I mean theories of meaning
VI. (ii) Character CUD
according to rvhich certain singular terms refer direcily rvithout the me-
VII. Earlier Attempts: Index Theory DUT
diation of a Fregean ^9r'znas meaning. If there a,resuch terms, then the
ViII. Monsters Begat by Elegance 510 proposition expressed by a sentence containing such a term would involve
IX. Argurnent for Principle 2: True Demonstratives 512 individuals directly rather than by way of the "individual concepts" or
IX. (i) The Arguments o r.t
"manners of presentation" I had been taught to expect. Let us call such
IX. (ii) The Fregean Theory of Demonstrations 5r4 putative singular terms (if there are any) d,irectly referenlial lerms and
IX. (iii) The Fregean Theory of Demonstratives 516 suclr putative proposiiions (ifthere are any) singular proposilions. Even
IX. (iv) Argument Against the Fregean Theory of
if English contained no singular terms rvhose proper semantics rvas one
Demonstratives 516 of direct reference, could we determine to introduce such terms? And
X. Fixing the Reference vs. Supplying a Synonym 518 even if rve had no directly referential terms and introduced none, is there
X. (i) Reichenbach on Token Reflexives 519 a need or use for singular propositions?
XI. The I\{eaning of Indexicals 520 The feverish developrnent of quantified modal logics, more genera"lly,
XII. Dihat 521 of quantified intensional logics, of the 1960s gave rise to a metaphysical
XIII. Contexts, Truth, and Logical T\uth 522 and episternological nalaise regarding the problem of identifying individ-
XIV. Summary of Findings (so far): Pure Indexicals uals across worlds-what, in 1967, I called the problem of "Trans -lVorld
XV. Further Details: Demonstratives and Ifeir Lines." This problem was really just ihe problem of singular propo-
Demonstrations 524 sitions: those which involve individuals directly, rearing its irrepressible
XVI. AlternativeTreatrnentsofDemonstrations 527 head in the possible-world semantics that rvere then (ar.rd are now) so
XVII. EpistemologicalRemarks 529 popular.
XVIII. The Formal System It was not that a,ccording to those semantical theories any sentences
XIX. Remarks on the Formal System 546 of the languages being studied \,\r'elethemselves taken to express sin-
XX. Adding 'Says' 553 gular propositions, it lvas just that singular propositions seemed to be
XXI. Russell on Egocentric Particulars and Their
needed in the ar.ralysisof the nonsingulat propositions expressed by these
Dispensabiliiy 557 sentences. For example, consider
XXII. On Proper Names 558
(0) lr(.Fc A -EFn).
IFx
482
484 David l(aplan Demonstratives 4B5
(under some assignment of an individual to the variable 'x'), we must while Frege's solution: that a and p, ihough referring to the sane thing,
first determine rvhether the proposilion expressed by its component do so by rvay of different senses,rvould be bloclied. Also: because of the
fact that tite component of tl.re proposition is being determined by the
Fx
individual rather than vice versa, rve have something lilie a violation of
(under an assignment of an individrral io the variable 'x') is a necessary tlre famous Fregean dictum lhal lltere is no roq.d doclr from denotation
proposition. So in the coulse of analyzing (0), u'e are requiled to deter- to sense fpropositional corrponent]. (Recently, I have come to ihinli thai
mine the proposition associated with a formula containing a free vati- if l'e countenance singular propositions, a collapse of Frege's intensioual
able. Now free variables under an assignu-rentof values are paradigms ontology into Russell's talies place.)
of rvlrat I lrave been calling direclly referenlial terms. Iu determinilg I can drarv son-relittle pictures to give you an idea of the trvo liinds
a sema,ntical value for a formula containing a free variable rve rnay be of semantical theories I want to coutrast.
given a. aalue for the variable-that is, an iudividual drarvn from the
universe over rvhich the variable is taken to range-but nothing more.
A variable's first and only meaning is its value. Therefore, if rve ale to
associaie a proposilion, (not merely a truth-value) with a formula con- Fregean P ictul.e
taining a free variable (with respect to an assignment of a value to the
variable), that proposition seems bound to be singular (even if valiant
PROPOSITIONAL COMPONENT
attempts are made to disguise this fact by using consiant functions to (a concept, something
s.^."
imitate individual concepts). The point is, that if the component of the ^"{..*tlo.f"lr{
propositiou (or the step in the construction of the proposition) u'hich
corresponds to the singular terrn is deterrnined by the individual and
lr
""'":,"1'1""/
- ^\'" .oar -dl- $"/ | :
Xl:"",i::Ti:ffL1i"""
the individual is directly determined by the singular term-rather than ..t1pc.f**"tZt | 3 1' ti" relation is, in general.
t,he individual being determined by the component of the propositiou,
rvhich is directly determined by the singular telm-then rve have rvhat I
q:+'l*".*{"i'f,cf
-*^.*oi.oi/ d
l{ :-".-':liil"i"hi"ii'"i*r'.r.rr"
the concept,i.e., who, uniquelv,
I I- :"0":
has the qualities)
call a siugular proposition. [Russell's semantics rvas lilie the sernautical " o\-,/ |
theories for quantified intens.ional logics that I ha.ve described in that
LANGUAGE / > 9tUOtVtnUer,
n
altlrouglr no (closed) sentence of Principia AIo.Ihematica was taiieu to denotes
( s i n g u l a r t e r m )' , - , .
stand for a singular proposition, singular plopositious are the esserltial ['Ihrs relation is defined
building blocks of all propositions.] as the product of the other
The most impoltant hold-out against semantical theories that le- two telations)
quired singular propositions is Alonzo Church, the great modern cham-
pion of Frege's semantical theories. Church also advocates a version of
quantified intensional logic, but rqith a subtle difference thai fiuesses the
need for singular propositions. (In Church's logic, given a sentential for-
mula cont,aining free variables ancl given an assignment of values to the
variables, no proposition is yet determined. An addiiional assignnrent
..
of "senses" to the free variables r:rust be rnade before a proposition cau
be associated rvith the forrnula.) It is no accident that Churcll rejects i6!
tlirecl reference sernantical theories. For if there rvere singular. terrns
rvhich refelred directly, it seems liliely that Frege's problen.r: horv can
',t = -a =
fr- , if true, differ in meaning from a', could be reilstated,
-t
Direct Reference Picture not determine the extension of ihe terms. And fina.lly I(riplce's Prince-
ton lectures of spring 1970, later published as Naming and Necessily,a
were just beginning to leak out along with their strong attack on the
PROPOSITIONAL COMPONENT Fregean theory of proper names and their support of a theory of direct
reference.
As I said earlier, I was intrigued by the semantics of direct refer-
o\*'o' ence, so when I had a sabbaiical leave for ihe year 1970-71, I decided to
.o
worlc in the area in which such a theory seemed most plausible: demon-
LANGUAGE
(singular term)
""::Y refefs
INDIVIDUAL
stratives. In fall 1970, I wrote, for a conference at Stanford, a paper
"Dthat."s Using Donnellan's ideas as a starting point, I tried to de-
velop the contrast between Fregean semantics and the semantics of di-
rect reference, and to argue that demonstratives-although lhey could
be treated on a Fregean model-rvere more interestingly treated on a
(This relation is determined direct reference nodel. Ultimately I came to the conclusion that some-
by the conventions or rules thing analogous to Donnellan's referential use of a definite description
of the language) could be developed using my nerv demonstrative, "dthat." In the course
of this pa1>erI groped my way to a formal semantics for demonstratives
rather diffelent in conception from those that had been oflered before.
(These pictures are noi eniirely accurate for several reasous, among
In spring 1971, I gave a series of lectures at Princeton on the seman-
them, that the contrasting pictures are meant to account for more than
just singular terms and that the relation marlied 'refers' may already tics of direct refereuce. By ihis time I had seen a iranscript of Naming
and. Necessity and I tried to rela.te some of my ideas to I(ripke's.6 I also
involve a liind of Fregean sense used to fix the referent.)
had rvritten out the formal semantics for my Logic of Dernonstratives.
I won't go into the pros and cons of these trvo views at this iime.
That surnmer at the Irvine Philosophy of Language Inst,itute I lectured
Suffice it to say that I had been raised on Fregean semantics and was
again on the semantics of direct reference and repeated sorne of these
sufliciently devout to wonder whether the kind of quantification into
lectures at various institutions in fall 1971. And there the matter l.ras
modal and episternic conte-\ts that seemed to require singular proposi- stood except for a bit of updating of the 1971 Logic of Demonstratives
tions really made sense. (N{y paper "Quantifying In" can be regarcled notes in 1973.
as an attempt to explain auag such idioms for epi3temic contexts.)2 I now thinli that demonstratives can be treated corlectly only on a
But there were pressures from quarters other than quantified in- direct relereuce model, but that rny earlier lectures at Princeton and
tensional logic in favor of a semantics of direct reference. First of Irvine ou direct refereuce senrantics were too broad in scope, and that
all there was Donnellan's fascinating paper "Reference and Definite the most importaut aud certainly the most conviucing part of rny theory
Descliptions."3 Then there rvere discussions I had had $,ith Putnam is just the logic of demonstratives itself. It is based ol just a ferv quite
in 1968 in rvhich he argued rvith respect to cer.tain natural kirrd terms
like 'tiger' and 'gold', that if their Eregean senses rvere the kind of thing {Saul I(r'ipke, "Narning a[d Necessit).," ln Scrnant;ca oJ Notural Lansuase,
that one grasped rvhen one undelstood the terms, then such senses could ed. G. Harrnan and D. Davidson (Dordrechi: Reidel, 1972); revised edition pub-
Iislred as a set)arate monoslap\, Nornins and. Neeessits (Oxford: Basil Blackrvell,
19SO). Referertces ale to the levised edition.
2David I(aplan, sDavid I(aptan,
"Quantifying h," Ssnthese19 (1968): 178-214; replinred iu Lrre " D t h a t , " i n S y n t a r a 7 . d .S e z . a n l j c s , v o l . 9 , e d . P . C o I e ( N e w Y o r k :
Philosophy o! Lansrase, ed. A. P. Martinich (Oxford; Oxford Universiry pr.ess, 'ti
Acadert]ic Press, 1978); reprinied in Nfartillich, oI). cit.
-1 s s s ) _
rI(eith
6Althougl' rhe central ideas of my theoly had been rvorked out before
I becarne
Donnellan, "Reference and De6nite Descri1:tions," Phitosophicat Redeu 7s fannliar rvirh ffantns ond. Ne.cts;ty , I have enilusiaslically adopted the 'analytical
(1966): 281-304; replinted in Martinich, op. cit. apparaLus'and sorne of the terFrinology of that briuiant l'or.k.
.;iF
488 David I{aplan Demonstratives 489
simple ideas, but the conceptual apparatus turns out to be surprisingly I. Introduction
rich and interesting. At least I hope that you will find it so.
I believe my theory of demonstratives to be uncontrovertable and latgely
In this work I have concentrated on pedagogy. Philosophically, there
is Iittle here that goes beyond the Summer Institute Lectures, but I have uncontroversial. Tl.ris is noi a tribute to ihe power of my theory but a
iried, by limiiing the scope, to present the ideas in a more compelling concession of its obviousness. In the past, no one seems io have follou'ed
way. Some nerv material appears in the two speculative sections: XVII these obvious facts out to their obvious consequences. I do that. lVhat
(Epistemological Remarks) and XX (Adding 'Says'). It is my hope that is original with me is some terminology to help fix ideas when things get
complicated. It has been fascinating to see how interesting the obvious
a, theory of demonstratives will give us the tools to go on in a more
c o n s e q u e n c e so f o b v i o u s p r i n c i p l e s c a n b e . 7
sure-footed way to explore lhe de re propositional attitudes as well as
other semantical issues.
II. Dernonstratives, Indexicals, and Pure Indexicals
'a theory of demonstratives', but that
I tend to describe my theory as
is poor usage. It stems from the fact that I began my investigations
by asking what is said when a speaker points at someone and says,
word 'he', so used, is a demonstrative, and
"He is suspicious."s The
the accompanying pointing is the requisite associated demonstration. I
hypothesized a certain semantical theory for such demonstratives, and
'dthat'', and stipulated tltat its
then I invented a nerv demonstrative,
semantics be in accord rvith my theory. I rvas so delighted with this
'dthai', that lvhen
methodological sleight of hand for my demonstrative
'I', 'norv', 'here', etc.-
I generalized the theory to a.pply to rvords lilie
rvords rvhich do nol require an associated demonsiration-I continued
to call my theory a 'theory of demonstratives' and I referred to these
u'ords as'demoustratives'.
That terminological practice conflicts with what I preach, and I rvill
try to correct ii. (But I tend to backslide.)
The group of rvords for which I propose a semantical theory includes
'he', 'his', 'she', 'it', the demonstrative
the pronouns 'I', 'my', 'you',
pronouns'that','this', the adverbs'here','now','tomorrow','yesterday',
'present', and others. These tvords have uses
the adjectives 'actual',
other than those in which I am interested (or, perhaps, depending on
how you individuate words, we should say that they have homonyrns in
'he' and 'his' are
rvhich l am not interested). For example, the pronouns
used not as demonstraiives but as bound variables in
i sSee
"Dihat," p.320 in tr,larlinich.
,..,;
-t
For wl'rat is a man profiied, if he shall gain 'that vine'). But it is clear that one can
the whole world, and lose his own soul? distinguish a demonstrative
with a vacuous demonstration: no referent; from a demonstrative with
What is common to the words or usages in which I am interested no associated demonstration: incomplete.
is that the referent is dependent on the context of use and that the All this is by way of contrasting true demonstratives with pure index-
meaning of the word provides a rule which deiermines the referent in icals. For tlre latter, no associaled d,emonslralion is required,, and any
terms of certain aspects of the context. The term I now favor for these demonstralion supplied is either for emphasis or is irreleaanl.rl Amons
lvords is 'indexical'. Other authors have used other terms; Russell used the pure indexicals are'I', 'now', 'here' (in one sense), 'tomorrow', anJ
'egocentric particular' and
Reichenbach used 'token reflexive,. I prefer others. The linguistic rules lvhich govern t[ei?" use fully determine the
'inde.xical'(rvhich, referent for each context.12 No supplementary actions or intentrons are
I believe, is due to Pierce) because it seems less theory
laden than the others, and because I regard Russell's and Reichenbach,s necded. The speaker refers to himself rvhen he uses 'I', and no pointing
theories as defective. to another or believing that he is another or intendiug to refer to another
Some of the indexicals requite, in order to determine their referents, can defeat this reference.13
an associated demonstration: typically, though not invariably, a (visual) trIichael Bennett has noted that some indexicals have both a. pure
presentation of a local object discriminated by a pointing.e These in- and a demonstrative use. 'Ilere' is a pure indexical in
dexicals are the true demonstratives, and 'that,'is their paradigm. The
demonstralirre (an expression) refers to that which the demonslrolian I am in here
demonstrates. I call that which is demonstrated the,demonstratum'.
