LOGICAL QUESTIONS CONCERNING

THE CONCEPT OF THE EMPIRICAL SELF

JEREMY WALKER
McGill University

About the concept of the empirical self one can make in-
quiries of two kinds. (I) One can inquire if it is a possible
concept. Here one inquires (a) whether the words 'the em-
pirical self refer to some entity that satisfies the criteria for
being classed as a concept: and (b) whether, if so, they
identify a concept that can exist as a concept for us. (a) I do
not know what are the general conditions that determine
whether or not a phrase refers to a concept. I do know, how-
ever,that one necessary condition is that there must be no
logical inconsistency attached to the phrase. Hence, if the
words 'the empirical self refer to a possible concept, it must
not be logically inconsistent to predicate empiricality and
selfhood of one & the same entity. Most of my paper discus-
ses this question, in its various details. (b) Traditionally one
has distinguished between a priori concepts and empirical
concepts. Although this distinction is not clear, one can at
least say that what distinguishes an alleged a priori concept
is that its content is not, or not wholly, given by experience,
even though it may be occasioned by experience. I am enough
of an empiricist to believe that it is a necessary truth that
there can be not a priori concepts of empirical entities.
Hence, the concept of the empirical self must be an empirical
concept. To show that it is a possible concept-for-us is there-
fore to show that the content of this concept is derivable
from experience. (It need not be ordinary, average experi-
ence). If one could establish points (a) & (b), one would
have established that the empirical self is a possible concept.

73
 (II) One can also, and independently,. inquire whether this
concept -assuming it to be possible-- has any actual appli-
cation (s). It does' not follow from the fact -if it is one-- .
that it is a possible concept, that it is a concept which actually
applies to any existing entity. Here I discover an aporia. Let
us suppose, as is logically possible, that the concept of the
empirical self, while itself a possible concept, is at the same
time empty: there are no empirical selves. Now let us consider
the logical status of this latter proposition. If we say that it
is itself necessarily true, we seem to be implying that the
concept of the empirical self is logically inconsistent, which
contradicts an assumption we have made. Hence we are
forced to construe this proposition as, if true, contingently
true, and hence the record of an empirical fact. But it
appears to me that when philosophers (since Hume) have
claimed that there is (is not) an empirical self, they have
intended their claims to have the status of philosophical
truths. Hume's observation concerning his alleged 'self' was
not intended merely as a piece of contingent autobiography.
Hence I am forced to suppose that philosophers have intended
their claims as a priori, hence necessary, truths. If so, the
normal logical process of (I) establishing the possibility of
a concept and then (II) indeperuIentlyestablishing its appli-
cation (if any) has not been conceived as itself applicable
in the case of this particular concept that we are examining.

In order to be classifiable as an 'empirical self', any entity

must satisfy two criteria: (A) It must be classifiable as
'empirical', that is, as an empirical entity, (B) It must be
classifiable as a 'self'. These form necessary, but by no means
sufficient, conditions. One further, extremely important, con-
dition can be seen if we reflect that the use of the phrase
'the empirical self' implies -as Russell showed- that there
is precisely one empirical self. This, if strictly construed,
implies that there is no more than one empirical self in the
whole world. However in normal context it clearly implies

74
only that just on empirical self is to be assigned to each
person, or each individual human being. It is not sufficient,
therefore, to show that each person possesses an empirical
self -assuming we could show that. In addition, one must
show that no person prossesses a plurality of such selves.
Ordinarily, I think, this part of the question is simply begged,
as when philosophers or psychiatrists refer to 'the self. They
do not justify the phrase.

