Cogito, Ergo Sum

Author(s): W. von Leyden

Source: Proceedings of the Aristotelian Society, New Series, Vol. 63 (1962 - 1963), pp. 67-82
Published by: Wiley on behalf of The Aristotelian Society
Stable URL: http://www.jstor.org/stable/4544671
Accessed: 28-07-2015 18:02 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact [email protected].

The Aristotelian Society and Wiley are collaborating with JSTOR to digitize, preserve and extend access to Proceedings of the
Aristotelian Society.

http://www.jstor.org

This content downloaded from 157.253.50.50 on Tue, 28 Jul 2015 18:02:01 UTC
All use subject to JSTOR Terms and Conditions

Meeting of the AristotelianSociety at 21, BedfordSquare,London,W.C.1,

on 26th November,1962, at 7.30 p.m.

IV-COGITO,

ERGO SUM

By W. VON LEYDEN

IN a recentpaper,ProfessorHintikkahas raised,often forcefully

and clearly,a numberof validpointsin his analysisof Descartes'
Cogito as a performatoryutterance.' None the less, I am not
convincedthat his analysisis itself free from difficulties. Just as
he charges Descartes with compressingseveral differentargumentsinto the apparentlysimpleformulationcogito,ergosum,so
mighthe be chargedwith mergingsome separateissuesinto each
other and into his apparently simple interpretationof the
Cartesiandictum. I intend to substantiatethis claim by first
pointingto the particularissueswhichin my viewHintikkahasnot
clearly distinguishedfrom each other. I will then advance a
numberof argumentswith a view to bringinginto the open the
complexityof the probleminvolvedin Descartes'Cogito.
1. Difficulties in interpretingthe Cogito as performatory

It would seem, in the first place, that for Hintikka one

importantreasonwhy Descartes'Cogitois to be interpretedas a
performatoryutteranceis that, if I say that I do not exist, my
sayingthis showsthat what I say is false; henceI refuteor defeat
myself by utteringthis sentence. Now it is certainlycorrectto
maintainthat the surmisethat I do not exist,thoughnot formally
is absurdin the sense of self-stultifying;and
self-contradictory,
it is also true that I could not utterthe words" I do not exist",
unlessI existed. But does it follow that for eitheror for both of
these reasons neither " I exist" nor " I am thinking" can be
regardedas a descriptivephrase,and that they must insteadbe
classifiedas performatory
utterances? Afterall, the denialof " I
exist " or " I am thinking" is absurdin the senseof self-defeating
even on the assumption-which Descarteswould seem to have
1Jaakke Hintikka, " Cogito, Ergo Sum: Inference or Performance?",
ThePhilosophicalReview,Vol. LXXI, i (January,1962),pp. 3-32.

This content downloaded from 157.253.50.50 on Tue, 28 Jul 2015 18:02:01 UTC
All use subject to JSTOR Terms and Conditions

68

W. VON LEYDEN

made-that both " I exist " and " I am thinking" are statements
describingor reportingfacts truly and that both " I do not
exist" and " I am not thinking" are statementsdescribingor
reportingfacts falsely. Accordingly,it mightbe arguedthat the
indubitabilityof Descartes'Cogito insightis not necessarilydue
to its performatorycharacternor that, becauseI cannot intelligently,i.e., withoutgivingriseto an " existentially" inconsistent
or self-defeatingstatement,deny " I exist " or " I am thinking",
such a denial must be tantamount to at once enteringinto
and not entering into a commitmentor the performanceof
an act.
Secondly,it is not clear to me whetherHintikkawishes to
interpretas performatorythe whole of the formulacogito, ergo
sum, or only the first part, i.e., the phrasecogito, or again only
the latterhalf, i.e., the phrasesum. The three types of interpretation mergeinto each otherthroughouthis paperin a confusing
mannerand are also, to my mind, each open to criticism. That
Hintikkaintendsthe first of these interpretationswould appear
from the title of his paperand the way in whichhe often speaks
of the performatoryaspectof " the Cogito" or of the performatorilyinterpreted" cogitoargument" (e.g., pp. 23, 27, andsect. 13
ad fin.) Assuming this is his intrepretation,one might ask
whether the issuing of the performativeutterance " cogito"
or alternatively" I exist" is not by itself sufficientto show the
truthintendedby Descartes. For is it not preciselyby the mere
utteranceof " cogito" or alternativelyof " I exist" that the
characterof Descartes'insight may be said to express itself?
Besides,if the whole Cartesianformulabe acceptedas performative in character,an unsatisfactoryconsequencewould appearto
be that by issuingthis utteranceone would both explicitlyassert
somethingwhichis true,i.e., that one exists,and at the sametime
be expectednot to assertanythingat all but merelyto implyor
to showthat one exists.
On the otherhand,if Hintikka'sintentionbe to interpretonly
the word cogito as performative,as would appearfrom what he
says on e.g., pp. 17 and 31-32, he would be splittingup the
meaning of a formula which Descartes obviously regardedas
uniformin character. Besides,would not on this interpretation,