A demonstrative without an associated demonstration is incomplete. and is a demonstraiive in
The linguistic rules which govern the use of the true demonsrrarrves
'thai','he', In iwo rveeks, I will be here [pointing at a city on a map].
etc., are not sufficient to determine their referent in all con-
texts of use. Sorrrething else-an associated demonstration-must be
provided. The Iinguistic rules assume that such a dernoust,ratron ac- treatment we would assign as referent for 'he' whatever the speaker intend.ed, Lo
dernonstraie. I intended the exalnple to exemplify a failed demonstration, rhus.
companies each (demonstrative) use of a demonstrative. An incomplete a case in which the spealier, falsely believing the flower to be some man or other
demonstlative is not aacuous lilie a,n improper definite descriptiou. A in disguise, but having no particular rnan in mind, and certainly not intending to
demonstra"tive caz be vacuous in various cases. For example, rvhen its refer to anything other than tha.t man, says, pointing at the florver, ,'He has been
associated demonstration has no dernonstratum (a halluciuation)-or follo*ing me around all day."
trI have in nnnd such cases
t h e r v r o n g k i n d o f d e m o n s t r a t , u m( p o i n t i n g t o a f l o r v e r a n d s a y i n g , h e ' as pointing at oneself while sayins'I' (emphasis) or
pointing at someone else while saying 'I' (irr€l€vance or madness or whar?).
in the belief thai one is pointing to a lran disguised as a florverlo)-qt 12Tlr"r" a.e certain uses of pure indexicals that rniaht be caUed 'rnessages
Iecorded
too many demonstrata (pointing to tlvo intertivined vines ald saying Ior later broadcast', which exhibit a special uncertainty as to the referelrt of ,here'
-"H"r,-r and 'norv'. If the messase: "I arn not here norv" is recor.ded oD a telephone
.- d.-"""t.", y also b€ opportune and require no special action on referre<t uo by ,uc,w,-is rtre
the sp€aker's part, as when someone shouts ,'Stop that man" while only one man
.::] answerins device, ir is to be ElIin?I-fE-atTIiFTfri.
time of playback rather than the time ot recording. Donnellan has suggested that
is rushing toward the door. My rror.ion of a d€rnonsrrarion is a theor.e rical conceDt. if there were typically a signiRcant lag between our produclion of speech and its
I do not, in the pr.esent work, underrake a deraited ,operaiional'analysis of riris audition (for example, if sound tr.aveled very very slo\vly), our languase rnigtrt
notion alihough iher.e are scartered remarks relevanr to ihe issue. I do consir.ler,
contain two folrns of 'norvr: one for the time of produciion, another for the tirne of
in XVI below, sone alternativ€ th€oretical rreat.nents of denlonsrr.arions.
1oI arn awa.e ( 1) that in aome audiiion. The indexicals 'here' and 'norv' also suffer from vagueness regarding the
languaaes the so-cs ed rnasculine sender proloun may
be appropriate for flowers, but. it, is nor so in Enslish; (2) thata backgrormd sror.y size of the spalial and temporal neighborhoods to rvhich they refer. These facts do
can be provided that will make poinrirA at the Bower a contextually appr.opr.race, not seem to lne to slur the difierence between demonstratives and pure indexicals.
13Of co..t"" it is certain intentions on the part of the speaker that
though deviant, way of referrina to a man; for exanrple, if we are ralkiirl of great make a partic-
lrybridizers; and (3) that it is possible to trear rhe example as a, rcJerential use ular vocable the first person singular pronoun rather a nicknarne for lrving. I\4y
of the dernonsirative 'he' on the model of Donnetlan's referenrial use of a definite semantical theory is a theory of word meaning, not speaker's meaning. It is based
description (see "Reference and Definite Descriptions"). Under the r.efetenriat use on l;nguistic rules known, explicirly or implicitly, by all competent users of the
lanauage.
-t
We mean rather that given a use of the expression, we may ask of tohat rvould yield a constituent which is complex although it would determine
has been said whether il would have been true or false in various counter- the same object in all circumstances. Thus, (1), though a rigid desig-
factual circumstancesr and in such counterfactual circumstances, rvliich nator, is not directly refelential from this (metaphysical) point of view.
are the individuals relevant to determining truth-value. Thus we must Note, however, that every proposition which contains the complex ex-
distinguish possible occasions of use-which l call contexls-frorn possi- pressed by (1) is equiralenl to some singular proposition which contains
b l e c i l c u n - i s t a n c e so f e t a l u a l i o n o f l v h a l w a s s a i d o n a g i v e u o c c a s i o n o f just the number three itself as constituent.20
use. Possible circuurstances of evaluation I call circurnstances ot, soltle- The semantical feature ihai ,f wish to highlighi in calling an exples-
times, just counlerfaclual silualions. A directly referential terrrr nra! sion_dtt9eily-r2f3r&ehgl is not the fact tbaL it designates the same object
desigrrate dilTerent objects rvhen used in diflerent conlexls. But when in every circurnstance, but Lbe way rn which it designates an object in
evaluating what was said in a given context, only a single object rvill be -6*pression
any circ-umstance- Such an is d deaice of diiect reference.
relevant to the evalua[ion in all circumstances. This sharp distinction This does oes not-imply. that thai !t it hash.a9no"qg -conventionally fixed semantical rrrles
betrveen cozleels o/ use and circwnstances of eraluation must be kept
deiet:Frjge*r"lq refereDt in each context of use; quite the opposite.
in mind if rve are to avoid a seeming conflict between Principles 1 and There are r u i " " * 6 i c l , d e t e_ r_ m i n e- : t l r c r e f e r e r r t i n c a . h . o n -
' '.-.-- i
2.18 To look at the matter from another point of view, once rve lecognlze l e x c o f u s e" " ' b, rg. iti.l t. h
J ._^ r,. -, " i-* "
.pt-is a l l . T [ e r u l e s d o o[ provide a complet wltich
tl.reobviousness of both principles (I have not yet argued for Principle 2) logelher wilh a circumslance of eualualion yields an objecl. They just
the distinction between contexts of use and circumstances of evaluation
p r o o i d , ea n o b je c t .
is forced upon us.
If rve keep in mind our sharp distinction between contexts of use and
If I may rvax metaphysical in older to fi,r an image, let us ihinli of the
circumstances of evaluation, we will not be tempted to confuse a rule
vehir:les of evaluation-the what-is-said in a, given context-as proposi-
which assigns an object lo each conlert with a 'complex' which assigns
t i o n s . D o n ' t t h i n h o f p r o p o s i t i o n s a s s e t s o f p o s s i b l er v o r l d s , b u t r a t l r e r
an object to each circumslance. Fot example, each context lLas an agenl
as structured entities looking something lilie the sentences rvhich express
(loosely, a speaker). Thus an appropriate designation rule fot a directly
thern. For each occurrence of a singular tenn in a sentence there lvill be a referential term would be:
c o r r e s p o n d i n gc o f l s t i t u e n t i n t h e p r o p o s i t i o n e x p r e s s e d .T h e c o n s t i t u e n t
of the ploposition determines, for each circumstance of evaluation, the (2) L.r each possible context of use the given term refers to the
object relevant to evalu.rting the ptoposition in that circumstance. In agent of the context.
general, the constituent of the proposition will be some sort of complex,
But this rule could not be used to assign a relevant object to each cir-
constructed from various attributes by logical composition. But iu the
c u m s t a n c e o f e v a l u a t i o n . C i r c u m s t a n c e so f e v a l u a t i o n d o n o t , i n g e n e r a l ,
case of a singular term which is directly referential, the constituent of
have agents. Suppose I say,
the ploposition is just the object itself. Thus it is that it does notjust
lurn oul that ihe constitueut determines the same object in every cir- f3) I do not exist.
cumstance, the constituent (corresponding to a rigid designator) just is
the object. Tltere is no delermining to do at all. On this picture-and Under n'hat circumstances lvould rrl/ro,I said be true? It rvould be true
this is real/y a picture and not a theory-the in circumstances in rvhich I did not exist. Among such circumstances
delinite description
are those in lvhich no one, and thus, no spealiers, no agents e-\ist. To
(1) The nf(Snorvis slighi A n'?= 9) v (-Snorv is slight n search a circumstance of evaluation for a spealier in order to (mis)apply
22=n + 1)l1e rule (2) would be to go off on an irrelevant chase.
people (possibly including me) nowadays seem to have viels rvhiclt allow that.
13I think it likely thar 'snorvis rvl,ite'may be necessaly.
it was just the failure to noiice rhis disrinciion rhar le.-l co
a failure to recognize Principle 2. Some of the hisLory and consequences of Lhe 20I am ignoring plopositions expressed by sentences containina episiemic operators
corltlation of Cor1text and CircumsNance is discussed in seciion VIt. or others for rvhich equivatence is not a sumcieni condition for interchange of
1eI would
have used'snow is white', but I wanted a coniinsent ctause, and so rlany
496 David Kaplan
Demonstratives 497
Three paragraphs ago I sketched a metaphysical picture of the struc- will be regarded as rigid designators in the mod,ified,sense of an expres-
ture of a proposition. The picture is tahen from the semantical parts sion which designates the same thing in all possible worlds (irrespective
of Russell's Principles of Malhernalics.2l Two years later, in "On De- of whetller the thing exists in the possible world or not).24 llowever, as
noting,"22 even Russell rejected that picture. But I siill like it. It is already noted, I do not regard all rigid designators-not even all strongly
not a part of my theory, but it well conveys my conception of a directly rigid designators (those that designate something that exists in all pos-
referential expression and of the semantics of direct reference. (The pic- sible worlds) or all rigid designators in the modified sense-as directly
ture needs some modification in order to avoid difficulties which Russell referential. I believe that proper names, like va,riables, are directly ref-
later noted-though he attributed them to Frege's theory rather than erential. They a,re not, in general, strongly rigid designators nor are
his own earlier theory.)23 they rigid designators in the original sense.25 \/hat is cha,racteristic of
If we adopt a possible worlds semantics, all directly refereuiial terms directly referential terms is that the designatum (referent) determines
2lBeltrand Rnssell, The Principles oJ Mathenotics the propositional component rather than the propositional component,
(London: Allen & Un|in, 19O3).
22Beltrand Russell, along with a circumstance, determining the designatum. It is for this
"On Denoting," A4ind 14 (1905): 475-93.
23Here is a dificulty in RusseU's 1903 picturc that has some histor.ical i[terest. reason that a directly referential term that designates a contingently ex-
Consider the proposition expressed by the senience, 'The centre of mass of the isting object will still be a rigid designator in the modified sense. The
Solar System is a point'. Call the proposition, 'P'. P ha-: in iis subject place a propositional component need not choose its designatum from those of-
certain complex, expressed by the de6nii,e description. Call the complex, 'PIexy'.
We can describe Plexy as "the cornplex expressed by 'the center of rnass of the
fered by a passing circumstance; it has already secured its designatum
solar system'." Can we produce a directly relerential term which desianates Plexy? before the encounter with the circumsiance.
Leaving aside for the noment the controversial quesiion of whether 'PIexy'is such When we think in terms of possible world semantics this fundamental
a term, let us irnagine, as Russell believed, that we can directly refer to Plexy
distinction becomes subliminal. This is because the style of the seman-
by afnxing a kl,ttd ot meaning marhs (on the analogy oI quotation marks) to the
description itself. Now consider the sentence 'athe center of mass of the solar
tical rules obscures the distinction and makes it appear thai directly
systernm is a point'. Because the subject of this sentence b direcily referential referential terrns differ from ordinary definite descriptions only in that
and refers to Plexy, the proposition the s€ntence expresses will have as its subj€ct the propositional component in the former case must be a conslazl func-
constituent Plexy itself. A moment's reflection will reveal that this proposition is
simply P asain. But this is absurd since the two sentences sp€ak about radicaUy
tion of circumstances. In actual fact, the referent, in a circumstance, of
different objects. a directly refereniial term is simply independenT of the circumstance and
(I believe the foresoinS argument ties behind some of the larsely incornprehensi- is no more a function (constant or otherwise) of circumstance, than my
ble arAurn€nts nounted by Russell aaainst Freae in ',On Denoring,,, though tlaere
are certainly other difficulti€s in that argurnent.
action is a function of your desires when I decide to do it whether you
It is not surpr.ising rhar RusseU
tlrere confused Flege's iheory with his own ol Pr;tuc;ple oJ Alathcmotics. Tine like it or not. The disiinciion that is obscured by the style of possible
nrst footrote of "On Denoting" asserts that the two theories ale ,,very near.ly the world semantics is dramaiized by the structured propositions picture.
same.,') That is part of the reason rvhy I lilie it.