(A) It is not easy to say what conditions must be fulfilled

in order that an entity be classifiable as 'empirical' - as an
empirical entity. Let us begin by remembering that for the
predicate empirical to have sense (content), there must be
some other contentful predicate with which it is logically
contrasted: some contentful predicate of the general form
non-empirical. This maybe a predicate such as a priori, or
transcendental. But no philosopher who regards all predicates
of the latter range as empty of sense can regard the predicate
empirical as.having sense.
I shall say that at least two conditions that must be ful-
filled in order that an entity be classifiable as 'empirical'
are (i) that it be identifiable by empirical methods, .& (ii)
that it be identifiable uiith' some empirical entity _.- even if
the latter be only itself. But I find that I cannot say what are
themselves to be identified as empirical methods. I do not
know what criteria must be fulfilled in order that we be fully
justified in speaking of empirical methods. Certainly 'em-
pirical methods' enither means nor refers to the same as does
'methods of empirical science'. If I say that tables are empir-
ical entities then I am implying that they are entities such
that they can be identified as tables by empirical methods.
But we do not identify tables as such by the methods of any
empirical science. Looking at the table from a variety of
viewpoints, touching it, rubbing it, knocking upon it, and lift-
ing it, are perfectly standard empirical methods for iden-
tifying it as a table. Further, it appears to me that we may

75
 speak of internal observation as a kind of empirical method
of identification, as, for example, when I suffer from a tooth.
ache but have not yet put my mind to locating the precise
tooth in which I am suffering. Putting my mind to this task
is an empirical method, as far as I can see. Yet concerning
the idea of 'verification by internal observation' there are
many notorious problems which Wittgenstein, among others,
has exposed: & it is well to remember that some philosophers
would not even agree with me that there does, or can, occur
any such phenomenon as 'internal observation'. Third, 'em-
pirical' in 'empirical method' does not refer to 'perceptual',
if perceptual methods are supposed to be contrasted with in-
tellectual methods. That an. entity can be identified only by
what may be loosely called intellectual methods does not
prove that it is not an empirical entity - does not prove, for
example, that it is a transcendental entity. Hence, if it were
a fact that the self could be identified only by intellectual
methods (thinking, reasoning, speculating, etc.}, that would
not have any tendency to prove that the self so identified was
a transcendental self. The justification for this claim is that a
cardinal technique employed in constructing the empirical
sciences is the use of the intellect in the formation of theo-
retical concepts and the construction of theories. Such sciences
are not.Iess empirical since they are also theoretical. I must
admit, though, that the theories & theoretical constructs of the
empirical sciences are empirically validated,· in some sense,
by means of their logical connexions with concepts immedi-
ately descriptive of actual phenomena of observation & mea-
surement.·Such theoretical.concepts, then, maybe called also
empirical only insofar as they play an essential role in the
explanation of such phenomena. I do not, therefore, argue
that because 'the empirical self' is a. 'theoretical concept'
- ..if it is--· it is therefore also an empirical concept. To show
this, it has to be shown in addition how this concept functions
as an essential part of some general schema of explanation
of certain phenomena. I believe this intention underlies the

76
employment of this, and related, concepts by most psychol-
ogists, psychiatrists, and certain sociologists.

(B) What criteria must an entityfulfil in order to be clas-

sifiable as a self? I should begin by observing that the pecu-
liar difficulty one meets in answering this question is due to
the fact that, in a certain sense, there is no concept of the
self. What I mean by this is that the concept of the self is a
philosophers' concept: an invented concept. There are pos-
sibly as many different concepts of the self as there are
philosophers who have philosophized about the self. In a
certain sense, therefore, the answer to the question I have
posed can only be arbitrary. Nevertheless, I think it is pos-
sible to identify four criteria which must be fulfilled by any
concept of the self, if that concept is to be identifiable as
a concept of the self. However, if someone replies that I am
doing no more than specify my own concept of the self, in
claiming these four criteria as essential, I do not clearly
yet know how to answer him.
In order to be a self an entity must (1 )be continuous over
a certain period of time, (2) possess an individual identity
-. be individuatable-, (3) suffer change, or be capable of
suffering change, of a certain type, over a period of time,
& (4) possess consciousness .at specifiable times and over
certain periods of time. Notice the reference to time. My
criteria clearly imply that there can be no such entity asa
non-temporal self, even abstracting from the question of the
logical possibility of such a concept. Notice also that in my
definitions 1 abstract from the question of the mortality of
the self or its possible survival of death. (I will, however,
later say some things that bear on this question.') 1 may be
accused, thus, of begging the question of the possibility of the
existence of a personal deity - a god with self-hood. This
accusation is unfounded. My criteria permit the existence of
either (a) a temporal, & extremely long-lived, deity, or (b)
a temporal, & infinitely long-lived, deity. It is only when the