This content downloaded from 157.253.50.50 on Tue, 28 Jul 2015 18:02:01 UTC
All use subject to JSTOR Terms and Conditions

69

COGITO, ERGO SUM

contraryto Hintikka'sclaim (p. 17), the status of " I think",

which Descartes considered unique, become vulnerable to
competition from other performatoryutterancessuch as " I
promise" or " I apQlogize", since the issuing of any of these
mightin some sensebe held to implythat the sentence" I exist "
is true? If howeverit is concededby Hintikkathat the statusof
" I think" is unique, it is difficultto see how he can hope to
establishthis by merelyclassifyingthe sentenceas performatory,
i.e., withoutat the sametime showingit, on independentgrounds,
to be eitherthe only one of its kind or alternativelyuniquewithin
its kind. Of these two alternatives,however,while the former
is difficult if not impossible,the latter necessitatesa further
analysis of the sentence,over and above its interpretationas
performative.
Yet, in the third place, if Hintikka'sintention, as appears
from sections(6), (7) and (13, ii) of his paper,is to interpretonly
the sentence ego sum or " I exist " as performatory, the argu-

ments he adducesto this effect, and particularlyhis point that

it wouldbe absurdor pointlessto denythis sentence,shouldmake
his interpretationalso apply to the phrasecogito or " I think".
For the inconsistencyor absurdity of uttering the sentence
" I am not thinking" does not seemto differfromthat of uttering
the sentence " I don't exist", nor does the sentence " I am
thinking" seemto be less self-verifyingor intuitivelyself-evident
than the sentence " I exist". In fact, the sentences " I am
thinking" and " I exist " share the important characteristics
that, though they are not logicallynecessaryand thereforemay
be denied without self-contradictioneven by the person
who happens to utter them, they are nevertheless certain
and indubitablein that their truth follows from their being
doubted or denied by the person utteringthem, so that their
denial is absurd in the sense of being self-defeating. Hence
the certainty of " I am thinking" can be establishedby the
same criterionas that of " I exist".2 Perhapsthe fact that the
two sentencesare in this sense independentof one another was
one reason for Descartes' frequent insistence on the noninferential nature of his Cogito.
2

Cf. A. J. Ayer, The Problemof Knowledge(London, 1956), pp. 46-7.

K2

This content downloaded from 157.253.50.50 on Tue, 28 Jul 2015 18:02:01 UTC
All use subject to JSTOR Terms and Conditions

70

W. VON LEYDEN

Moreover,a difficultywouldnow also ariseas to the interpretation of ergo sum: is this to mean the fact that I exist or that
I am certainthat I exist? In other words, is Descartes'Cogito
intendedas the embodimentof an ontologyor an epistemology?
Here, I think, Mr. Hintikkais guiltyof a certainconfusion. He
says " the word cogito refers to the ' performance ' (to the act

of thinking)throughwhich the sentence'I exist' may be said

to verify itself" (p. 17). But exactlywhat verifiesthe sentence
" I exist " ? Surely,on Descartes'view, if fully expanded,it is
only becausethe performanceof thinkingshowsor makesexplicit
that I exist, that the sentence" I exist " can be held to be verifiable. I shouldthinkthat, if Descartes'Cogitois to be in any way
interpretedas performative,the gain of such an interpretation
wouldbe preciselythe possibilityof sayingthat the act of uttering
" I think" in the first place shows the fact of my existenceor
gives it to be understoodthat I exist, beforeit can subsequently
serve as the verificationof the sentence" I exist", or help to
demonstrateto myselfthat I knowthat I exist. In my view the
whole point of the Cogitois not so muchthat the indubitableor
certainknowledgeof my own existencedependson my thinking,
as that my existenceis indubitablebecausemy thinking(whichis
indubitable)dependson it or presupposesit. This point is both
obscured and made ambiguous when Hintikka talks of
" Descartes'sintuitiveidea of the dependenceof his existenceon
his thinking" (p. 22). Clearly what depends on Descartes'
thinking is his certain knowledge of his existence, while his
thinkingitself of coursedependson his existence.
In this connexion, I believe, it is also importantto avoid
as Hintikkatends to do (pp. 18 ff.), the role of
underestimating,
introspectionand consciousnessin Descartes'cogito argument.
Descartes'celebratedpassage from scepticismto certaintywas
determinedby the fact that doubt,even total doubt,was for him
not so much the resultof an unshakeablebelief as a short-lived
implicationof his philosophicalmethod.3 The function of this
methodologicaldoubt was to help in fixing the Archimedean
3 See my " Descartes and Hobbes on Waking and Dreaming", Revue
Internationale de Philosophie, Vol. XXXV (1956).