The solutio4 to the dimculty is simple. Regard the ,object'places of a singular
proposition
Some directly referential terms, like proper names, ma,y have no se-
as marked by some operaiion rvhich cannot mar-li a cornplex. (There
always rvill be some such operation.) For example, suppose that no complex is mantically relevant descriptive meaning, or at least none that is specific:
(repr€sented by) a set containing a single mernber. Then rve need only add {. . .} to that distinguislies one such ierm frotn another. Otliers, lihe the index-
mark the places in a singular proposition *hich conespond ro dirccity refer.enrial icals, may have a limited kind of specific descriptive meaning relevant
terrns. l /e no longer need worry about confusing a conplex with a propositional 'dthat' terms (see
constituent co.responding
to the features of a context of use. Still others, like
to a dir.ectly refer na t€rrn bec.ause no complex wiU
have the form {x}. In particular, Plexy I {Plexy}. This technique can also be below), nay be associated with full-blown Fregean sensesused to fix the
used io resolve another confusion in RusseU. He araued thar a sen!ence containing referent. But in any case, the descriptive meaning of a directly referen-
a nondenoiing directly referential term (he would have called it a noqdenoting
'logically proper tial term is no part of the propositional content.
name') woutd be meaningless, presumably because rhe purporred
sinaular proposition would be incomplete. But the braces themsetves can fill ou|
the singular pr.oposition, and if they contain nothina, no more anomalies need ?4ThiB is the
Jq.si rc's€ of footnoie 16.
result than rvhat the development of Free Logic has already inur.ed us ro. 2sTlris is ihe second sense of footnote 16.
498 David l{aplan
Demonstratives 499
It is impossible that I do not exist. I The actual circumstances. here. and now ate such that it is
possible that in Pahistan in five years only those who, in the
Ilere is another example to show that the descriptive meaniug of an il first, are located at the second, during the third, are envied.
indexical rnay be entirely inapplicable in the circumstance of evaluation.
lVhen I say, Although this may not be the most felicitous form of expression, its
meaning and, in particular, its symboliza,tion should be clear to those
I lvish I lvere not speaking now.
familiar with quantified intensional logics. The pronouns, 'the first', 'the
TIre circumstances desired do not involve contexts of use and agenls second', and'the third' are to be represented by distinct variables bound
wlto are not speaking. The aclual context of use is used to determine to existential quantifiers at the beginning and identified lvith 'the actual
the relevant individual: me-and time: nora-and then we query the circun-rstan ce', 'here', and 'norv'respectively,
various circumstances of evaluation with respect to l/r.al individual and
(5) (lru) (:p) (:t) [a,=the actual circumstance A p=here At=nou
lhat tr'rle.
A O In Pakistan In five years Vc(e is envied * c is located
Here is another example, not of the inapplicability of the descriptive ^+ -. 1..-:-- +:-
dL y uulttt6 L t t t "u. ), l\ l
meaning to circumstances but ofits iuelevance. Suppose I say at le, "It
rvill soon be the case ihat all that is now beautiful is faded." Consider But such transformations, when thought of as representingthe claim
wha,t \aas said in the subsentence, I
that indexicals take primary scope, do not provide an allernaliue to
Principle 2, since we may still ask of an utterance of (5) in a context c,
AII tha.t is norv beautiful is faded. I when evaluatingit with respecito an arbitrary circumstance,to rvhat do
I
I wish to evaluate tllat content at some near future time 11. \,\rhat is the the indexicals'actual', 'here', and 'norv'refer. The ansrver,as ahvays,is:
relevant time associated rvith the indexical'low'? Is it the future time .a
the relevant featuresof the context c. [In fact, although (4) is equivalent
t1? No, it is lo, of course: the time of the conte.xt of use. to (5), neither indexicalsnor quantification acrossintensionaloperators
t is dispensablein favor of the other.]
,{t
.-l
I
I
il
content by a function from circumstances of evaluation to an appropriate the time of the circumstance. If we built the time of evaluation into
extension the contents (thus removing time from the circumstances leaving only,
rtle represen[a[ron Is a n] speak of con- say, a possible world history, and making contents specif,c as to iime),
tents in terms of it, but one should note that contents which are distinct it would make no sense to have temporal operators. To put the point
but equivalent (i.e., share a value in all circumstances) are represented another way, \f uhal is said is thought of as incorporating reference to
by the same intension. Among other things, this results in the loss of a specific time, or state of the world, or whatever, it is otiose to ask
my distinction between terms which are devices of direct reference and whether what is said would have been true at another time, in another
descriptions which lurn oul to be rigid designators. (Recall the meta- state of the world, or whatever. Temporal operators applied to eternal
physical paragraph of section IV.) I wanted the content of an indexical sentences (those whose contents incorporate a specific time of evalua-
to be just the referent itself, but the intension of such a content will be tion) are redundant. Any intensional operators applied to perfecl sen-
a constant function. Use of representing intensions does not mean I am tences (those whose contents incorporate specific values for all features
abandoning that idea-just ignoring it temporarily. of circumstances) are redundant.2s
A fited conlenl is one represented by a constant function. All di-
rectly referential expressions (as well as all rigid designators) have a 23The notion oI redundancy involved could be made precise. When I speak of
fixed content. [What I elsewhere call a stable conLent.f .f
building the time of evaluation into contents, or making conients sp€cific as to
Let us settle on circumslances for possible circumstances of evalu- time, or taking what is said to incorporate reference to a specific tirne, what I
'I am writing', in the present context
ation. By this I mean both actual and counterfactual situations rvith t have in rnind is this. Given a sent€nce S:
c, which of the followins should we take as the content: (i) the proposition that
respect to which ii is appropriate to ask for the extensions of a given 'proposition' that David
Dawid Kaplan is writins at 10 A.M. on 3/26/77 ' or (ii) the
well-formed expression. A circumstance will usually include a possible Kaplan is writing? The proposition (i) is specific as to time, the'Proposiiion'(ii)
state or history of the world, a time, and perhaps other features as l'ell. [the scare quotes refl€ct rny feeling that this is not the traditional notion of a
The amount of information we require from a circumstance is linked to propositior] is neuLral with respect to tim€. If we ta.ke the content of .s in c to be
the degree ofspecificity of contents, and thus to the kinds of operators
I (ii), we can ask wheiher it would be true at times other than the time of c. Thus
'proposition' a-s changing its truth-value over
we think of the ternporally neutral
in the language. t tirne. Note tha.t it is not just the non€ternal sentence .S that changes its truth-
Operators of the familiar kind treated in intensional logic (modal, value over tirne, but the 'p.oposition' itseu. Since the sentence S coniains an
temporal, etc.) operate on conients. (Since we represent contents by indexical'1', it will express difierent rpropositions' in difierent contexis. But since
t S contains ^o reiLporul indexical, the time of the contexi will not influence the
intensions, it is not surprising that inteusional operators operate on 'proposition'expressed. An alternatiwe fana more trad;tional] view i6 to say that
contents,) Thus an appropriate extension for an intensional operator the verb tense in S involves an implicit ternporal indexical, so ihat S i5 un<lerstood
I 'I am wrii,inA now'. If we take this point of view we will
is a function from intensions to extensions.2T A modal operator when as synonymous with '9':
take the content of 5 in c to be (i). In this case uhat i3 raid is eternal; it does not
applied to an intension will look at the behavior of the intension rvith over iime, although S will express ditrerent propositioru! at
I change its truth-value
respect to the possible staie of the world feature of the circumstances difierent times.
of evaluation. A temporal operator will, similarly, be concerned with There are both technical and philosophical issues involved in choosins between
a (i) and (ii). Philosophically, we may ask why the ternporal indexical should be
.{ taken to b€ implicit (lnakins the proposition eternal) when no modsl indexical
27 As we shall see, indexical operators such
as "Ii is now the case that," "It is act ually is taken to be iInplicit. After all, we cou,d understand S as synonymous wilh
I
the case that," and "dthat" (the last takes a term rather than a sentence as S": 'I arn actually wriling now'. The cont€nt of S" in c is not only etelnal' it
argurn€nt) are also intensional operators. They difier from the familiar operators is perfect. Ii3 truth changes neither through time nor Possibility. Is there some
in only trvo rvays: first, their extension (t he funclion from intensions to extensions ) aood philosophical reason for pleferling contents which are neutral wiih resp€cl
depends on context,, a4d second, rhey are directly referential (thus they have a fixed to possibility but draw nxed values from the context for all other leaiures of a
conient). I shall arsue below (in section VII: Monsters) that all operatoE that can possible circumstance whether or noi the sentence contains an explicit indexical?
be given an English readilg are 'at rnost' iniensional. Note that when discussing (It rnay b€ that the traditional view was abetied by one of the delishtful anornalies
issues in terms of the forrnal representations of the model-theoretic semantics, I of the loaic of indexicals, namely that S, S', and S" are all logically equivalent!
tend to speak in terlrE of intensions and intensional operators rather than contenrs See Rernark 3, p. 547.) Technically, ve must note thai intensional operators must,
and conleni operaior.s. if they ar'e not to be vacuous, opelaLe on conLents which are neutral with respect
l
!
504 David Itaplan
I
I
Demonstratives 505
What sorts of intensional operators to admit seems to me largely uncomfortable with some of its seeming consequences(for example, that
a matter of language engineering. It is a question of which features of
the world might be in a state qualitatively exactly as it is, but with a
what rve intuitively think of as possible circumstances can be sufficiently permutation of individuals).
well defined and isolated. If rve wish to isolate location and regard it as a
It is hard to see how one could think about the semantics of indexicals
feature of possible circumstances vt'ecan introduce locational operators:
'Trvo niles north it is the case that', etc.
Such operators can be iterated
and can be mixed with modal and temporal operators. Hor.vever,to malie
I and modality rvithout adopting such a view.
is correct, rouahly, that words and phrases have a kind of d€sc.iptive rneanina or.
sense which at one and the same iiine constitutes their cognitive signi6cance and
VII. Earlier Attempts: Index Theory
iheir condiiion3 of applicability.
I<ripke says repeaiedly
The following picture seems to emerge. The meaning (character) of a,n
in Naning and, N.ce$it! that he is only providing a
picture of how proper narnes refer and that he does not have an exact theory. indexical is a function frorn coniexts to extensions (subst,ituting for fixed
His piclure yields some startling resulte. In the case of indexicals we do have a contenis). The meaning (content, substituting for fixed characters) ofa
rathet precise theory, which avoids the ditficulty of specifying a chain of corununi- nonindexical is a function from circumstances to extensions. From this
cation and which yields many anatogous resulis. Iu tacing the vastly more dif6cult
problerns associaied with a theory of reference for proper nanres, the theory of
point of view it may appear that the a.ddiiion of indexicals requires no
indexicals rnay prove uselul; if only to shorv-as I believe-that prop€r names are ne'u logic, no sharp distinction between contexts and circumstances, j ust
not indexicals and have no meanins in the sense in which indexicals hav€ m€an- tlre addition of some special new Jealures ('contextual' features) to the
ins (namely a 'coanitive content'which fixes the relerences in all contexts). [TLe
issues that arise, involvin6 token reflexives, homonymous words wit.h distinci char-
circumstances of evaluation. (For e,xample, an agenl to provide an in-
acter, and hornonyrnous token reflexives lviih the sam€ character ar.e best saved terpretation for'I'.) Thus an enlarged view ofintension is derived, The
for later-much laier.l intension of an expression is a function fiom certain factors to the ex-
32see gza of Rudolf C;.',.p'"
,l./"o"t"e and, Necessiry (Chicaso: University of tension of the expression (with respect to those factors). Originally such
Chicaso Press, 19a7).
factors rvere simply possible states of the world, but as it was noticed
508 David Kaplan Dernonstratives 509
that the so-called tense operators exhibited a structure highly analo- and f is a time-(6) will be false. In fact, (6) is true only with respect
gous to that of the modal operators the factors with respect to which to those indices (ro,x,p,t) which are such that in the world history tu,
an extension was to be determined were enlarged to include moments r is located a,t p at ihe time 1. Thus (6) fares about on a par with
of time. When it was noticed ihat contextual factors were required to
(7) David Itaplan is in Portland on 26 N4arch 1977.
determine the extension of sentences containing indexicals, a still rnore
general notion was developed and called an "index." The extension ofan (?) is empirical, and so is (6).
expression was to be determined with xespect to an index, The intension But here we have missed something essential to our understanding
of an expression was that function which assigned to every index, the of indexicals. Intuitively, (6) is deeply, and in some sense, which we
extension at that index. will shortly make precise, universally, true. One need only understand
the meaning of (6) to know that it cannot be uttered fa,lsely. No such
The above example supplies us with a statement whose truth-
guarantees apply to (7). A Logic of Ind.eticals which does not reflect this
value is not constant but varies as a function of i e J,
intuitive difference between (6) and (7) has bypassed something essential
Tl.ris situation is easily appreciated in the context of time-
to the logic of indexicals.
dependent statements; that is, in the case where l represents
What has gone wrong? We have ignored the specia.l relationship
the instant of tirne. Obviously the same statement can be
between 'I', 'here', and 'norv'. Here is a proposed correction. Let the
true at one moment and false at another. For more general
class of indices be narrowed to include only the proper ones-namely,
situations one must not think of the i € f as anything as
those (tu, c,p, t) such that in the world rit, a is located at p at the time t.
simple as instants of time or even possible worlds. In general
Such a move may have been intended originally since improper indices
we will have
are lilie impossible lvorlds; no such contexts coultl exist and thus there
i: (ro,t,p,a,...) is no interest in evaluating the extensions of expressions with respect to
them. Our reform has the cotrsequencethat (6) comes out, correctly, to
where the index i has rrrany coordinales; for example, u., is a be logically true. Now consider
uorld, t is a lime, p = (c , y, z) is a (3-dimension aL) posilion (8) E I am here now.
in tlre world, a is an agenl, etc. All these coordinates can
be varied, possibly independently, and thus affect the truth- Since the contained sentence (namely (6)) is true at every proper index,
values of statements which have indirect references to these (8) also is true at every proper index and thus also is logically ttue. (As
coordinates. [From the Aduice of a prominent logician.] would be expected by the aforementioned principle of modal generaliza-
iion. )
A sentence { was taken io be logically true if true at every index But (8) slrould not be logically true, since it is false. It is certainly
(in every 'structure'), and nd was taken to be true at a given index nol necessary that I be here now. But for several contingencies, I would
(in a given structure) just in case / was true at every index (in that be rvorliing in my garden now, or even delivering this paper in a location
structure). Thus the farniliar principle of rnodal geueralization: if ! /, outside of Portland.
then f E@, is validated. The diffi.culty, here, is the attempt to assimilate the role of a conled
This view, in its treatment of indexicals, was technically wrong and, t o t l r a t o f a c i r c u m s l a n c e . T h e i n d i c e s ( w , c , p , t ) t h a t r e p r e s e n tc o n t e x t s
more importantly, conceptually rnisguided. must be proper in order that (6) be a truih of ihe logic of indexicals, but
Consider the sentence the indices that represent circumstances must iuclude improper ones in
order ihat (8) nol be a logical truth.