77
concept ofetemity is not construed as the concept of infinite
time that my criteria rule out the possibility of the existence
of a deity who is both eternal and possesses self-hood: but
despite the fact that some thinkers have employed a 'non-
temporal' concept of eternity (e.g. Kierkegaard), I cannot
find that I understand them.
(1) To be a self, an entity must necessarily be continuous
in its existence (persistence) over a certain period of time.
That is to say, no entity whose existence in time is not con-
tinuous, for example instantaneous, can possibly be a~'self.
This is because the concept of the self is the concept, inter
alia, of a principle for assigning different phenomena to one
'focus', or, more perspicuously, a principle for assigning
different events to one 'focus'. It is a principle for collecting
events: and, perhaps, also a principle for individuating events.
I am here speaking of its. logical function. When I speak of
the concept of the-self as the concept of a certain principle,
I wish to imply that this principle is to be taken as non-
phenomenalistic: the self is 'more' than that bundle of events
to which it serves as principle of unity or collection. It can-
not be defined extensionally, by reference to these events
alone. (This claim does not imply, however, that principle
which is the self in its logical aspect is a noumenal, or trans-
cendental, existent. It may, for example, be an entity which
can be inferred inductively from those events whose principle
it is - as it perhaps appears in the work of Ryle.) No in-
stantaneous entity, then, can be a 'self, since where an entity
is instantaneous it appears to me impossible to distinguish
between its essential & accidental attributes: or, to express
the point otherwise, impossible to distinguish a principle of
unifying assignation from the events so unifyingly assigned.
And in the case of the self, I am implying, these distinctions
must be capable of being made. (I am aware of a deep insuf-
ficiency in these arguments: I do not knowhow to remedy it.)
To say this is, I think, something that can profitably be
expressed by saying that, in my conception, the self must

78
be understood as having a substantial role, in some sense of
'substance'. For we cannot speak of substance without speak-
ing of essential attributes, & of the distinction between es-
sential & accidental attributes. (It is at this point that the
contemporary notions of roles & role-playing, in their rela-
tions to the self, become relevant.)
To claim that the self must be continuous over a certain
period of time is, however, not perspicuous. For there are
very many varying concepts of continuity over time, or of
temporal continuity. Now we may admit that the mode of
continuity of the self, whatever it is, cannot so to speak
transcend the mode of continuity of time itself. (I cannot
argue for this: it seems clearly true to me.] Thus, for exam-
ple, if time is densely ordered, the self cannot possess Dede-
kindian continuity: if time is merely discrete, the self cannot
possess density, and so on. On the other hand, it appears to
me that, whatever the mode of continuity of time itself, the
self cannot possess a mode of continuity which is, so to
speak, inferior to that of time. In fact time & the self must
possess exactly the same sort of continuity. Although this
appears to me to be true, it does not solve the problem with
which we are faced, since it is notorious that there is no one
self-evidently true model for time. That is to say, it does not
appear possible (yet) to decide what kind of logical series
is instanced by time. Hence we cannot say what kind of con-
tinuity over time the self possesses (if there are any selves).
Nevertheless, I want to express my doubt whether any 'punc-
tiform' concept of continuity -even that of Dedekindian
continuity, which is punctiform, in the sense that it is in-
stanced by the series of the real numbers- can possibly
fit the concept of the self. I want to express, in as tentative a
manner as I know how, my feeling that the concept of the
self is the concept of an entity which possesses some kind of
'absolute', non-punctiform, continuity. I must again admit
that I do not know clearly what these words mean.
Here I must introduce a further consideration whose bear-