This content downloaded from 157.253.50.50 on Tue, 28 Jul 2015 18:02:01 UTC
All use subject to JSTOR Terms and Conditions

COGITO, ERGO SUM

71

point4of his metaphysics,and this was the fact that, even though
I may doubt, dream,be deceived,or think wrongly,so that what
I am thinkingabout may be false, it is neverthelessindubitable
that I am consciousof somethingwheneverI doubt, dreametc.,
and thereforeI mustthinkand exist. Accordingly,thoughI fully
agreewith Hintikkawhen he says (pp. 18-19), " the reasonwhy
Descartes could not doubt his own existence is in principle
exactlythe sameas the reasonwhy he could not hope to mislead
anybodyby saying'I don't exist ' ", I doubt whetherhe is right
in claimingthat " the one does not presupposeintrospectionany
more than the other" (p. 19). In my view, as I shall try to
explain later, Descartes'Cogito insight is not wholly separable
fromwhatusedto be calledactsof introspectionor self-awareness.
And it seemsto me that it is only by disregardingthis aspectof
the questioncompletelythat Hintikkahas been able to lay so
much stress on an over-allinterpretationof the Cogito as performative.
Before advancingmy own suggestionfor an analysisof the
Cogito,however,I may add here that in connexionwith such an
analysisit is important,more so even than Hintikkawould seem
to recognize,to distinguishclearlyand persistentlybetweenthree
differentissues, namely (a) that of ascertainingthe meaningof
the Cogitoas intendedby Descartes;(b) that of ascertainingwhat
the Cogito may mean in connexion with any particularphilosophicalissue, or what it has been held to mean in one or the
to which
otherof the hundredsof discussionsand interpretations
this Cartesianprinciplehas givenrise;and (c) that of ascertaining
whatit must meanin orderfor it to be acceptedas a proposition
one knowsto be true or a piece of reasoningone can show to be
valid.
2. The indispensability of the first-person pronoun " I" in the
formulation of Descartes' insight

It has beena commonplacein discussionsof Descartes'Cogito

to arguethat his use of the phrase" I think", from his own as
4Second Meditation,ad. init., ThePhilosophicalWorksof Descartes, ed.
Haldane and Ross (Cambridge,England, 1931), Vol. I, p. 149; Reply to
Bourdin'sseventhset of Objections,Vol. II, p. 271.

This content downloaded from 157.253.50.50 on Tue, 28 Jul 2015 18:02:01 UTC
All use subject to JSTOR Terms and Conditions

72

W. VON LEYDEN

well as from a more generalpoint of view, can mean no more

than "thinking is going on " or "there is a thought now ".
That any referenceto a mind or self is illegitimateon Descartes'
own premissesis indicatedby the fact that whenhe firstthought
of the cogito foimula or uttered it as the one self-certifying
propositionsupplantinghis allegedtotal doubt,he couldnot then
have used the word " I " in this formulato describea substance
or spiritualentitythat enduredas the unitarybasis of his mental
life, for all this would have had to remainfor him a matterof
absolute doubt. Then there was of course Hume's scepticism
regardingthe mind and personalidentity, and his inability to
prove that thoughts require a thinker. Accordingly,Russell
suggested5that the word " I " shouldbe omittedaltogetherfrom
the Cartesianformula. Now this interpretativemove would
depriveDescartes'formulaof neitherits prima facie plausibility
nor its importance. For if his originaldoubt was twofold, i.e.,
whetheranythingexistsandwhetherthereis any true knowledge,
the incontestableaffirmationof such a fact as that there is a
thoughtnow would disposeof both doubts,withoutreferenceto
a self or the employmentof the word " I ".
None the less, it appearsthat what Descartesclaimedas his
insightcannotbe made whollyplausibleunlessit is formulatedin
terms of a first-personsingularsentence. His own reason for
preferringto expressthe cogito statementin the first ratherthan
the third person was not merely that it would have been
grammatically incorrectto say " il suis ", or fallaciousto talk
of " thought thinking", or that the Frenchlanguagemade no
provisionfor the phrase" it thinks"6 in analogyto the German
"es blitzt".