(6) I am here now.
If one wishes to stay with this sort of index theory and blur tl.re
It is obvious that for many choices of index-i.e., for rnany quadruples conceptual difference between context and circurnstance, the minimal
\tu,x,p,t) rvhere u-ris a possible world history, * is a person, p is a place, requiLement is a system of double indexing, one index for context and
510 David I(aplan Demonstratives 511
another for circumstance. It is surprising, looking back, that rve (for I than use it, rve can, of course, operate directly on it. Carnap once
was among the early index theorists) did not immediately see that double pointed out io me how important the difference betrveen direct and
indexing was required, for in 1967, at UCLA, Hans Kamp had reported indirect quotation is in
his rvork on'now'33 in rvhich lte had shorvn that double indexing rvas
required to properly accommodate temporal indexicals along with the Otio said "I am a fool."
usual temporal operators. But it was four yeans before it was realized Otto said that I am a fool.
that this was a general requirenrent for (and, tn a sense, the key to) a
Operators like'In some contexts it is true that', rvhich attempt to meddle
logic of indexicals.
witlr character, I ca,ll monsters. I claim that none can be expressed in
However, mere double indexing, without a clear conceptual under-
English (without snealcing in a quotation device). If they stay in the
standing of what each index stands for, is still not enough to avoid all
metalanguage and confine their attention to sentences as in
pitfalls.
In some contexts "l am not tired norv" is true
VIII. Monsters Begat by Elegance they are rendered harmless and can even do socially useful work (as
d o e s , ' i s v a l i d ' [ s e eb e l o w ] ) .
N,Iy liberality with respect to operators on content, i.e., intensional op-
I have gone on at perhaps excessive lengih about monsters because
erators (any feature of ihe circumstances of evaluation that can be well
they have recently been begat by elegance. In a specific application
defined and isolated) does not extend to operators which attempt to
'In some contexts it is of the theory of indexicals there will be just certain salient features of
operate on character. Are there such operators as
true that', rvhich rvhen prefixed to a sentence yields a truth if and only a circumstance of evaluation. So we may represent circumstances by
if in some context the contained senlence (not the content expressed by indexed sets of features. This is typical of the model-theoretic way. As
already indicated, all the features of a circumstance rvill generally be
it) expresses a content that is true in the circumstances of that conte-\i?
required as aspects of a context, and the aspects of a context may all be
Lei us try it:
features of a circumstauce. If not, a little ingenuity may make it so.35
(0) In sorne contexts it is true that I am not tired norv. 3sRecall that in a palticular lormal theory the fea.tures of a circurnstance must
iuclucle all elements rvith respecl to which there are conlent oPerators, and lhe
For (9) to be true in the present conte-xt it suffices that some agent of aspects of a context must include aU elements .lviih respect to which lhere ate
'O', 'tr" and
some cootext not be tired at the timeof that context. (9), so intelpreted, indexicals. Thus, a language r*ith both th€ usual modal oper.atols
an indexical modal operator'It is actually the case tlrat' will contain a possible
has nothing to do rvith me or the present moment. But this violates
world history feature in its circumstances as well as an analogous asPect in its
Principle 2! Principle 2 can also be expressed in more theory laden rvay contexts. If a ci.currrstance is an aspect of a context, as seerns necessary for the
by saying tlrat indexicals always take primary scope. If this is true-and denniiion of trulh, then we onty need worrv sbout asPects of contexts that are
it is-then no operator can control the character of the indexicals within not, features of circumstances. The most prorninent of these is the 49.n, of the
context, requi!'ed io interpret the indexical'I' In order to supPlv a corresponding
its scope, because they will simply leap out of its scope to the front of we must tteat contents in such a way that we
nonvacuous f€atur.e to circurrstances
the operator. I an-r not saying rve could not construct a lauguage with can ask whether they are true for valious agents. (Nor charactcrs lnind vou, but
such operators, just that English is not one.34 And such operators cozld contenfs.) TIris can be done by rePreseniing the ager.t bv 6 tueurral-a ielm which
n o l b e .a d d e d l o i l . plays the syntacticat !.ole of 'I' but Sets an interpretation onlv w;th respecl to a
circurnstance. Let a be a special variable that is not subjeci to quanlification and
There is a lvay to control an indexical, to keep it from taliing 1>titrrar:y let 6 be a variable not in the language. Our variable a is ihe neutral. We wish io
scope, and even to refer it to another context (this amounts to changiug iniroduce conienl operators which affect the agent place and which can be i!erated.
'a.R6' for 'b is an uncle of o'.
its clraracter). Use quotation rnarks. If we nlenlion the inde.rical rather Let -R be a relation between individuals, for example
Then we rnay intcrpret the operator OR{ as (36)[aRD n (]a) (6 = a n C)1. If C is
33Publislred in 197r as 'Norv'," Theoria, 'o *alks', ORd cornes to 'an uncle of a wa!ks'. The indexical'I' can be represe-nted
"Fomal Properties of
3aThotr,."ot, alleges a counrelinstance: 'Never put off uniil tonlorr'o\v what )ou can by an operator OI for which
'o-R6' is just 'I=b' The result should be that Ord is
do today'. What should one say about ihis? equivalent io repl&cing the neutral a by the indexical'I'.
512 David Kaplan Demonstratives 513
We could then represent contexts by the same indexed sets we use to text provided that Paul-in whatever costume he appeared-were still
represent circumstances, and instead of ha,ving a logic of conlecls and residing in Princeton.
circumslances we have simply a 7uo-d,irnensional logic of ind.emd sels.
This is algebraically very neat and ii permits a very simple and elegant IX. (i) The Argurnents
description of certain important classes of characters (for example, those
which are true at every pair (i,t), though the special significance of I am arguing that in order to determine what the truth-value of a propo-
the set is somehow diminished in the abstract formulation).36 But it sition expressed by a sentence containing a demonstrative uould, be nnd,et
also permits a simple and elegant introduction of many operators which other possible circumstances, the relevant individual is not the individual
are monsters. In abstracting frorn the distinct conceptual roles played lbat roould laoe been demonstrated had those circumstances obtained
by contexts of use aud circumstances of evaluation the special logic of and the demonstration been set in a context of those circumstances, but
indexicals has been obscured. Of course restrictions can be put on the rather tlre individual demonstrated in the context wh\ch did, generate
two-dimensional logic to exorcise the monsters, but to do so would be the proposition being evaluated. As I have already noted, it is char-
to give up the mathematical advantages of that formulation.3T acteristic of sentences containing demonstratives-or, for that matter,
any indexical-that they may express different propositions in different
contexts. We rnust be wary ofconfusing the proposition that would have
IX, Argument for Principle 2: True Demonstratives been expressed by a similar uiterance in a slightly different context-
I return now to ihe argument that all indexicals are directly referential. say, one in rvhich the demonstratum is changed-rvith the proposition
Suppose I point at Paul and say, that was actually expressed. If we keep this distinction in mind-i.e., we
distinguish Pat and Mike-we are less likely to confuse what the truth-
He now lives in Princeton, New Jersey.
value of the proposition ccluolly expressed would have been under some
Call whal I said.-i..e., the content of my utterance, the proposition possible circumstances with what the truth-value of the proposition that
expressed-'Pat'. Is Pat true or false? True! Suppose that unbeknownst would haue 6een expressed would have been under those circumstances.
to me, Paul had rnoved to Santa Monica last week. Would Pat have When we consider the vast array of possible circumstances with re-
then been true or false? False! Now, the tricky case: Suppose that spect to which we might inquire into the truth of a proposition expressed
Paul and Charles had each disguised themselves as the other and had in some context c by an utterance r, it quickly becomes apparent that
switched places. If that had happened, and I had uttered as I did, then only a small fraction of these circumstances will involve an utterance of
the proposition I utould haoe expressed would have been false. But in the same sentence in a similar context, and thai there must be a way of
tha.t possible context the proposition I uould haue expressed is not Pat. evaluating the truth-value of propositions expressed using demonstra-
That is easy to see because the proposition I would hate expressed, had tives in counterfactual circumstances in which uo demonstrations are
I pointed io Charles instead of Paul-call this proposition 'Mike'-not taking place and no individual has the exa,ct cha,ra,cteristicsexpLoited in
on|y would haae been false but actually is false. Pat, I would claim, the dernonstration. Surely, it is irrelevant to determining whether rvl.ratI
would still be true in the circumstances of the envisaged possible con- said rvould be true or not in some counterfa.ctual circumstance, whether
36See, for exarnple, Krister Sea€rbera, Modal Losic," Jotrtual oJ
Paul, or anyone for that matter, looked,as he does now. AII that rvould
"Two-dirnensional
Philotoph;cal Losic 2 (r973\t 77-96. Seaerberg does metamathematical work in be relevant is where he liues. Therefore,
his article and makes no special philoaophical clairns about its sianificance. That
has been done by others. (T3) the relevant features of the demonstratum q?ra den,onslra-
3TThere is one other difficulty in identifyinS the class of contexis with the class of lzm (compare, the relevant features of the x Fx qua llte t
circumstances. The special relationship between the indexicals ,I', 'here', ,now'
seerns to require that th€ agent of a context be at the location of ihe coniext Fc)-narnely, that the speaker is pointing at it, that it has
durin8 the tirne of the context. But this restriction is not plausible for arbitrary a certain appearance! is presented in a certain way-cannot
circurnstances. It appears that this approach will have difficulty in avoiding rhe be the essential characteristics used to identify the relevant
problems of (o) and (s) (section vII).
individual in counterfactual situations.
-,1l
,i.
514 David Kaplan
Demonstratives 515
These two arguments: the distinction between Pat and \,Iike, and con-
the same demonstratum but distinct manners of presentatioo. It is this
sideration of counterfactual situations in which no demonstration occurs,
diflerence between the sense of Hes and the sense of Phos that accounts,
are offered to support the vierv that demonstratives are devices of direct
the Fregean claims, for the informativeness of the assertion.
reference (rigid designators, if you will) and, by contrast, to reject a
It is possible, to pursue the analogy, lor a demonstration to have no
Fregean theory of demonstratives.
demonstratum. This can arise in several ways: through hallucination,
through carelessness(not noticing, in the darltened room, that the sub-
IX. (ii) The F\'egean Theory of Demonstrations ject had jumped off the demonstration plat,form a few moments before
In order io develop the latter theory, in contrast to my own, we turn first the lecture began), through a sortal conflict (using the demonstrative
to a portion of the Fregean theory which I accept: the Fregean theory phrase tthat F , where -F is a common noun phrase, while demonstrat-
of demonstrations. ing sornething which is not an F), and in other ways.
As you know, for a Fregean the paradigm of a meaningful expres- Even Donnellans's important distinction betlveen referential and at-
sion is the definite description, which picks out or denotes an individual, tributive uses of definite descriptions seems to fit, equally comfortably,
a unique individual, satisfying a condition s. The individual is called the case of demonstrations.38
lbe denolaliott of the definite desciiption and the condition s we may The Fregean hypostatizes demonstrations in such a way that it is
ideniify wibh the sense of the definite description. Since a given individ- appropriate to ask of a given demonstration, say Phos, what would it
ual may uniquely satisfy several distinct conditions, definite descriptions have demonstrated under various counterfactual circumstances. Phos
with distinct senses may have the same denotation. And since some con- and Hes might have demonstrated distinct individuals.3s
ditious may be uniquely satisfied by no individual, a definite description \4/e should not allow our enthusiasm for analogy to overwhelm judg-
may have a sense but no denotation. The condition by means of which ment in this case. There are some relevant respects in which descrip-
a defirrite description picks out iis denotation i,s llte rtan.ner of presen- tions and demonstrations are disanalogous. First, as Da,vid Lervis has
lation of the denotation by the definite description. pointed out, demonstrations do not lrave a syntax, a fixed formal struc-
The Fregean theory of demonstratives claims, correcily I believe, iure in terns of rvhose elements we might try to define, either directly
that the analogy between descriptions (short for'definite descriptions') or recursively, the notion ofsense.40 Second, to diflerent audiences (for
and demonstrations is close enough to provide a sense and denotalion exarnple, the spealier, those sitting in front of the demonstration plat-
analysis of the 'meaning' of a demonstration. The denotation is tire form, and those sitting behind the demonstration platforrn) the same
den.ronstratum (that which is dernonstrated), and it seems quite nat- demonstration may have difierent senses. Or perhaps we should say
ural to regard each demonstration as presenting its demonstratum in that a single performance may involve distinct demonstrations from the
a particular manner, which we may regard as the sense of the dernon- perspective of distinct audiences. ("Exactly like proper names!" says the
stration. The same individual could be demonstrated by demonstra- Fregean, "as long as the demonstra,ium remains the same, these fluctu-
tions so different in manner of presentation that it would be informative ations in sense are tolerable. But they should be avoided in the system
to a competent auditor-observer to be told that the demonsrrata were 38I have written elsewhere, in appendices vlI and VIII of
"Bob and Carol and Ted
one. For e,rample, it ntight be infor.mative to you for me to tell you and Alice," of these malters and won't pursue the topic now.
Itrat 3eIt could then be proposed that demonstraiions be individuated by the principle:
d1 = d2 if and only iI, for all appropriate circu:nrstances c, the demonstratum of
That [poiut,ingto Venus in the morning sliy] is identical rvith d1 in c = the demonstratum of d2 in c. An alternative principle of individuation
is that the sarne dernonstration is being perfo|med in trvo dilTerent contexts if the
that [pointing io Venus in tlie evening sky]. standard audience can't determine. hom the demonstration alone. whether the
confexts al.e distinct or identical. This makes the individuation ol demonsrrations
(I would, of course, have to speak very slorvly.) The trvo demonstra- more epistemolotical than the metaphysical proposal above.
tions-call the first one 'Phos'and the secondone'Hes'-rvhich accon.t- {oAlthouah recent wor.k on computer perception has aLtempted to idcnlify a syntax
panied the two occurrencesof the demonstrativeexpression,that, have of pictu.es. See P. Suppes and Rottmayer, "Automata," in Ho,nd6ook ol Percep-
t;oz. vol. 1 (Ne$/ York: Acadernic Press. 1974).