79
 ing puzzles me. It is a notorious feature of human existence
that our conscious life has periods of intermission: of sleep
and unconsciousness, for example. One might therefore ar-
gue, if one made the strong claim that the self as such neces-
sarily possesses consciousness, that during such periods the
self ceases to exist. Indeed, on that assumption, one must so
argue. But it appears to me unnatural to say that the self
ceases to exist during periods of sleep & unconsciousness, for
example. On the other hand, if someone wished to say that
it is unnatural to say that during these periods the self still at
that time exists, I do not know what to say to him. In fact
I am forced to suppose that the question. "When does the
self exist, during life?", is a nonsense question. But I do not
see why.
I find at this point an even odder aporia in my mind. If we
consider for a moment the hypothesis that the self possesses
a recurrent mode of temporal existence -that although there
are periods of intermission in its existence, still the same self
turns up again next morning, so to speak-, I find myself
inclining to say that it is as if between one 'occurrence' of
the self & another no time had elapsed. It is as if, although the
self really ceased to exist during the intermission between
the two periods of its occurrence, nevertheless during that
intermission it continued to exist in some non-temporal man-
ner. There may be at least three considerations which drive
me to say such a peculiar thing. (i) Generally, I find it im-
possible to admit that an entity can exist in a genuinely dis-
continuous manner. It may be said that there is no problem:
the thing, x, goes out of existence at one moment of time t, and
reappears in existence at another later moment of time t'.
There is admittedly the problem of providing satisfactory
criteria for identifying the thing which reappeared in exis-
tence at t' with the thing that disappeared at t. However, I
cannot find that I can make sense of the ontological picture
drawn in this manner. (ii) There is the phenomenon of
memory. This is well known to be central to the problem

80
of the self, but in this connexion the phenomenon that im-
presses me is our normal capacity, on waking or recovering
from unconsciousness, for example, to remember the events
that preceded that interruption of our consciousness. This is
not solved by simply remarking that there are persistent
conditions of the brain & nervous system which in some un-
known manner contain our memories. The problem is how
the consciousness of a certain past event can recur at a later
time, after a period of the total interruption of the conscious-
ness, & I must confess this problem puzzles me totally. Mem-
ory relates to my aporia in another way, too. If one analyses
the concept of memory as the concept of a mode of conscious-
ness of some past event, then granted that past events are
genuinely past, i.e, no longer exist, I find myself puzzled to
conceive how a consciousness existing at a certain time t' can
embrace an event that occurred at some other time t, A fur-
ther difficulty is that in remembering something, it appears
to me, we in a certain sense experience that past event as
present. It is as if time did not exist for memory. (iii) Leav-
ing these aporiai aside, I find that the third -and relatively
simple--, consideration that influences me is that selves must
be ascribed to persons, and while consciousness certainly
suffers lapses, the person who is the bearer of this conscious-
ness certainly continues to exist during hia/her periods of
unconsciousness.
I now leave the subject of continuity, in order to tum to
the concept of the identity of the self.
(2) The empirical self must possess an identity, and more
particularly an individual identity. (Atoms of a certain ele-
ment each possess identity without, however, possessing in-
dividuality.) Therefore it is necessary that there be criteria
for individuating selves. These criteria playa dual role. (i)
They serve to individuate each self as the particular self it is.
(ii) They serve, at the same time, to identify it distinguish-
ingly from all other selves.
I here put forward a thesis: that the self is individuated

81
by its acts. One might think that it is better to say that the
self is individuated by the acts of the person - the person
to whom the self belongs. I do not accept this as the equivalent
of my version, however.For it seemsto me that there can be a
person & the acts of that person even where there is no self
that can be adscribed to that person - for example, in certain
disordered psychological conditions. It is only when a person
has a self that hisjher acts can serve to identify hisjher self.
I understand 'acts' in a rather strange way. I do not only
include the overt bodily behavior of the person in question, in-
cluding hia/her speech. I wish to include also that person's
thoughts, wishes, unconscious wishes, dreams, parapraxes,
etc. All of these are to be counted as 'acts of the self' by my
criterion. Notice that it is from its identification by its acts
that the self derives its empirical content - it is from the
fact (if it is a fact) that the concept of the self is the concept
of an entity for whom the identifying criteria are provided by
empirical 'acts' that it follows that it is itself necessarily the
concept of an empirical entity.
In including thoughts, wishes, unconscious wishes, etc.,
under the category of 'acts', I am merely supposing that to
know someone is to know what his/her outer & inner life is.
This is what we normally do suppose. It is important, how-
ever, to avoid a mistake in interpreting this thesis. The mis-
take to which I am referring is that of assuming that the
relationship of outer to inner is identical with the relation-
ship of the empirical to the transcendental. That this is a
mistake has, I think, been sufficiently shown by the work
of Wittgenstein and Ryle, so I need not go further into it
here. Now I earlier claimed a non-phenomenalistic account
of the self. We have, therefore, to resist a phenomenalistic
account of the relation of the self to its acts. What is the
proper relationship? My view is extremely unsettled, but
I wish to propose that the relation of the self to its acts is
an instance of the general relation between Essence and Ap-
pearance. Essence and Appearance are definitely not here to