Nor is it easy to concur in Collingwood's7 belief

that Descartes'paiticularformulationwas due to the desire to

emphasizethe concretehistoricalfact, the fact of one's actual
5 Cf. BertrandRussell, An Outlineof Philosophy(London, 1927), p. 171;
History of WesternPhilosophy(London, 1946), p. 589.
6 For the suggestion of this phrase see G. C. Lichtenberg, Werke,ed.
Goldschmit (Stuttgart,1924), p. 78, and Wittgenstein'sview as reportedby
G. E. Moore, " Wittgenstein'sLectures in 1930-33 ", Mind, Vol. LXIV
(1955),pp. 13-14. Cf. also P. F. Strawson,Individuals(London, 1959),p. 95,
n. 1.
7 R. G. Collingwood,SpeculumMentis (Oxford, 1924), p.
202.

This content downloaded from 157.253.50.50 on Tue, 28 Jul 2015 18:02:01 UTC
All use subject to JSTOR Terms and Conditions

COGITO, ERGO SUM

73

presentawareness,as the root of science. Descartes'reasonwas

rather,one mightsay, semantical. He realized,like St. Augustine
and a numberof mediaevalthinkersbefore him,8that, just as I
cannot doubt concerningthat of whichI am conscious,9because
statesof awarenessaretheirown guarantee,so aretherea number
of first-person sentences, e.g., " I am thinking ", " I live ", " I

know some Englishwords", etc., whichit would be in a certain

senseabsurd,or at any rateveryrepugnantto whoevermakesthe
statement,to deny or to contradict. Whateverit is, for instance,
that I see whenI claimto see a hawk,therecan be no doubtthat
in one sense, though in one sense only, my utterance" I see a
hawk" is true, even if thereis no hawkto be seen, and that one
could not in this sense be said to " see falsely . Similarly,the
statement " I am thinking" as also the statement " I live "
are both in a senseincontrovertible
and self-guaranteeing.However, the peculiarreason for this here is that to doubt or deny
eitherstatementpreciselyprovesit to be true: we thinkin saying
that we do not think and we cannotbut exist while denyingour
existence.
It is perhapschieflybecauseof this logicallyuniquestatusin
whichcertainfirst-personforms of assertionfind themselvesthat
DescartesrejectedGassendi'schargethathis cogito formulawas a
syllogismin whichthe universalmajorpremiss," he who thinks,
exists" (qui cogitat est) had to be supplied.10 For to accept

Gassendi'spointwouldhavemeantplacingthe two premissesand

the conclusionof the proposedsyllogismall on the same logical
level, therebylosing sight of the indubitablenature of certain
first-personutterancesor the self-verifyingnatureof certainforms
of knowledge. Admittedly,the gain in this case would have
been to make the conclusion concerningone's own existence
appeardemonstratively
certain. Descartes,however,though he
8 Cf. M. Chastaing, " Consciousnessand Evidence", Mind, Vol. LXV
(July, 1956).
9 Second Meditation, Works,ed. Haldane and Ross, Vol. I, p. 153; also
Oeuvres,ed. Adam and Tannery(Paris,1910),Vol. VII, p. 443. Foradevelopment of this point in the textbook of Cartesian logic see the Port-Royal
Logique,ou l'art de penser(Paris, 1662, 3rd ed. 1668),pp. 380-1.
10Reply to Second Objectionand Letter to Clerselier,ed. Haldane and
Ross, Vol. II, pp. 38 and 127.

This content downloaded from 157.253.50.50 on Tue, 28 Jul 2015 18:02:01 UTC
All use subject to JSTOR Terms and Conditions

74

W. VON LEYDEN

sometimesused the word " conclusion"'1 for his cogitoformula,

always consideredthe principleit involved as a truth known
immediatelyor by itself(per se nota).'2 He consideredit as such,
becausehe claimedit to be a basic or primaryprinciple'3and
this claim he could not have maintainedwith the principle
depending,like a ratiocination,on the acceptanceof a major
premiss. But neithercould he have declaredthe principleto be
certainin the sense of being self-certifyingwithout making use
of the semanticalrulethat the sentence" I am thinking" cannot
be denied by the person utteringthe sentencewithout thereby
evincingits truth.
Hence the retentionof the first-personsingularpronounin
the formulationof Descartes'basic principle,though admittedly
not of the assumptionthat the use of this pronounindicatesa
self or mind,wouldseemnecessaryat leastso far as his immediate
purposeis concerned.
3. The question of the logical truth of the Cogito