516 David I(aplan Demonstratives 517
of a demonstrative science and should not appear in a perfect vehicle of ciated demonstration. Both theories allow that, the demonstration can
communication." ) be regarded as having both a sense and a dernonstratum. My theory, the
direct reference theory, claims that in assessingthe proposition in coun-
IX. (iii) The Fregean Theory of Dernonstratives terfactual circumstances it is the actual demonstratum-in the example,
Let us accept, tentatively and cautiously, the Ftegean theory of demon- Paul-that is the relevant individual. The Fregean theory claims that
strations, and turn now to the Fregean theory of demonstratives.4l the proposition is to be construed as if the sense of the demonstration
According to the Fregean theory of demonstratives, an occurrence of were the sense ofthe demonstrative. Thus, in counterfactual situations
a demonstrative expression functions rather like a place-holder for the it is tlre individual that uould. have been demonstrated that is the rele-
associated demonstration. The sense of a sentence containing demon- vant individual. According to the direct reference theory, demonstraiives
stratives is to be the result of replacing each demonstrative by a con_ are rigid designators. According to the Fregean theory, their denotation
stant whose sense is given as the seuse of the associated demonsiration. varies in different counterfactual circumstances as the demonstrata of
An important aim of the Fregean theory is, of course, to solve Frege,s the associa,teddemonstra,tion would vary in those circumstances.
problem. And it does that quite neatly. You recall that the Freqean The earlier distinciion between Pat and tr{ike, and the discussion
accounted for the informativeness of of counterfactual circumstances in which, as we would norv put it, tl-re
demonstration would have demonstrated nothing, argue that with re-
That [I{es] = ihat lPhos] spect to the problem of associating propositions with utterances the
in terms of the distinct senses of Hes and Phos. Now we see that the direct reference theory is correct and the Fregean theory is wrong.
senses of the trvo occurrences of 'that' are identified with these two I have carefully avoided arguing fot the direct reference theory by
using modal or subjunctive sentences for fear the Fregean would claim
distinct sensesso that the ultimate solution is exactly like that given by
that the peculiarity of demonstratives is not that they are rigid designa-
Frege originally. The sense of the left ,that' differs from the sense of the
right'that'. tors but that they always take primary scope. If I had argued only on
the basis of our intuitions as to the truth-value of
IX. (iw) Argurneut Against the Fbegearr Theory of If Charles and Paul had changed chairs, then lie (Delta)
Dernoustratives rvould not norv be living in Princeton
Let us return now to our original example:
such a scope interpretation could be claimed. But I didn't.
He [Delta] now lives in Princeton, New Jersey The perceptive Fregeans among you will have noted that I have said
nothing about, how Frege's problem fares under a direct refereuce theory
where 'Delta'is the name of the relevant demonstration. I assume that
of demonstratives. And indeed, if 'that' accompanied by a demonstra-
in the possible circumstances described earlier, paul and Charles hav_
tion is a rigid designator for the demonstratum, ihen
ing disguised themselves as each other, Delta would have demonstrated
Charles. Therefore, according to the Fregean theory, tl.re proposition I that (I{es) - that (Phos)
just expressed, Pat, would have been false under the counterfactual cir-
cumstances of the switch. But this, as argued earlier, is wrong. There_ Iooks like two rigid designators designating the same thing. Uh Oh! I
fore, the Fregean theory of demonstratives though ii nicely solves Frege,s will return to this in my Epistemological Remarks (section XVII).
problem, is simply incorrect in associating propositions.uitL r.,rt"ru,r""".
Let me recapitula.te. We compared trvo theories as to the or.onosition
e x p r e s s e db y a s e n t e n c ec o n r a i n i n g a d e m o n s t r a t i v e a l o n g * i r h a u a s s o -
alThe
Fr.egean theory of demonstrations is not a part of my obvious and unconrr.over-
sial iheory of indexicals. On the contrary, ih has the faEcinrarion of the specularive.
518 David I{aplan
Demonstratives blg
-=
ET
.;lE
$
ia
522 David I(aplan Demonstratives 523
tioning as its demonstration. (Unless you hold a Fregean theory of If you try out the notion of truth on a few examples, you will see
demonstratives, in which case its meaning is as stipulated above.) t,hat it, is correct. If I now uiter a sentence, I will have uttered a truth
Norv we can come much closer to providing genuine synonyms. just in case what I said, the content, is true in lrlese circumstances.
'I' means the same as 'dihat As is now common for intensional logics, we provide for the notion of
lthe person who utters this
a struclure, comprising a family of circumstances. Each such structure
token]'.
will determine a set of possible contexts. Truth in a structure, is truth
(The fact that this alleged synonymy is cast in the theory of utterances in every possible context of the structure. Logical truth is truth in every
rather than occurrences introduces some subtle complications, which structure.
have been discussed by Reichenbach.)
t
c
524 David I(aplan Demonstratives 525
"N{ontgomery Furth is tired on 2Z trIarch 1972.,, But there is another (1) A demonstration is a way of presenting an individual.
sense of meaning in which, absent lexical or syntactical ambiguities, trvo
occurences of the sozne word or phrase rnus, mean the same. (Other_ (2) A given demonstration in certain counterfactual circumstan-
wise how could we learn and communicate with language?) This sense ces would have demonstrated (i.e., presented) an individual
of meaning-which I call characier-is what determineJ the content of other than the individual actually demonstrated.
arl occurrence qf a word or phrase in a given context. For inde_ticals,
(3) A demonstration which fails to dehonstrate any individual
the rules of language constitute the meaning in the sense of characler.
might have demonstrated one, and a demonstration which
As normally expressed, in dictionaries and the like, these rules are in_
demonstrates an individual misht have demonstrated no rn-
complete in that, by omitting to mention that indexicals are directly
dividual at all.
referential, they fail to specify the full content of an occurrence of an
indexical. So far we have asserted thai it is not an essential property of a given
Three important features to keep in mind about these two kinds of demonstration (according to the Fregean theory) that it demonstrate a
meaning are: given individual, or indeed, that it demonstrate any individual at a.ll.
1. It is this feature of demonstrations: that demonstrations which in fact
Character applies only to words and phrases as types, con_
tent to occurrences of wotds and phrases in contexts. demonstrate the same individual might have demonstrated disiinct indi.
viduals, which provides a solution to the demonstraiive version of Frege's
2. Occurrences of two phrases can agree in content although problem (why is an utterance of'that [IIes] = that lPhos]' informative?)
the phrases differ in character, and two phrases can agreein analogous to Frege's own solution to the definite description version.
character but difer ir content in distinct contexts. There is some theoretical lattitude as to horv rve should regard such
other features of a demonstration as its place, time, and agent. Just
The relations)rip of character to content is someihinE lilie to fix ideas, let us regard all these features as accidental. (It may be
that traditionally regarded as tlre relationship of sen-seto helpful to think of demonstrations as types and particular performances
denotation. character is a way of presentingco;rtent. of therl as their lol'ens). Then,
I
demonstrations was simply in order to establish the follJwing structural given by the structure but by the evaluation rules.)
features of demonstrations:
1. A demonstration, when set in a context (i.e., an occurrence XVI. Alternatiwe Treatments of Demonstrations
of a demonstration), determinesa cotrLent. The foregoing development of the Fregean theory of demonstrations is
{3Since,
remarkect earlier, rhe speaker and ditrerenr rnenbers of rhe audience not inevitable- Alicl.raelBennett has proposed that only places be demon-
".
aeneraUy have differcnL perspectives on ihe demons!ration, ii may apr)ear sl;ahily strata and that we require an explicit or implicit common noun phrase
differ.ent to each of them. Thus each may take a stighrly cliftererit j"r,.or,"c..t.io, to accornpany the demonstrative, so that:
to have been perfor.rned. Insofar as the aaeni arrd gir"r, conre\t can
d j l f e r i n l o c a t i o n , i h e I o c a t i o n o f a c o n t e x r i s t h e l o" "caai r. i. or .n. o l" f t h"e a g e n t . a eW e s h o u l d n o r , o I c o u r s e , f o r g e t L h e m a n y d i s a n a l o g i e s n o t e d e a l l i e r . n o r l a i l t o n o t e
Therefor.e
the demonstratum of a given demonsrrarion s€t in a given contixl rvill be the that though a description is associaied rvith a palticular characl.er by linguistic
individual, if any, rhereby demonsirared h.om rhe speaker's point
of vierv. conlrenlion, a demonstration is associated Nith ;t.' character by tatu'e.
|
528 D.uid Itaplan
Demonstratives 529
:t
4
ti
530 David Kaplan
Demonstratives 531
i
532 David Kaplan Demonstratives 533
could be in the same tot,al cognitive state and still, as we would say, time is presented, as 'yesterday' or as 'this March 26'. If I see, reflected
believe difierent things. This doesn't prove that the cognitive conient in a window, the image of a ma,n whose pants appear to be on fire, my
of, say, a single sentence or even a word is to be identified with its behavior is sensitiveto whether I think,'His pants are on fire'or'N4y
character, but it strongly suggests it. pants are on fire', though the objeci of thought may be the same.
Let me try a different line of argument. We agree that a given con- So long as Frege confined his attention to indexical free expressions,
tent rnay be presented under various characterg and tha,t consequently and given his theory of proper names, it is not surprising that he did
we may hold a propositional attitude toward a given content under one not distinguish objects of thought (content) from cognitive significance
character but not under another. (For example, on N4arch 27 of this (character), for that is the realm of fixed, chatacter and thus, as already
year, having lost track of the date, I may continue to hope to be finished remarlied, ihere is a natural identification of character rvith content.
by this March 26, without hoping to be finished by yesterday.) Now Frege does, horvever, discuss indexicals in two places. The flrst passage,
instead of arguing that character is what we would ordinarily call cog- in which he discusses 'yesterday' and 'today'I have already discussed.
nitive significance, Iet me just ask why we should be interested in the Everything he says there is essentially correct. (He does not go far
character under which we hold our various attitudes. Why should rve be enough.) The second passage has provoked felv endorsements and much
interested in that special kind of significance that is sensitive to the use
skepticism. It too, I believe, is susceptible of an interpretai,ion which
of indexicals; 'I', 'here', 'now', 'that', and the like? John Perry, in his rnakes it esseniially correct. I quote it in full.
stimulating and insiglitful paper "Frege on D€monstratives" asks and
answers this question. [Perry uses 'thought' where I would use 'object Norv everyone is presented to himself in a particular and
of thouglrt' or 'content', he uses 'apprehend' for 'believe' but nole llt.al primitive way, in which he is presented to no one else. So,
ollt.er psllchological terbs would, yield analogous cases. T have talien a when Dr. Lauben thinks that he has been wounded, he will
few liberties in substituting my own terminology for Perry's arcl have probably talie as a basis this primitive way in which he is
added the emphasis.] presented to himself. And only Dr. Lauben himself can
Why should lve care under rvhat character someone appre- grasp thoughts determined in this way. But norv he may
hends a thought, so long as he does? I can only slietch the want to communicate with others. He cannot communicate
barest suggestion of an answer \ete. We use Lhe ntanner a thought which he alone can grasp, Therefore, if he now
of presenlalion, llte characler, 1o indiaiduale psycltological says 'I have been wounded', he must use the 'I' in a sense
slates, in ezplaining and predicling action. 7t is the manner that can be grasped by others, peihaps in the sense of 'l.re
of presertation, the character and not the thought appre- who is speaking to you at this moment', by doing which he
hended, tliat is tied to human action. When you and I have rlakes the associated conditions of his utterance serve for the
beliefs under the common character of 'A bear is about to expression of his thought.56
attack me', we behave similarly. We both roll up in a ball
What is the particular and primitive way in which Dr. Lauben is
and iry to be as still as possible. Different thoughts appre- presented to himself? What cognitive content presents Dr. La,uben to
hended, same characier, same behavior. When you and I himself, but presents him to nobody else? Thoughts determined this
both apprehend ihat I am about to be attacked by a bear, way can be grasped by Dr, Lauben, but no one else can grasp lAol
rve behave differently. I roll up in a, ball, you run to get
thought determired ir liol way. The anslver, I believe, is, simply, tliat
help. Same thought apprehended, different characters, dif-
Dr. Lauben is presented to himself under the character of'I'.
ferent behaviors.5s
A sloppy thinker might succumb to the temptation to slide from
Perry's examples can be easily multiplied. lvly hope io be finished an aclinorvledgement of the privileged perspecliae we each have on our-
by a certain time is sensitive to how the content corresponding to the selves-only I can refer to me as'I'-to the conclusions: first, that
55John Perry, p. 494. 56Gottlob
"Frege on Demonstlatives," :' Flege, "The Thought: A Logical Inquily," p. 298.