82
be construed in a Platonistic sense: of that I am clear. Wheth-
er the relationship is to he construed in a Hegelian, or an
Aristotelian, sense I am not in the least clear. I earlier spoke
of the self as a 'substance' & using this term I was conscious,
and wished you to be conscious, of the Aristotelian rever-
berations of the term. I do not claim to understand what
Aristotle means by essence or substance (to ti en einai,
ousia), but his account in the Metaphysics impresses me as
bearing the marks of convincing truth. Again, although I
find myself in almost complete darkness when reading Hegel's
account of the relation between inner and outer (in the
Phenomenology), it impresses me as saying something which,
if could understand it, would be convincinglytrue.
There is one difficulty -a problem of practice as well as
of theory- which I should like to mention before leaving
this topic. In proposing that it is in the acts of the self that
the self is identified, it appears to me to he necessary to be
able to distinguish between those acts that are essential at-
tributes of the self, so to speak -that is to .say, if we can
speak of acts as attributes-, and those that are its accidental
attributes. This is a distinction that finds some place in every-
day life. It also plays an important rolein our thinking about
ourselves and our everyday behavior. But I do not know
how we can distinguish non-circularly between the essential
& the accidental attributes, or acts, of the self.
(3) The empirical self must suffer change, or be capable
of suffering change, over a given period of time. This implies
three things. (a) The self must be allowed to grow over time.
(b) It must, perhaps, be allowed to grow into a self from a
condition of non-self (pre-self, rudimentary self). (c) It must
he permitted to develop over time. There seem to me two
reasons for saying this. The first is that the ordinary concept
of the self is of something which does typically develop over
the course of a human life. The second is an a priori argu-
ment, over whose force I am entirely doubtful. The proof is
from the consideration of the concept of an essentially static

'83
self. Allowing, what I have not yet argued but will come to
bellow, that the self is essentially self-conscious, this is the
concept of an 'entity which possesses self-consciousnessbut is,
at the same time, without the possibility of change due to
self-consciousness. That self-consciousness allows the pos-
sibility of change appears to me to be more than a contingent
truth. I find the idea of a being who possessed self-conscious-
ness, but whose self-consciousnesswas powerless to act as a
cause, or occasion, for self-change, to be incomprehensible.
In order to discuss the question of the self's development,
we must ask what criteria must be fulfilled, in general, for us
to be able to say of any entity X that it has developed into
entity Y. For I take it that all development is of something
into something -- whether or not it is something else, The
point ofthis is that we must be able to distinguish something's
developing into something from its having changed into that
thing. Set out in logical schema, to say that X has changed
into Y appears to me to imply (i) that X no longer exists,
and (ii) that Y is not X - Y is different from, other than,
X - Y cannot be identified as X. If change is other than
development, these two necessary conditions of change will
not, in general, be necessary conditions of development. In-
deed, it seems plausible that we reach the concept of develop-
ment precisely by relaxing these two conditions. I say 'relax-
ing', rather than 'abandoning', since I am not sure that we
can quite say, in interpreting the notion of development, that
(i) X still exists or that (ii) Y = X. For example, it does
not seem to me that, considering the example of the acorn
and the oak-tree, we can quite say (i) the acorn still exists
'in' the tree, or (ii) the oak-tree is, in a sense, the acorn.
Nevertheless, it appears to me equally crass to say that there
has been, so to speak, a complete lacuna of existence between
the acorn and the tree, or that there is no principle leading
us to assimilate the acorn and the tree. My present view is
that we must say that X and Yare the same thing, but with-
out converting this into the proposition: X = Y. In fact, there