However, Descartes obviously wished the two statements

"I am thinking" and "I am" to be not only certain but
logically certain, not only psychologicallyor semanticallyindubitablebut a priorior necessarilytrue.'4 Now, while a priori
or logicallynecessarystatementsare such that their negationis
logicallyimpossibleor self-contradictory,
my being here or the
fact that I am thinkingnow are contingentfacts and could therefore, especiallyif the personalpronoun is understooddescriptively, be deniedwithoutself-contradiction.It is true that they
could not be denied meaningfullyat any time by myself. For
Oeuvres,ed. Adam and Tannery,Vol. IX, pp. 2, 27; Vol. VIII, p. 7.
Cf. here also P. D. Huet's CensuraphilosophiaeCartesianae(Paris,
1689), Vol I, p. 11, and P.-S. Regis' reply in Re!ponseau livre (Paris, 1696),
pp. 48-53.
13 See Spinoza's Principiaphilosophiaecartesianae(Amsterdam,
1663),
PartI, Prolegomenon,for the correctemphasison the basicnatureand unique
status of the Cogitoin Descartes'doctrine.
14 " I am, I exist, is necessarilytrueeach time that I pronounceit, or
that I
mentallyconceiveit " (my italics), SecondMeditation,ed. Haldaneand Ross,
Vol. I, p. 150. See also A. J. Ayer, 7heProblemof Knowledge,pp. 45 ff.,
and in Analysis,Vol. 14, No. 2 (December,1953).
12

This content downloaded from 157.253.50.50 on Tue, 28 Jul 2015 18:02:01 UTC
All use subject to JSTOR Terms and Conditions

COGITO, ERGO SUM

75

how could such denialsactuallybe madeif they weretrue? The

factthenthattheycan be madeshowsthemto be false. None the
less, such denials on my part are absurdor inconsistent,not in
the sense that they are formallyself-contradictory,'5
but only in
that they are self-refuting,self-defeating, or self-stultifying:
though they are false, and though we can indeed deduce that
whatthey assertmustbe false,they arenot logicallyfalse. Hence
the statements" I am thinking" and " I exist" are likewisenot
logicaltruths. Both Leibnizl6and Kantl7wererightin this sense
in regardingDescartes'Cogitoas a factualor empiricaltruthonly.
Yet it is clearthat Descartespersuadedhimselfto find in it more
than a purelyfactualor semanticalcertainty.
The only alternativeinterpretation,then, for the sake of
satisfyingDescartes,would be to suggestthat the statement" I
exist" is necessarilytrue becauseit is logically implied by the
statement" I am thinking". However,to say this would be to
suggest that the meaningof the first statementmerelyrepeated
that of the second. For neitherstatementmightbe saidto follow
logicallyfrom the other,if " I exist " meant somethingdifferent
from " I am thinking" (as would particularlybe the case if
" I am thinking" is understoodas meaning" thereis a thought
now ") andif we acceptHume'snotion that no one fact or event
occurringat a givenmomentpointsto, or provesthe existenceof,
any otherfact or event in the past, present,or future. So let us
considerthe view that " I exist " does not differin meaningfrom
"I am thinking" and is thereforededuciblefrom it.
In explanationof this view it has been urged that the word
"I " in Descartes'formula,just as any personal pronoun or
propername, " is an index sign which cannot be meaningfully
used exceptto referto an existentparticular. And to say ' this
existent particularthinks, thereforeit exists' conveys no more
15
The first to have clarifiedthis point, I think, was ProfessorAyer in his
Analysis article, p. 30, and in the Problemof Knowledge,p. 46. See also
J. Passmore,PhilosophicalReasoning(London, 1962), p. 60, and Hintikka,
loc. cit., pp. 10-18.
16 NouveauxEssais, Bk. IV, Ch. ii, Sect. 1, ad fin., Ch. vii, Sect. 7.
17 Critiqueof Pure Reason,Refutation of Idealism, 2nd ed. (Konigsberg,
1787), p. 274; in Werke,ed. Hartenstein(Leipzig, 1867), Vol. III, p. 197.

This content downloaded from 157.253.50.50 on Tue, 28 Jul 2015 18:02:01 UTC
All use subject to JSTOR Terms and Conditions

76

W. VON LEYDEN

new informationin the conclusionthan does ' this is a blackcat,

thereforeit is a cat',"18 Now I believethat such an interpretation of the Cogito would have appeared unsatisfactoryto
Descartes for two reasons. In the first place, if the cogito
formulais regardedas tautologousand thereforetrivial,it could
not be takento assertor describeanything,as it wouldbe entirely
concernedwith symbols. The statement" I exist" would likewisebecomeuninformativeand consequentlytherewouldbe little
force left in the word " ergo ", or " therefore ". It follows that

the characterof the whole folmula, which Descartesclaimedas

an insight or discovery,would become modifiedto an extent
incompatiblewith his originalpurpose.
4. Is the Cogito an " analytic triviality"?