::
:'
iT
534 David I(aplan Dernonstratives 535
ilris perspective necessarily yields a privileged picture of what is seen sense of the demonstration is the sense of the indexical, but the sloppy
(referred to), and second, that this picture is what is intended when one thinker commits an additional error in believing t,hat such senses are ln
'I'). These conclusions,
malies use of the privileged perspective (by saying any way necessarily associated with uses of pure indexicals. The slide
'I' provides the
even if correct, are not forced upon us. The character of from privileged perspective to privileged picture is the sloppy thinker's
acknowledged privileged perspective, whereas the analysis of the content original sin. Only one who is located in the exact center of the Sahara
'I' provides for (and needs) no privileged 'here', but aside from
of particular occurrences of Desert is entitled to refer to that place as that,
pictures. There may be metaphysical, epistemological, or ethical reasons the place may present no distinguishing features.ss
why I (so conceived) am especially impo anl to myself. (Compare: The sloppy thinker's conclusions may have another source. Failure
why n,ow is an especially important time to me. It too is presented in to distinguish betrveen the cognitive significance of a thought and the
a particular and primitive way, and this moment cannot be presented thought itself seems to have led some to believe that the elements of an
a.t any other tirrre in the same way.)57 But the phenomenon noted by object of thought must each be directly accessible to the mind. From
Frege-that everyone is presented to himself in a particular and primitive this it follorvs that if a singular proposition is an object of thought,
way-can be fully accounted for using only our semantical theoly. the thinlier must somehow be imrnediately acquainted with each of the
Furthermore, regarding the firsi conclusion, I sincerely doubt that individuals involved. But, as we have seen, the situation is rather dif-
'I', a particular, ferent from this. Singular propositions may be preseuted to us under
there is, for each of us on each occasion of ihe use of
primitive, and incommunicable Fregean self-concept which rve tacitly characters which neither imply nor presuppose any special form of ac-
express to ourselves. And regarding the second conclusion: even if Cas- quaintance witl.r the individuals of the singular propositions. The psy-
tor were sumciently narcissisiic to associate such self-concepts rvith his chological states, perhaps even the epistemological situations, of Castor
every use of 'I', his twin, Pollux, whose mental life is qualitatively iden- and Pollux are alike, yet they assert distinct singular propositions when
tical with Castor's, would associate the some self-concept with iis every they each say 'N,Iy brother was born before me'. IIad they lived at dif-
(matching) use of 'I'.58 The second conclusion would lead to the absurd ferent times they might still have been situated alilie epistemologically
'I', ihey do not thereby dis-
result that when Castor and Pollux each say seSo far, we have tirnited our atteniion to the 6rst tluee sentences of the quotation
tinguish themselves from one another. (An even more astonishing result
from Frege. How are we to account for the second part of Frege's remarks?
is possible. Suppose that due to a bit of self-deception tlte self-concept Suppose Dr. Lauben wants to conununicate his thought without disiurbing.its
held in comn.ron by Castor and Pollux fits neither of them. The second cognitive content. (Think of trying to tell a color-blind person that the green
conclusion then leads irresistibly to the possibility that when Castor and light should be replaced. You would have to find another way of cornrnunicating
Pollux each say 'I' they eaclrrefer to a tlrirJ partyl) Dhat you wanted to g€t across.) He can't corrurlunjcate l,{al thought with tha,
significance, so, he himself would have to aitach a nonstandard signiffcance to 'I'.
The perceptive reader rvill have noticed that the conclusions of the Here is a suaaestion. He points at his auditor and uses the dernonstr.a.tive
'you'.
'I'are not unlilte those of the
sloppy thinlier rega.rding the pure indexical If w e neglect fine differences in p e r s p e c t i v e , t h e d e m o n s t r a i i o n w i l l h a v e t h e san,e
Fregean regarding true denonstrat,ives. The sloppy thinlier has adopted character for all present and it ceriainly will have the same demonsi.rat.um for all
'I' is synonymous rvitii 'this person' presenh, tlrerelore the demonstrat.ive will have the sarne cllara.cter a.nd conteTLtlor
a d.enronslraliuetheory of intlericois: all present. The indexical 'now' will celtainly have the sane character and contert
'norv' n'ith 'this
falong with an appropriate subjectiue demonstration], for all present. Thus
'the pelson who is speaking to you
boinis] now' rvill have
t i n - r e ' , ' h e r e ' r v i t h ' t h i s p l a c e ' f e a c h a s s o c i a t e dw i t h s o m e d e m o n s t r a t i o n ] , character and content lor all those present. Unforlunately the conient
etc. Like the Flegean, the sloppy thinlier errs in believing that the is not that of 'I' as Dr. Lauben standardly uses it. He needs a demonstraiive like
'dthat'to convert the description to a telm with a nxed content. He chooses the
s?At other times, earlier and later, rve can know it only externally, by description as demonstrative'he', with a lelative clause construction to nake clear his intention.
'he
it were. But no\v we are directly acquainted with ii. (I believe I owe this pojni to Now, if Dr. Lauben uses '1' wilh the nonsiandard meaning usually attached to
John Perry.) who is speaking to you lpoints] now' he will have found a way to corrunuuicale
53 Unless, of course, the self-concept involved a bit of direct rcference. In rvltich case his original thought in a folm vhose cognitive significance is cornmon to all. Very
(when direct relerence is adnitied) uhere seerns no need for the *hole ilreory of clcver, Dr. Lauben.
Fregean self-concepcs. Unless, of course, direct reference is limited to iterns of [Perhaps it is poor pedagogy to join this fanciful interpretation of the second
direct acquaintance, of which more betow. part ol the passage with the serioue inierpretation of the REt part.]
*
536 David l(aplan
Demonstratives 532
while asserting distinct singular propositions in saying 'It is quiet here 'a =
a and B, / will be uniformative to anyone appropriately situ-
now'. A kidnapped heiress, locked in the trunk of a car, knorving nei-
ther the iime nor where she is, may think'Ib is quiet here now'aud the ated, epistemologically, to be able to use these terms.62 I hope that
indexicals will remain directly referential.6o my discussion of the two kinds of meaning-content and character-will
have shown the Direct Acquaintance Theorist that his views are not
E. Corollary 2 Ignorance of lhe referent d,oesnol d,efeat llte directly the inevitable consequence of the admission of directly referential terms,
referenlial characler of indexicals. From the point of view of a lover of direct reference this is good, since
the Direct Acquaintance theorist admits direct reference in a portion of
From this it follows that a special form of knowledge of an object is language so narrow that it is used only by philosophers.63
neither required nor presupposed in order that a person may entertain I have said nothing to dispute the epistemology of the Direct Ac-
as object of thought a singular proposiiion involving that object. quaintance theorist, nothing to deny that there exists his special kind of
There is nothing inaccessible to the mind about the semantics of di- object with which one can have his special kind of acquaintance. I have
only denied the relevance ofthese epistemological claims to the semantics
rect reference, even rvhen the reference is to that which we know only by
of direct reference. If we sweep aside metaphysical and epistemological
description. What allows us to take various propositional attitudes to-
pseudo-explanations of what are essentially semantical phenomena, the
wards singula,r propositions is not the form of our acquaintance wiih the
objects but is rather our ability to manipulate the conceptual apparatus result can only be healthy for all three disciplines.
of direct reference.6r Before going on to further examples of the tendency to confuse meta-
physical and epistemological matters with phenomena of the semantics
Tlre foregoing remarks are aimed ai rcfuting Direct Acquainlance
of direct reference, I want to briefly raise the problem of cognilive dy-
Theories of direcl reference. According to such theories, the question
zamics. Suppose that yesterday you said, and believed it, "It is a nice
whether an utterance expresses a singular propositiou turns, in the first
day today." lVhat, does it mean t,o say, today, thai you have retained
instauce, on the speaker's knowledge oJ lhe referent raiher than on the
that beltef? It seems unsatisfactory to just believe the same content
form of lhe reference, If the speaker lacks the appropriate form of ac-
under any old character-rvhere is the relenlion?64 You can'l believe
quaintance with the referent, the utterance cannot express a siugular
proposition, and any apparently directly referring expressions used must 62For sorne consequences of tl s viet, wiih regard to the interpretation
of demon-
be abbreviations or disguises for something like Fregean descriptions. stratives see "Bob and Carol and Ted and Alice," appendix VIl,
63'llhere is an obvious
connection betweed the fix in which the Direct Acquarncance
Perhaps the Direct Acquaintance theorist thought that only a theory Theorist finds himself, and /(ripte b problem: t,ow can
'a
=B' be informative if d
like his could permit singular propositions while still providing a solu- and B differ in neither denotation nor sense (nor, as I shall suAAest is the case {or
tion for Frege's problem. If we could direclly refer io a, given object in plop€r narnes, character)?
6aThe sort of case I have in rnind is this. I first think, r'His panrs are on fire." I
nonequivalent ways (e.g., as'dthat[Hes]' and'dthat[Phos]'), rve could
later r€alize, and thus corne to think "My pants are on fire." Still later,
not-so he thought-explain the difference in cognitive significance be- I decide that I was wrona in thinking ,,His pants w€re
"I arn he" and conclude
tween the appropriate insta,nces of 'a = qt and 'a = d. I{ence, the on fire." ff, in facl, I on he, have I retained, ny beliel that my pants are on ffre
objects susceptible to direct reference must not permit such reference in simply because I believe the sarne content, though under a diflerent character?
inequivalent rvays. These objects must, in a certain sense, be u'holly lo- (I also deny that content under the former, but for change of tense, character,)
When I first thought "My pants are on fire," a certain singular proposition, call it
cal and completely given so that for any two directly coreferential terms 'Eek'; was the object
of ihought. At the later stage, both Eek and its negaiion are
60Can the heiress plead believed by me, In this sense, I stiU believe rvhat I believed before, namely Eek.
that she could not have believed a singular proposirion
But this does not capture my sense of reiaizins a ,elt€l: a sense that I associate
involving the place p since when thinking 'her€' she didn't lno@ she was at p, that
with saying ihat some people ha,ve a very rigid cosnitive siructure whereas othe!.s
she was, in fact, unacquainted with the place p? Nol Ignorance of the r-€fereni is
are wery flexible. It is temptina to say that cognitive dynarnics is concelned not
61This with retention and change in rvhat is believed, but with retention and change in the
makes il sound as if an exact and conscious masiery of sernanrics is pr.er'e.l-
uisite io having a singular proposition as object of thought, I wilt h.y to nnd a characters under which our beliefs are held. I think that this is basically correct.
But it is not obvious to Ine what relation between a character under rvhich a betief
belt€r rvay Lo express the point in a succeeding draft.
is held at one time and the set of characters under rvhich beliefs are held at a larer
il
538 David Kaplan Demonstratives539
that content under the same character. Is there some obvious standard tr(dthatfa] = dthat[B])
adjustment to make to the character, for example, replacing loday with
geslerd.ay? If so, then a person like Rip van lVinkle, who loses track of may be true, although
time, can't retain any such beliefs. This seems strange. Can we only
reloin beliefs presented under a fixed character? This issue has obvious dthat[a] : dthat[B]
and important connections with Lauben's problem in trying to com-
municate the thought he expresses lvith 'I have been wounded'. Under is not logically t,rue, and is even logically equivalent to the contingency,
what character must his auditor believe Lauben's thought in order for
d=p
Lauben's communication to have been successful? It is important to
note that if Lauben said 'I am wounded' in the usual meaning of 'I', (I call / and ry' logically equivalent when ! * / is logically true.)
there is no one else who can report what he said, using indirecl dis- These cases are reminiscent of l(ripke's case of the terms, 'one meter'
course, and convey the cogniiive significance (to Lauben) of rvhat he
and'the length ofbar z'. But where I{ripke focuses on the special episte-
said. This is connected with points made in section VIII, and has inter-
mological situation of one who is present at the dubbing, the descriptive
esting consequences for the inevitability of so-called de re constructions
meaning associated rvith our directly referential term dthat[a] is carried
in indirect discourse languages which contain indexicals. (I use 'indirect
in the semantics of the language.65
discourse' as a general term for the analogous form of all psychological How can something be both logically true, and thus cerlain, and
verbs.) conlin.genl at the same time? In the case o{ indexicals the answer is
A prime example of the confusion of direct reference phenomena with easy to see.
metaphysical and epistemological ideas was first vigorously called to our
attention by Saul I(ripke in Naming and NecessiTy. I wish to parallel E. Corollary 3 The bearers of logical Trulh and of conlingency are d,if-
his remarlis disconnecting lhe a priori and the necessorg. entilies. It is the character (or, the senlence, if gou prefer)
ferenl
Tl.re form of a prioricily that I will discuss is that of logical trulh (in that is logically lrue, prod,ucing a True conlenl in eaery conlert. Bul
the logic of demonstratives). We sarv very early that a truth of the logic ii is llre content (|he proposilion, ifyou will) thal is conlingenl or
of demonstratives, like "I am here now" need not be necessary. There necessary.
are many such cases of logical truths rvhich are not necessary. lf a is
any singular term, then As can readily be seen, the modal logic of demonstratives is a rich
and interesting thing.
e = dthatla]
65 A case of a seerningly difierent, kind is that, of ihe logical equivalence between an
is a logical truth. But arbitrary sentence C and the result of prelixing eiiher or both oI the indexical
operators, 'ii is actually the case thai' (symbotized'n') and 'it is now ihe case
'(d *
U(a = d[nat,Lo]) that' (syrnbolized 'N'). lfhe biconditional ANdl is losicalty true, bur
prenxing either 'O' or its tenrporal count€rpart can Iead to falsehood. (This case
was adverted to in footnote 28.) It is interesiing to noie, in ihis case, thal the
is generally {alse. We can, of course, also easily produce the opposite parallel between modal and temporal modifications of sentences carries over to
efiect. indexicals. The foregoing claims are verified by the formal system (sections xVIII
and xlx, see especially Rernark 3). Note that the lorrnal system is constructed
time would constitute retaining the original belief. Where indexicals are involved,
in accordance wiih Carnap's proposal thai the intension of an expression be that
for the reasons Aiven below, \re cannot simply require that ihe very same char.acter the €xtension of the expression rvitlr
{unction which assiens to each circurrstance,
still appear at the later i,ime. Thus lhe problern of cognilive dynarnics can be put been thouaht to insure that
respect to that circurnstance. This has cornmonly
like ihis: what does it m€an to say of an individual who at one time sincer.ely logically equivalenl expressions have the same intension (Church's Aliemative 2
asserted a sentence containing indexicals that at some later tirne he has (or has among principles of individuation for the notion of sense) and that logically true
rot) chansed. /ris rnind with lespect to his assertion? What sentence or sentences express the (unique) necessaly proposiLion. Homervork Problem: What
must he be willinc to assert ai the later tinle? weni rvr.onE here?