84
 is no single variable term that we may substitute for the
expression 'thing', nor for the phrase 'the same thing'. That
proposition cannot be schemasized logically. And we have to
add, in order to make our proposition more perspicuous, that
X and Yare the same thing at different stages of its develop-
ment - 'its' development. Hence, it appears that the defini-
tion of the concept of development by means of a schematic
criterion is liable to be circular, as is the implicit definition
proposed above. More generally, it seems to me that formal
logic does not, so to speak, quite fit such entities - develop-
ing entities. I must admit I have not pursued this matter
deeply, and I am not aware of any work there may recently
have been upon this problem. The answer is not to be found,
I am sure, in trying to construct formal logics that reject the
Law of. Contradiction, or the Law of the Excluded Middle,
since such attempts merely appear to me to be deeply con-
fused.
Coming down to earth, we must next' examine the criteria
for the notion of 'the development of the self'. My only sug-
gestion here is that we may identify these with the criteria
for the development of the personality'. I do not know what
the latter are. Does it follow that self = personality? I do
not believe so: for I think that the relationship of the self
to the acts performed by the person in question is not identical
with the relationship of the personality of that person to
his/her acts. But here is another point at which I have not
pushed my analysis further.
Let us say three things in general about the criteria for
self-development. (1) Some features can appear, others can
vanish. (2) It follows, if we are to be able to speak of iden-
tity, that there must therefore be an identifiable continuity
of features. (3) And there must, in addition, be a principle
for linking new features to the old self. These are, in a way,
quite general conditions for the application of the notion of
development to anything whatever. But in respect of the con-
cept of the self, I suggest that we should not do better, than

85
adopt, as a guiding light, the. principle of potentiality & ac-
tualisation of potentialities. Whether this is, so to speak, a
natural actualisation, or a Hegelian actualisation involving
the actualisation ofself-consciousness, I am not certain. Nev.
ertheless, I believe that the Aristotelian metaphysics still
provides us with certain ideas that are useful, and perhaps
indispensable, in these discussions.
If we identify the growth of a self from a non-self (a pre-
self or rudimentary self) with its growth from a potential self,
this maybe a second place in our puzzles where such meta-
physics is of assistance.
(4 ) That a self must possess consciousness that is tem-
poral in both the sense that it exists at certain times and the
sense that it exists over certain periods is, probably, the most
important and most difficult of all these assertion concerning
the concept of the empirical self.
I begin by making two assertions. (a) Unless a being pos-
sesses consciousness it cannot be conceived to possess a 'self'
- at least an empirical self. (One might wish to hold that
even a stone partook in the transcendental Self of the uni-
verse.) The proof of this proposition appears to me to be
very short & simple, but I am correspondingly doubtful of its
validity, as of its power to convince. It is that if we consider
the concept of a nonconsciousbeing, we cannot assign to such
a being a self, since we cannot more fundamentally assign
to it an essential principle of individuality unless it is that
essential principle of individuality which applies to organic
beings. To sub-organic beings it appears to me that we can-
not assign any such essential individuating principle. In fact,
at this point it seems to me that we can assign substance, in
a useful sense, only. to such beings as are either organic or
super-organic. However only the substantial principle of in.
dividuality of a super-organic -that is, an essentially con-
scious- being can be conceived under the category of self.
It is to natural organic beings in particular that the Aris-
totelian category of substance applies: and I do not think