But secondly,the interpretationin questionis by no means

incontestable. It dependsfor its acceptanceon the assumption
that the word" I ", in its two usesin Descartes'formula,logically
refersto one and the samething. If it can be shown, at least on
Descartes'own premissesand perhapsalso independentlyof his
own view, that this need not be so, the chargethat his famed
cogito, ergo sum is merely an " analytic triviality ", and in

additionthe claimthat the formulais the expressionof a logically

necessarytruth,mustfall to the ground.
The questionshouldbe firstapproached,as is naturalI think,
from Descartes'own point of view and particularlyin connexion
with that stage of his argumentwherehe showsthat he can pass
from any of the basicallyunguaranteedforms of knowledgeto
the indubitablefact of his own existence. His procedurehere
dependson the notoriouslycontroversialdistinctionbetweenthe
act and the contentof awarenessand on the notion that an act of
awarenesscan be madethe contentor objectof furtherreflective
acts. He finds that his self or ego always re-emergesas active
and as the life-centreof any subsequentdoubt or thought concerningitself or anythingelse, and that thereforehis self or ego
must be real and immunefrom doubt. Considernow Descartes'
18D. J. O'Connor, John Locke (London, 1952), pp. 111-12; also J. R.
Weinberg, An Examination of Logical Positivism (London, 1936), pp. 1834.

This content downloaded from 157.253.50.50 on Tue, 28 Jul 2015 18:02:01 UTC
All use subject to JSTOR Terms and Conditions

COGITO, ERGO SUM

77

statement cogito or " I am thinking ". It expresses a case of

self-awareness:it is to the effectthat I know or am consciousof

myselfthinking. If we distinguishagainin this case betweenthe
act and the content or object of self-awareness,we might say
that it is not so muchwhen I am thinkingas whenI conceiveof
myself thinking,or make myself and my thinkingan object of
thought,that I must exist and can know that it must be true to
say " I exist ".19 Hence it might be suggestedthat the word
" I "', which occurstwice in Descartes'cogito formula,has two
differentsenses, without howeverindicatingtwo " I "s or even
two persons. The two senses may be describedas " the use as
object" (Me) and " the use as subject" (I or Ego).20
In orderto substantiatethis argumentit is essentialto show
that a case for the two senses or uses of the word " I ", which I

have indicated,can in fact be made out.

Thereare at least two ways in which it is possibleto draw
a distinctionbetweenthe two senses of " I ". One was seized
upon by existentialistsand phenomenologists,who have always
hailedDescartesas one of theirchief forerunnersand his Cogito
as a basictruth.21Onlybriefreferenceto theirdistinctionneedbe
madehere. On the one hand,they argue,thereis in everyperson
a real self or ego, somethingunique, spontaneous,and active.
This actualfountain-heador core of a man's existencecan itself
neverbecomean objectof knowledge. It eludesany attemptat
apprehension,by whichit is constantlypresupposed. Whenever
1I That Descartes' phrase " cogito " stands for the content, not the act,
of awareness clearly emerges from Thomas Reid's formulation: " In this
state of universal doubt, that which first appearedto him to be clear and
certain, was his own existence. Of this he was certain, because he was
conscious that he thought, that he reasoned,and that he doubted." (Essays
on the IntellectualPowersof Man, Edinburgh,1785,Essay II, Ch. viii, p. 129.)
Of course to say " I exist " in its turn shows that I am now thinkingabout
my own existence,therebymaking this an object of thought; but this in its
own turn againpresupposesmy existenceand thereforethe truthof " I am ".
20 Ishould point out that these phrases, as used here, differ in meaning
from Wittgenstein'suse of this terminologyin TheBlue Book (Oxford,1958),
p. 66.
21
Cf K. Jaspers, Descartes und die Philosophie (Berlin, 1937);
J.-P. Sartre, Existentialismeest un Humanisme(Paris, 1946),pp. 63-4 (Engl.
tr., London, 1948, p. 44). -

This content downloaded from 157.253.50.50 on Tue, 28 Jul 2015 18:02:01 UTC
All use subject to JSTOR Terms and Conditions

78

W. VON LEYDEN

I observeor reflecton my own being in introspection(or rather

retrospection),I apprehendmyselfonly as an objectof consciousness, a " Me ", but never recapture the original " I ", myself

as somethingreallyexisting,as the subjectof experience. On the

other hand, as the object of self-knowledgethe self becomes
somethingaltogetherdifferent. It loses all the immediacyand
independence,the vividness and activity of the real centre of
consciousness:it turns from somethingabsoluteinto something
relative. Sartrewouldin fact say that consciousnessis definable
as " 'etre-pour-soi"(being-for-itself).22By this phrase Sartre
means somethingwhichis what it is not, and is not what it is.23
This existentialistapproach and terminologyis perhaps best
summedup in Jaspers'pithysaying" I am not whatI know, and
I don't know what I am."24
5. The concept of " I " and " higher order actions"