I
540 David Kaplan Demonstratives 541
It is easy to be taken in by ihe effortless (but fallacious) move from XVIII. The Forural Systern
certainty (logical truth) to necessity. In his important article "Three
Just to be sure we have not overlooked anything, here is a macl-rine
Grades of Modal Involvement,"66 Quine expresses his scepticism of the againsi which we can test our intuitions.
first grade of modal involvement: the sentence predicate and all it stands
for, and his distaste for the second grade ofmodal involvement: disguis-
The Language LD
ing the predicate as an operator 'It is necessary that'. But he suggests
that no new metaphysical undesirables are admitted until the third grade The Language LD is based on first-order predicate logic with ideniity and
of modal involvement: quantification across the necessity operator into descriptions. We deviate slightly from standard formulations in using
^n ^nan.cnian.a two sorts of variables, one sori for positions and a second for individuals
I must protest. That first step let in some metaphysical undesirables, other than positions (hereafter called simply'individuals').
falsehoods. All logical truths are analytic, but they can go false rvhen
you back them up to'E'. Prirnitive Symbols
Ore other notorious exan-rpleof a logical truth which is not necessary,
Primitive Symbols for Two Sorted Predicate Logic
I e-tisi,
0. Punctuation' (, ), [, ]
One can quickly verify that in every context, this character yields a true
proposition-but rarely a necessary one. It seems likely to me that it was 1. Variables:
a conflict betrveen the feelings of contingency and of certainty associated
with this sentence that has led to such painstaking examination of its (i) An infilite set of individual variables: V;
'proofs'. It is just a
truth of logic! (ii) A n i n f i n i t e s e t o f p o s i t . i o nv a r i a b l e s : V o
Dana Scott has remedied one lacuna in this analysis. lVhat of the
premise 2. Predicates:
His discovery was that the premise is incomplete, and that the lasi five 3. Functors:
words
(i) An infinite number of rn-n-place i-functors (functors rvhich
up the logic of demonstratives form terms denoting individuals)
(ii) An infinite number of tn-n-place p-functors (functors which
had been lost in an early manuscript version.6T form terms denoting positions)
4. Sentential Connectives: n, V, -, -, *
5. Quantifiers: V, 3
6 6P r o c e e d i n g t o l t h e
XI Intcrnationat Consrcss ol Ph;losoph! 14,6s-81; reprinred
-_in
\V. V. Quine, Wa!s ol Para.dax (New York: Random House, 1966). 6. Definite Description Operator: the
" / A g a i n , i l i s p r o b a b"lAy e a p e d a g o g i c a l m i s i a k e t o n i x t l i s p l a y f u l p a r a g r a p h w i r h t h €
Precedrng serrous one. 7. Ideniii,y; =
45
.:',
542 David Itaplan Demonstratives 543
A
It 6, t areformulas,ihen (/ A',/r),@v 4,),-A, @ - r/.'),(6 * t) 6. P is a nonempty set (the set of positions, common to all rvorlds)
are formulas
7. T is the set of integers (thought of a.s the limes, common to a.ll
If / is a formula and a € yiuyp, then Va/ and 3al are formulas worlds)
544 David I(aPlan Demonstratives 545
8. Z is a function which assignsto each predicate and functor an Deffnitionr For the following recursive definition, assume that c e C, /
appropriate inlension as follows: is an assignment, t e T, ar.d 1De Wl
(i) If rr is an rn-n-predicate,7, is a function such that for each 1. If a is a variable, lalat- = f(a)
t €T and u eW , Z"(t, w) g (U^ x P") 2. l " r , _ t r a 1 . . . q ^ A t . . . 8 " i t r ( 1 a 1 1 " 7 . - . . . 1 0 " 1 "et ,Z- "l ( t , u )
(ii) If ry is an rn-n-place i-functor, 7n is a function such that for
e a c hl e T a \ d . u € 1 1 , 7 , ( t , u ) € ( U / u { t } ) ( ' / - x P " ) ( N o t e :
3. U n is neither'I'nor 'Here'(see12,13 below),then
t is a completely alien entity, in neither U oor P, which rep- ( Ir(t, w)((la11.1,-. . . lg.b,-)),
resentsan 'undefined' value of the function. In a notrnal set ifnoneoflo;1"11''19*l"t'-
theorywe can take I to be {U ,Pl.) lnor...c^1r...p^l"t'- = I
t*l!"
(iii) If a is an rn-a-placep-functor, Zn is a func_tionsuch that for ( t, othenvise
e a c ht € T a o d w € W , I n ( , w ) € ( P u {I})(u-xP')
(i) F"1,_(6^l) itr tsa,_6kts"1,_,!
L i e U i t r ( f t e T ) ( l u e W ) ( ( ; ) € 7 9 * ; . 1 ( 1u,r ) ) (it) ts.rtu-d itr - 1"164
etc-
10. If c€C, then (ca, cp) e Z1o"o;4(c1,cp)
(i) lf a €Vi, tlien 1"1,-Vol iff VieU, *.17r-d
11. If (i,p) € lu.^t"a(t,u), then (i) € I\*i"r(t,u) (ii) If o € Vp, then F.1r-Yog iff Ype?,l"1gr-d
(iii) Similarly for 3oS
Truth and Denotation in a Context
' 'tx o- (t If o € V;, then:
lVe write: tslJr- 4 for {, when tahen in the context c (under the
assignment / and in the structure 2l), is unieuei € U such that p"r-,,/, if
fthe
lrue with respecl 1o the time I and the Ithe a dl"rr- = { thereis such;
{E:
world ru. ( t, otherwise
(ii) Similarly for a € Vo
We write: l"l7 r- for The tlenolation o/ a, when tahen in the
context c (under ihe assignment / and in i 7. ts"1oa=9 itr lal"1r.- l?l"tu
the structure 4), wilh respecl lo the tirne
I and the world rl 8. (D tsa,..6 ifr Yw'ew,l"1e'6
-t
'21', and we will assume that the .B (ii) ts"ttuOd itr 1w'eW,*"Jt*,o
In general we will omit the supetscript !!
s t r u c t u r el i s ( C , W , U , P , T , 7 ) . o (i) l"t,-F 4 itr f t l € 7 s u c h t h a t t ' > t a n d 1 " 1 r , - d
(ii) +"lt.P o iff f r ' € 7 s u c h t h a t l ' < t a n d 1 " 1 1 , . 6
Definition: f i s a n a s s i g n m e n l ( w i t h r e s P e c tt o \ C , W , U , P ' T , I ) ) i t r : (t\\ ="ttuG4 ln L- -r
r-cl(t-r)uv
3hfz(heuv, k f"ePv" u f = hufz)
1 0 . (i) tsa...Nd itr
Definition: fi =U - {(",/("))}) u {(a,r)} (ii) ?crtuA6 ln
(i.e., the assignmentwhicli is just like / except that it assignsr to a) (iii) F"rt-yd itr
546 David I(aplan
Demonstratives 547
Definition: Rerrralk4: lf $ is closed, theu / is true in c (and 2{) iff for every
.,,tt assignment /, time t, and world - =1,-6.
(i) If / is a formula, {d}} = that function rvhich assigns to each t € 7
fx
and w€W, Ttuth, if l?t,-6, F a l s e h o o do t h e r w i s e .
""4 Deffnition: Where I is either a telm or a formula, llt.e Conlent of I
. ,lx in the conlerl c (in the slruclur"e 2l) is Slable ifffor every assignr.nent/,
(ii) If a is a term, I.]fi = that function which assigns to eacb t e T
{r}} i s a c o n s t a n tf u n c t i o n ( i . e . , { f } } ( / . t u ) = { f } 7 ( t ' , r o ' ) .f o r a l l / .
and wQW,lal"16.
l', ro, and w' in 2l).
548 David Kaplan Demonstratives 549
Rernark 5: Where { is a formula, a is a term, and B is a variable, The sublogic of LD which concerns only formulas of Stable Character is
each of the following has a Stable Content in every context (in every not identical with iraditional logic. Even for such formulas, the familiar
structure): ,4lV{, dthat[c], B, I, Here. Principle of Necessitation (if F d, then f tr/) fails. And so does its
If we were to extend the notion of Content to apply to operators, tense logic counterpart: if F d, then F (-,t'-d tt -F-4 A /). From
we would see that all indexicals (including N,,4,Y, and dthat) have a the perspective of LD, validity is truth in every possible conted. Fot
Stable Content in every context- The same is true of the familiar logical traditional logic, validiiy is truth in every possible circumstance. Each
constants although it does noi hold for the modal and tense operator.s possible context determines a possible circumstance, but it is not the
(not, at least, according to the foregoing development). case that each possible circumstance is part of a possible context. In
patticular, the fact that each possible context has an agent implies ihat
Rernark 6: That aspect of the meaning of an expression which deter- any possible circumstance in which no individuals exist will not form a
mines what its Content will be in each context, we call the Ch.aracter part of any possible context. Within LD, a possible context is repre-
of the expression. Although a lack of knowledge about the context (or sented by (A,c) and a possible circumstance by (U,r,rl,'). To any (21,c),
perhaps about the structure) rnay cause one to mistalie the Content of a there corresponds (4,,ca, c.u'').But it is not the case that to every (2[,t, tr,)
given utterance, the Character of each well-formed expression is deter- there exists a context c of% such that l=c1 and ur=cw. The result
mined by rules o{ the language (such as rules 1-13 on pages 545 and 5.16, i s t h a t i n L D s u c h s e n t e n c e sa s ' f a r E x i s t r ' a n d ' * 3 p L o c a t e d c , p ' a r e
which_are presumably known to all competent speakers. Our notalion valid, although they would not be so regarded in tradiiional Iogic. Ai
'{6}!,' for the
Content of an expression gives a natural notation for the least not in the neotraditional logic that countenances empty rvorlds.
Charicter of an expression, namely '{/}', Using the semantical developments of pages 543-46, we can define this
traditional sense of validity (for formulas wl.rich do not contain demon-
Defirrition: Wlrere f is either a term or a formula, the Character of I stratives) as follows. First note that by Remark 7, if d has a Stable
is that function which assigns to each structure 2{, assignment /, and Character,
context c of U, {f }I. .-,q,
,^
t-.Itu'r 'tL f ctf tilv
Deffnition: Where I'is either a term or a formula, lhe Characler of I Thus for such formulas we can define,
is Stable iff for every structure fl, and assignment /, the Character of I
(under / in E) is a constant function (i.e., {f}} = {f}\t, for all c, c, $ is lrue al I,u (in Z/ if for every assignmeut / and every
,4
in Z). conrex[ c, FcJrup
'Elc Exist c'is not valid evenin LD. Nor are its counterparts,
Ofcourse Rernark 11: Our 0-0-place i-functors a"re not proper narnes, in the
'--F-lz Exist r', and'-P-3c Exist c'. sense of Kripke, since they do not have a Stable Content. But they
This suggests that we can transcend the context-oriented perspec- can easily be converted by means of stabilizing influence of 'dtl-rat'.
tive of LD by generalizing over times and worlds so as to capture tllose Even dthat[a] lacks a Stable Character. The process by which such
possible circumsiances (X,1, u) which do not correspond to any possible expressions are converted into expressions with a Stable Character is
'dubbing'-a form of definition in which context may play an essential
contexts ({,c). We have the following result:
role. The means to deal with such context-indexed definitions is not
(iii) If d contains no demonstratives, available in our object language.
Frd itr Frpu(-F-d ^ -P-4 ^ 4). There would, of course, be no dificulty in supplementing our lan-
guage with a syntactically distinctive set of 0-0-place i-functors whose
Although our definition of the neotraditional sense of validity was ntoti- semantics requires them to have both a Stable Character and a Stable
vated by consideration of denronstrative-free formulas, we could apply it Content in every context. Variables already behave this way, whai is
also to formulas containing essential occurrences of demoustratives. To wanted is a class of constants that behave, in these respects, Iike vari-
do so would nullify the most interesting features of the Iogic ol demon- ables.
stratives. But it raises the question, can we express our llew sense of The difficulty comes in expressing the definition. tr{y thought is
validiiy in terms of the neotraditional sense? This can be done: that when a name, like'Bozo', is introduced by someone saying, in some
context c*, "Let's call the Governor,'Bozo"', u'e have a context-indexed
(iv) llpd itr FrANd definition of the form: A =c. a, where A is a, new constaut (here, 'Bozo')
and a is some term whose denotation depends on context (here, 'the
Remark 10: Rigid designators (in the sense of I(ripke) are terms rviih Governor'), The intention of such a dubbing is, presumably, to induce
a Stable Content. Since Kripke does not discuss demonstratives, l.risex- tlre semantical clause: for all c, {A}! = {a}".r . Such a clause grves A a
amples all have, in addition, a Stable Cl.raracter (by Remark 8) Iiriplie Stable Character. The context-indeiing is required by the fact that the
'rt 'definiens') may vary from context
claims that for proper names d, p may happen that a = B, though Content of a (the to context. Thus
not a priori, is nevertheless necessary. This, in spite of the fact that the the same semantical clause is not induced by taking either ,4 = a or
names o, p may be introduced by rneans of descriptions a', B' for which even ,4 = dihat[a] as an axiom .
a' = pt is not necessary. An atralogous situation holds in LD. LeL at, B' I think it is likely tha.t such definitions play a practically (and perhaps
be definite descriptious (without free variables) such that a1': Bt is rtot iheoretically) indispensable role in the growth of language, allorving us
a priori, and consider the (rigid) terms dthat[a/] and dthaifB'] rvhich are to introduce a. vast stock of names on the basis of a meager stock of
formed from them. We knorv that: demoustratives and some ingenuity in the staging of demonstratlous.
Perhaps such introductions should not be called'definitions'at all,
p(dthat[o'] = dthatlB'] F+ dt -Pt).
since they essentially enrich the expressive porver of the language. lVhat
a nameless man may express by 'I am hungry'may be inexpressible in
Thus, if a' = Bt ),s not a priori, neither is dthat[o/] = dthat[B']. But,
slnce: remote contexts, But once he says "Let's call me'Bozo"', his Content
is accessible to us all.
p (dthat[rr'] = dthat[B/] - tr(dthat[a'] = dthat[B']))
Rernark 12r The strongest form of logical equivalence between two
it may happen that dthat[a'] = dthat[B/] is necessary. The couverse
formulas 4 and, d' is sameness of Character, {6} = 16'}. This forrn of
situation can be illustrated in LD. Since (a = dthatfa]) is valid (see
synonymy is expressible in terms of validity.