86
that we can find in Aristotle any concept like the concept
of the empirical self: not even the idea of psyche plays this
role satisfactorily. (b) In the second place, I wish to assert
that unless a being possesses self-consciousness,or more prop-
erly self-awareness, it cannot be conceived as possessing a
self. Hence, although I am willing to grant that animals may
have personalities, I will not allow them selves, except insofar
as we may justifiably ascribe self-awareness to them.
Before proceeding, I wish to insert some reflexions here on
the concept of substance in relation to conscious beings and
organic but non-consciousbeings. In the latter case substan-
tiality may be assigned to a being on the grounds of an
evident principle of unity's being operative in the various
consecutive stages of that being's growth over time. This
statement is completely formal, and does not begin to specify
what kind of evident principle of unity is involved, or by
what criteria it is to be identified in reference to various
species of organic beings. What is, however, indubitably
true, I think, is that here the principle of unity, whatever
it is, is necessarily completely manifested in the various
phenomena of the stages of growth. Hence I find that here
there are no grounds for rejecting a phenomenalist interpre-
tation of the 'essence' of such types of being. Contrast the
phenomenon of conscious being. Here consciousness makes
all the difference. For it allows a distinction -a gap- be-
tween the substantial principle and the phenomena. (Consid-
er, for example, the familiar gaps between intention and
action, thought and speech, the phenomena or hypocrisy,
deceit, and self-deceit, and so on.) There are, therefore, in
this case grounds for rejecting the phenomenalist analysis as
giving an insufficient account of these familiar cases.
I have asserted that any being that possesses a self neces-
sarily possesses self-consciousness.Notice that it is entirely
illegitimate to argue from this proposition that the self is the
consciousness of this being, or that the self is its self-con-
sciousness. I do not therefore assert either of these last pro-

87
positions.
If possession of a self necessitates possession of self-con-
sciousness, does possession of self-consciousness necessitate
possession of a self? It might be thought on inspection that
this question resolved itself through an evident tautology. But
I wish to reject the converse assertion, since it appears to me
that the phenomena of the consciousnesses of children and
schizophrenics (etc.) afford good grounds for saying that a
being can possess self-awarness, of a kind, without being a
genuine self. What is guaranteed, however, by possession of
any mode of self-consciousneesis that one is a being of a
kind that normally has a self, and hence that, in a certain
sense, one is also already potentially a self. The sense of
'potentially' here is unusual: it does not necessarily refer
forwards to the future, since the schizophrenicmay be beyond
the point of no return, so to speak. He is not now capable
of ever becoming a genuine self. It is still true, however, that
it is not impossible that he might (now) have had a self, and
this is the sense of potentiality I mean.
Above I hastily rejected the idea that the self is to be
identified with consciousness. I want to end by giving an
argument to this effect. In general, I have earlier claimed,
criteria for describing and identifying the empirical self are
taken from behavior, in a broad sense which includes inten-
tional performance. I do not accept behaviorism in any of
its many logical forms: but I merely wish to point out that
even if one accepts behaviorism in its weaker form ('There is
consciousness, but its criteria are behavioral') one will still
not be able to identify the self and consciousness.For if one
tried to make this identification, there would remain no
defence against phenomenalism, and no possibility of picking
out from the mass of behavior such behavior as embodies
consciousness, and is thus the 'essential' behavior of the self.
Such selection would always necessarily remain quite ar-
bitrary.
Let me end by putting my four criteria into a somewhat

88
 new light. I do this by recalling that I have earlier claimed
that the concept of the empirical self is (formally) the con-
cept of a principle for assigning certain phenomena -a
certain selection of the relevant phenomena- to one 'focus'.
(1) The question of the continuity of the self now becomes
the question of the 'continuity' of a certain type of unifying,
explanatory, principle of personal behavior. If we take this
line, it is clear why I was so puzzled about the nature of the
'existence' of the self in its relation to time. It is well to be
puzzled about the 'existence' of principles of explanation of
temporal phenomena in their relation to time. (2) The ques-
tion of the identity of the self becomes the question of prov-
iding criteria for identifying such a focussing, explanatory,
principle of behavior. This illuminates my claim that the self
is identified by its acts, and gives a new meaning to the view
that the relation of the self to its acts is an instance --in
some sense- of that of Essence to Appearance. (3) To say
that the empirical self is essentially capable of temporal
development now becomes the rather different claim that a
certain principle of explanation is essentially capable of such
development. Given an entity -a person- otherwise iden-
tifiable independently of identifying his/her 'self - whose
behavior itself changes & develops over time in some relative-
ly coherent manner, we may well try saying that the funda-
mental cohesive principle of his/her behavior itself needs
adaptation to fit the new circumstances. (4) If the self is a
principle for assigning & explaining the behavior of an es-
sentially conscious being, it becomes immediately clear why
there is a temptation to identify the self with this 'conscious-
ness' itself, since in many cases the consciousnessdoes serve
an explanatory role in relation to the bodily phenomena. But
not in all cases.