Alternatively,one might say, followingProfessorRyle,25that

the concept of " I " is " systematicallyelusive", and that what
ordinarilypassesas self-awarenessand figuresso prominentlyin
Descartes'Cogitois adequatelydescribedas the performanceof a
"higher orderaction" upon a " lowerorderaction". A higher
oider actionfor him is one that involvesthe thoughtof another,
such as when I laugh at myselffor an awkwardnesson my part,
or furtherreflectupon my laughingat myself. One might then
arguethat we talk of self in two differentways, accordingto two
differentuses of the first-personsingularpronoun. The distinction is admittedlyobscuredas a resultof whatRylecallstheelusive
characterof "I ": we seem to be able to recognizethe basic
importanceand uniquenessof whateverit is that " I " stands
for, and yet there is a systematicand hence tantalizinglack of
22L'Etre et le Neant (Paris, 1943), Introduction and Part II. Cf. also
E. Husserl, CartesianMeditations(The Hague, 1960), pp. 25 ff.; also his
distinctionbetweencogitoand cogitatum,pp. 31 ff., and his view of the nature
of the alterationimplicit in " transcendental-phenomenological
reflection",
p. 34.
23

24
25

Op cit., p. 33.

Die GeistigeSituationderZeit (Berlin, 1932), 3rd ed., p. 147.

The Conceptof Mind (London, 1949),Ch. vi.

This content downloaded from 157.253.50.50 on Tue, 28 Jul 2015 18:02:01 UTC
All use subject to JSTOR Terms and Conditions

COGITO, ERGO SUM

79

successin apprehendingand describingit. "Like the shadowof

one's own head, it will not wait to be jumpedon. And yet it is
never very far ahead. It is too near even to be within arm's
reach.' 26

The explanationof the last point of courseis that everytime

one attemptsself-descriptionone adds a fact to be described.
None the less, if we regardthe statement" I caughtmyselfjust
beginningto dream" as expressingthe operationof a higher
orderaction upon a lower orderaction, not unlikereportingan
event, applaudinga performance,or replyingto a question,the
statementmust be of a logically differenttype from the statement " I was just beginningto dream". Hence the personal
pronounsemployedin the two sentencesare being used with a
differentlogical force.27
I suggestthat the secondof thesetwo attemptsat distinguishing betweentwo sensesof the word " I " has obviousadvantages
over the first, which is ratherobsolescentin character. On the
otherhand, I wonderwhetherRyle'sexplanationdoes fulljustice
to Descartes'originalintention. In the first place, there is an
existentialpre-conditionratherthan any kind of logical implication involved in the Cogito, and certainlythe " existentialist"
preoccupationin Descartes,as lateralso in Pascal,is undeniable.
For this reasonDescartes'point in the cogitoargumentseemsto
me to be somewhatdifferentfrom the Rylean point of view of
conceptualanalysis. For, as Ryle explains,higherorderactions
are " in one way or anotherconcernedwith" other actionsand
their descriptions "involve the oblique mention of other
actions."28He also says that the performanceof a higherorder
action " involves the thought of" another, lower order action.

He attemptsto safeguardthe meaningof the phrase" involves

the thought of" from certain misconceptions,29
using in two
passagesthe term " presuppose"30 in order to indicatein what
sense a higherorderaction " involves" the thought of another.
26
27
28

29
30

Ibid., p. 186.
Ibid., p. 190.
Ibid., p. 191.

Ibid., p. 192-3.
Ibid., pp. 191, line 35, and 192, line 7.

This content downloaded from 157.253.50.50 on Tue, 28 Jul 2015 18:02:01 UTC
All use subject to JSTOR Terms and Conditions