Remark 3), it is surely capable of being hnown a prioli. But if cvlacks
a Stable Content (in some context c), tr(cr = dthat[a]) rvill be false.
it
552 David I(aplan
Demonstratives 553
{d}={6,} itr F n[-r'-(ded')^-p-(de4')^ @* d')] 5. Special continuity conditions through time might be added for the
predicate'Exist'.
fUsing Remark 9 (iii) and dropping the condition, which rvas stated
only to express the intended range of applicability of p1, we have:
6, If individuals lacking positions are admitted as agents of contexts,
{6} = {6'} itr Fr(d .* //).1 Sincedefinitionsof the usual hind (as 3(iii) of page 543 should be weakened tot cp eP U {J}. It would
opposed to dubbings) are intended to introduce a short expressron as
no longer be the case that: pLocated I, Here. If individuals also
a mere abbreviation of a longer one, the Character of the defined sign
lacking temporal location (disembodied minds?) are admitted as
should be the sarne as the Character of the definiens. Thus, rvithin LD,
agents of contexts, a similar weakening is required of 3(ii). In any
definitional axioms must talie the unusual form indicated above.
case it would still be true that ! Exist L
for knowing-(or believing)-who.) The term is chosen to reflect the idea that of translating the language of thought). lVe control this by assum-
that the individual in question is identified with respect to some prior ing the direct discourse to be in tl.re language of the indirect discourse
or independent information-re-cognition-not immediaiely connected report. Second, as Carnap once pointed out to me, if Chrisiopher's dis-
with the current attribution.) Of the two elements the former ls sema,n- '6
course had the form A ,/t- even the strictest court would accept as
tical; the latter, frankly epistemological. The English idiom 'There is true the testimony,
rchristopher
said that ,/t A 6-. What logical trans-
someone such that Ingrid indirect-discourse-propositional-attitude-verb formations on the original discourse would be allowed in the report? (If
that ...he...'always implies ihat a singular proposition is the object Christopher says'3c n is round', may we report him as saying that 3y
of Ingrid's thoughi (and thus thai some directly referential term a oc- g is round?) We control this by allowing no logical transformations (*-e
curred in her inner direct discourse) and may sometimes imply (or only are explicating lilercl indirect discourse). Third, if in saying 'The circle
suggesi?) that Ingrid recognized, who cv is. I offer no analysis of the can't be squared' Christopher thought ihat 'can't'was synonymous rvith
latter notion.72 'should not'rather than 'cannot', should he
be reported as having said
In the first paragraph, I referred to a contlolled experiment. By
that the circle can't be squared? We control this by assuming that our
that I mean the following. Accepting the metaphor of "inner direct
speakers make no linguistic errors.
discourse evenis" and "indirect discourse reports" in connection with the
What ihen remains of the logic? Is the move from direct discourse
usual episiemic verbs, I want to examine the logical relations betu'een
these two. But the study is complicated by at least three factors which to Iiteral indirect discourse not simply the result of disquotation (and
obscure the issues I wish to bring to light. First, there is no real syntax decapitaliztion) plus the addition of 'that', as in:
to the language of thought. Thus, even in the case of the simplest Christopher says'the world is round'
thoughts the relation between the synta-x of the sentential complement ... Christopher says that the world is round ?
to the epistemic verb and the structure of the original thought is obscure.
But how then are we to report Dr. Lauben's saying,'I have been wound-
Second, in containing images, sounds, odors, etc., thoughi is richer than
ed'? Certainly noi as, 'Dr. Lauben says that I have been wounded'!
the language of the report. Might these perceptual elements play a
Even in this highly aniiseptic environment, the logic of sags should
role in determining logical relations? Third, thought ranges from the
provide us with a full measure of that baffling and fascinating de re
completely explicit (inner speech) to the entirely implicit (uucouscious
vetsus de diclo, notional versus relational, etc., behavior, And here,
bel.iefs which explain actions) and through a variety of occurreut and using the conceptual apparatus of the semantics of direct reference, we
dispositional forrns. lfhis makes it hard to pin down the rvhole direct may hope to identify the source of these antics.
discourse event. These three factors suggest tahing as a paradigm of
[I also hope to distinguish, in discussing r.eports of self- at tribution,
the relation between direct and indirect discourse-direct and iudirect c says lhal r is a fool, ftorn c sags-hitnself lo be a.fool.)
discoursel
Even when reporting the (outer) discourse of another, at least three
XXI. Russell on Esocentric Particulars and Their
obscure irrelevancies (for our purposes) remain. First, if Christopher
Dispensabilit y
speaks in a language difrerent from that of the report, we have again the
problem of translation (analogous to, though perhaps less severe than, In clrapter YII of Inqtiry Inlo Afeaning antl TrutL,?3 Russell gives a
series of atrocious arguments for the conclusion that "findexicals] are not
any issues of sufficient theoreiical interest io make ;t worth pursuina.
T2There is a considerable literaiure on this subject with inrportant contributions needed in any part of the description of the world, rvhether physical or
by
Hintikla, Castaneda and others. In connection with the proposal that
'a
knows
psychological." This is a happy uo-nonsense conclusion for an argument
who o is' can be symbolized '3'(4
krows that. ' = o)' , ii should be noted rhar o's that begins by remarking "A physicist will not say 'I saw a table', but
like Neurath or Julius Caesar, 'Otto saw a table'." [Why Julius Caesar
knowled8e of the logical trulh rdthat[d] = a- leads, simply by rh€ semancics of
'!c(o
direct ref€rence, to knows that r=d)r. This shows only ihat a recosnition would be provoked to say'Otto sarv a table', is unexplained.]
sense of knowing a singular proposirion is not definab)e, in thc obwious way, in
icrrns of a purely dircct rclcrencc sense oI knowing a singular Froposii;on. T3Bertrand Russell (London: AIIen & Unrvin, lgaO).
:
j
558 David Kaplan Demonstratives 559
let us examine Russell's conclusion without prejudice to his argu- negaiive, without including much supporting argumentation (several of
ment. [What follows is an outliue.] the omitted arguments seem both tedious and tendentious). My current
In brief, there are essentially two points. First: if we have both the inclination is to drop this whole section from the final draft.]
indexicals and an unlimited supply of unused directly referential proper A, uord, is an expression along with its meaning. When two expres-
names, and we ca,n do instantaneous dubbing, then in each conLext c sions have the same meaning, as with "can't" and "cannot", we call the
for any sentence / containing indexicals .we can produce a sentence d* tlvo words sgnongms. When two meanings have the same expression,
whose character is fixed and whose content is the same as that of / in c. we call the two words homongms. In the latter case we also say that
In this sense, if you can describe it with indexicals you can describe it the expression is atnbiguous. (Probably we would say that the word is
without.Ta There are problems: (i) things can change fast and dubbings ambiguous, but accept my ierminology for what follows.) In a disam-
take time, (ii) ihe indexicals retain a kind of epistemic priority. biguated language, semantics can associate meanings with expressions.
The second point is: given any prior collection of proper names, Even in a language containing ambiguities, semantics can associate a set
there will be things, times, places, etc., without a name. IIow do I say of meanings lvith an expression. But given an utterance, semantics can-
something about these unnarned entities? (E.9., how do I tell you that not tell us what expression was uttered or what langua"geit was uttered
your pants a.re on fire-now? It may be that nothing in sighi, including in. This is a presemantic task. lVhen I utter a particular vocable, for
us, and no nearby time has a name.) example, the one characteristic of the first person pronoun of English,
There are two cases. It seems most liliely that without inde-xicals you must decide what uord I have spoken or indeed, if I have spoken
some entities cannot even be uniquely d,escribed, In this case rve are any word at all (it may have been a cry of anguish). In associating a
really in trouble (unless Russell believes iu the identity of indesoibables rvord rvith my utterance you take account of a variety of {eatures of ihe
-objects lacking uniquely characterizing descriptions) because wiihout
context of uiterance that help to d.elermine rvhat I have said but that
indexicals rve cannot freely introduce nerv names. If every entity ccn be need not be any parl of what I have said. N'Iy egotism, n.ry intonation,
uniquely described, there is still ihe problem of not presenting ihe right my demeanor, may all support the hypothesis that it was the first person
content under the right character required to motiva.te ihe right action pronoun of Englisli. But these aspects of personality, fluency, and rnood
(recall the discussion on pages 532-33). The proposition expressed by are no part of any semantic theory of the first person pronoun. The
'the pants belonging to
the x .Fo are on fire at lh,e I Gl' is not the factors I have cited are not, of coutse, criterial for the use of the first
proposition I want to express, and certainly does not have the character person pronoun. What are the criteria? What would definiiively settle
I wish to convey.75 the question? I don't know. I think this is a very difficult question.
But among the criteria there must be some that touch on the utterer's
XXII. On Proper Narr. es intention to use a word in conformity rvith the conventions of a particu-
lar linguistic conmunity. For proper name words, in part because they
[Some thoughts on proper names from the perspective of the formal are so easily introduced, this aspect of the presemantic determination is
system are contained in Remark 11, page 551. lVhat follorvs is the r.nost especially important.
hastily written section of this draft. I sketch a vierv that is rnainlv According to the causal chain or chain of communication theoly,
there are two critical intentions associated with the use of the proper
?4I assume
here that proper Dames are noc indexicals. I araue tlre poinc rn sectron name word. One is the intention to use the word with the meauing
xxIl.
75Some inieresting arsuments given it by the person from whom you learned the word. The other
of a clifferent sort for the indispensabiliry of inciesicals
are aiven by Burae in "Belief De P"e," JDurnal oJ Phtlwoph! 74 (r927): 33S-62, is the contrary intention to crea,te (and perhaps simultaneously use) a
and by Bar-HiUel in his pioneering rvork, "Indexicat Expr.essions,,, .4/ind (195{). proper name lvord to refer to a given object iuespective of any prior
In conneciion rviih the arguments of Burge and Bar-Hillel it rvould be rnceresrrnA
meanings associated witll the expression chosen as a vehicle. One who
to check on some related empirical issues involving linguisric univer.sals. Do all
languages have a firsi person singular form? Do they all have all of the standar.d uses a proper name word with the first intention generally (bui not
indexicals? ahvays) believes that someone originated the vvord by using it rvith the
t,
560 David I(aplan
Demonstratives 561
IVIy airn in the foregoing is to emphasize how delicate and subile our knows that. But he doesn't. What is perhaps even more astounding is
analysis of the context of utterance must be for the presemantic purpose that I may introduce a new proper name wr:rd and send it on its jour-
ofdetermining what words, if any, were spoken. I do this to make plausi- ney. When it returns to me-perhaps slightly disiorted pl.ronologically
ble my view that-assuming the causal chain theory of reference-proper by its trip through other dialects-I can competently take it into my
names are not indexicals. The contextual feature which consists of the vocabulary without recognizing ii as the very same word! Shocking!
causal history of a particular proper name expression in the agent's idi- In earlier sections of this paper I have tried to show that many of the
olect seems more naturally to be regarded as determining what lvord lvas metaphysical and epistemological anomalies involving proper names had
used than as fixing the content of a single context-sensitive word. Al- counterparts involving indexicals, and further that in the case of index-
though it is true that two utterances of 'Aristotle' in different contexts icals these wonders are easily explained by an obvious theory. Insofar
may have difierent contents, I am inclined to attribute this difference as I am couect in regarding the anomalies as counterparts, the the-
to the fact that disiinct homonymous words were uttered rather than ory of indexicals may help to break down unwarranted resistance to the
a context sensitivity in the character of a single word 'Aristotle'. Un- causal chain theory. It may also suggest the form of a general semantical
lihe indexicals like 'I', proper names really are ambiguous. The causal and epistemological scheme cornprehending both indexicals and proper
theory of reference tells us, in telms of contextual features (including names. This is not the place to attempt the latter task; my purpose
the speaker's intentions) which word is being used in a given utterance. here is simply to show that it is not trivial.T8 Those who suggest that
Each such word is directly referential (thus it has a fixed content), and it proper names are merely one species of indexical depreciate the po*'er
also has a fixed character. Therefore, in the case ofproper name rvords, and the mystery of the causal chain theory.
all three kinds of meaning-referent, content, and character-collapse.
In this, proper naure words are unique. They have the direct reference
of indexicals, but they are not context-sensitive. Proper name rvorcls
are like indexicals that you can carry a,way from their original context
without affecting their content,. Because of the collapse of character,
content, and referent, it is not unnatural to say of proper names that
73The i.sues to b€ resolved by "a senerat semantical and epistemolosical scherne
they have no meaning other than their referent.
comprehending ... proper such as these. Is the {'ork of the causal
Some may claim that they simply use 'indexical' in a wider sense chain theory presemantic, as I have claimed? Do proper naInes have a kind of
than I (perhaps to mean something like 'contextual'). But we rnust be meanina oiher than reference? Does th€ caus.l chain theory its€U constitute a
wary of an overbroad usage. Is every ambiguous expression an indexical kind of rneaning for proper names that is analogous to character for indexicals
(but which, perhaps, gives all proper names che same neaning in th;s sense)? Are
because we look to utterer's intentions to disambiguate? Indeed, is er.ery
propernarnes words ofany particular lanaua8e? Is there synonymy between proper
expression an indexical because it might have been a groan? names that are expressed difierently (as the.e is between'can't'and'cannot')?
If the character and content of proper name words is as I have de- How should we describe the linguistic cornpetence of one who does not know that
scribed it (according to the causal theory), theu the informativeness of Hesperus is Phosphorus? Is he guilty of linguistic ellor? Should we say he does
ra=/, 'Hesperus'and'Phospholus'
not know what words he speaks? Does he know that
with a and p proper names, is not, accounted for in terms of
are difi€lent words? Are they? Is it really possible, as I claim, to accounl for
differences in either content or character. The problem is that proper. the semantics of indexicals without rnaking use of thc fuU conceptual r€sources
names do not seem to fit into the whole semantical and epistenological required to account fo. the semantics ol proper names? I raise these issues-aDd
scheme as I have developed ii. I claimed that a competent spealier linorvs there are othels-within the framework oI a hypothetical acceptance of the causal
chain theory. There are other issues, of a quite different kind, involved in trying
the character of words, This suggests (even if it does not irnply) that
to fill out sorne details of the causal chain theory itself. For exampte, if one rvho
if two proper names have the same character, t]re competent speaker has received some particular proper name expression, say, "Jarnes", hundreds of
limes, uses that expression attributively as a proper name, and has in rnind l1o
particular source, how do we decide which branch to follow back? The first set of
well, for example Russell's 'denotins phrases' which do not denoie. Bul it still issues seerns to Ine to be largely independent of the details of the relevant causal