80

W. VON LEYDEN

Now while for Ryle the performanceof a higher order action

involves the thought of anotherin the sense of logically " presupposing" it, Descartes' interest is focused upon a rather
differentkind of implication,namely that such a higher order
actionas for instancethe doubtingof a beliefor statementimplies
or presupposes,quite apart from the thought of this belief or
statement,an act of thinkingor awareness,and this in turn(even
if it is denied) the fact that he who thinks (or denies that he
thinks)exists. The meaningand force of " ergo" in the Cogito
formula lie preciselyin the fact that my existence is a preconditionof the truthof " I am thinking" and this in turn of the
truth of " I am doubting". Whereaslaughingat myselffor an
awkwardness,just like doubting something, presupposesthe
thought of a lower orderaction, an " object" upon which it is
performed(retrospectivelyit may be said), doubtingsomething
now on Descartes'view involves,in additionto any such lower
orderaction,the presentperformanceof an act of thinkingin the
sensethat, unlessdoubtingis thinking,it is impossibleor does not
occur. Similarly,for Descartes,thinkingor beingconsciousnot
only involves somethingto think about, to reflectupon or be
awareof, i.e., a lowerorderactionupon whichthe thinkingis an
operation,but in some furthersense also the fact of one's own
existence.
But secondly, in his distinctionbetween higher and lower
order actions Ryle might perhapshave gone further. He has
been criticised3lfor assumingthat any action or situationwhich
becomesthe " object" of a higherorderactionremainsunchanged
in the process. The assumptionis perhapsjustified in cases
wherethe higherorderactionhas to do with thingsotherthan a
person'sself. However,in a case of self-commentary,
as when I
reflector reportupon a recentfeeling of sadnessor a previous
attemptat memorizingon my part, the processof becomingthe
object of a higher order action is necessarilycorrelatedwith a
change of the notion of self from subjectto object. In saying
" I am laughingat myself for beingclumsy", I am of coursenot
talkingof two selves;nor am I, as can be seenfromthe grammar,
31

I. T. Ramsey, " The Systematic Elusiveness of ' I ' ", The Philosophical

Quarterly,Vol. V (1955), pp. 196 ff.

This content downloaded from 157.253.50.50 on Tue, 28 Jul 2015 18:02:01 UTC
All use subject to JSTOR Terms and Conditions

COGITO, ERGO SUM

81

referringto two " I "s or two " me "s: I am talkingin terms of

one I and one me (or myself). While the logical status of me
(or myself) is somethingobjectivein the sense that for " me "
we could readily substitutethe name of a person, the logical
statusof I can be describedas that uniqueareain self-awareness
which is " observationally elusive ", i.e., of which Hume could

never have obtained the eageily looked-for " perception". It

follows that no self-commentarycan ever be completely or
adequatelyaccountedfor, either at the moment when it is performed (for at that stage it cannot be the concernof itself and
hence no accountwould includeit), or indeedat any subsequent
momentwhen it may have becomethe objectof anotherhigher
order action (for by then it would have lost its actualityand
spontaneityand have becomea mereobjectof reflection).
6. Summaryand Conclusion

The upshotof my argumentthenis that fromDescartes'point

of view the distinctionbetweenthe two sensesof the word " I "
in his cogito foimula can be representedin even more radical
termsthan is allowedfor by Ryle. At any ratethe formulaas a
whole cannot be considereda tautology nor indeed a logical
truth. Accordingly,its weakness is threefold. If "cogito "
is taken in the sense of " thereis thinking", " I exist" does not
follow fromit nor of coursethat thereis a self or miDd. Neither
in fact does there seem to remain anythingconclusiveor selfcertifyingin the formula. On the other hand, if the cogito
formulais understoodin its ordinarysenseof " I think,therefore
I am ", to say " I think" as also to say " I exist", though in
some sense indubitableand true, is not necessarilytrue, since a
denial of eitherstatement,though self-defeating,is not formally
self-contradictory. However, to regard the whole formula
" I think, thereforeI am" as logicallycertainand necessaryis
possible only at the price of renderingit uninformativeand
trivial-a price which Descarteswould not have been prepared
to pay and which,as I have triedto show, no one can in fact be
compelledto pay.
I mayaddherethatin my viewthe suggestiveness
of Descartes'
Cogitoandthe challengeit has presentedto subsequentgenerations

This content downloaded from 157.253.50.50 on Tue, 28 Jul 2015 18:02:01 UTC
All use subject to JSTOR Terms and Conditions

82

W. VON LEYDEN

of philosophersarise partly no doubt from the subtlety and

complexityof its implications,but also partlyfrom its inherent
difficulties. And thoughin fact we may do it justiceby explaining
its subtletyand complexity,we cannot,in any of the senseswhich
Descartesmay have had in mind, strictlyspeakingestablishor
vindicate it. As regardsmy foregoing criticism of Professor
Hintikka'sanalysis,I will summarizethis here by saying that,
thoughthe cogito insight,or one aspectof it, may be interpreted
as performatoryalongthe lineshe has indicated(withthe proviso
howeverthat in orderto be effectivesuch an interpretationmust
be free of certainambiguities),I see no reasonfor believingthat
Descarteshimselfunderstoodthe Cogito exclusivelyin this way,
nor indeedthat it mustbe so interpretedin orderto be plausible.
For it seemsto me that whateveris plausible,true or indubitable
in the Cartesianformulais so even withoutHintikka'ssuggested
analysis.

This content downloaded from 157.253.50.50 on Tue, 28 Jul 2015 18:02:01 UTC
All use subject to JSTOR Terms and Conditions