89
RESUMEN

Con relacion al concepto empirico del yo hay dos tipos de pregun-

tas posihlese (I) Se pregunta uno si es un concepto posible: (a) si
las palabras 'el yo empirico' se refieren a alguna entidad que satis-
faga los criterios para ser clasificada como 'un concepto', y (b) si
esto es asi, si elIas identifican un concepto para nosotros.
(a) No se si haya condiciones generales que determinen 0 no si
una frase se refiere a un concepto, Lo que si se es que una con-
dicion necesaria para la posibilidad de un concepto es que no sea
logicamente contradictorio predicar de una entidad la empirisidad
y la mismidad.
(b) Soy 10 bastante empirista para creer que es una verdad ne-
cesaria el que haya conceptos no a priori de entidades empiricas,
esto es, conceptos que sean derivados a partir de la experiencia.
(II) Uno se puede preguntar si este concepto -asumiendo que
es posible-s- tiene alguna aplicacion actual. Por reducci6n al absurdo
suponemos que no hay yos empiricos, y suponemos que esto es una
verdad necesaria, ello implica que el concepto del yo empirico es
contradictorio, 10 cual contradice nuestra eupcsicien inicial.
Para que cualquier entidad sea clasificada como 'yo empirico' debe
satisfacer dos criterios: (A) Debe ser clasificable como una entidad
empirica y (B) debe ser clasificahle como 'yo'. Estas constituyen
solo las condiciones necesarias, otra condicion muy Importante es
ver si reflejamos que el uso de la frase 'el yo empirico' implica
--como Russell mostr6-- que hay precisamente 8010 un yo empirico.
En los contextos normales esto implica que a cada persona 0 ser
humano individual se le asigna solo un yo empirico. Para ello no es
suficiente mostrar que cada persona posee solo un yo empirico --su-
poniendo que se pudiese mostrar-, ademas se debe mostrar que
ninguna persona posee una pluralidad de tales yos, Considero que al
menos dos condiciones se deben satisfacer para que una entidad
sea clasificahle como 'empirica': (i) que sea identificable por me·
todos empiricos y (ii) que sea identificabIe con otra entidad empi-
rica - aun cuando esta ultima sea solo si misma,
(B) Para que una entidad sea clasificable como yo debe (1) ser
continua durante cierto periodo de tiempo; (2) debe poseer una
entidad individual --ser individualizable-c-; (3) debe sufrir cam-
bio, 0 ser capaz de sufrir cierto tipo de cambio durante un periodo
de tiempo, y (4) debe poseer conciencia en tiempos especificahles

90
 y durante ciertos periodos.
(1) La existeneia de esa entidad debe ser continua durante un
cierto periodo. La funcien IOgica del concepto de si mismo con-
siste en ser un concepto inter alia, es un principio por el cual Be
asignan diferentes eventos a un mismo 'focus', es un principio para
coleceionar eventos y, tal vez, es tamhien un prineipio para indivi-
dualizarlos.
(2) EI yo empirico debe poseer una entidad individual. Los cri-
terios para individualizar al yo juegan un doble papel: (0 sirven
para individualizar cada yo como el particular yo que es y (ii)
sirven al mismo tiempo para identificarlo distinguiendolo de todos
los demas yos.
Sostengo la tesis de que el concepto del yo es el concepto de
una entidad cuyos criterios de identificacion se basan en sus actos
empiricos. La entidad referida por tal concepto se individualiza por
sus actos, La relacion del yo con sus actos es un caso de la relacion
general entre Ia esencia y la apariencia.
(3) EI que el yo empirico sufra cambio 0 sea capaz de sufrir
cambio durante un periodo dado de tiempo implica tres cosas: a)
Que al yo Ie debe ser permitido crecer durante el tiempo, b) Se Ie
debe permitir, tal vez, crecer dentro de un yo partiendo de la con-
dicion de un no yo. c) Se Ie debe permitir evolucionar en el tiempo.
En relacion a los criterios para el desenvolvimiento del yo hay que
mencionar tres: 1) Algunos caracteres pueden aparecer y otros des-
aparecer. 2) Se deduce que podemos hablar de identidad, esto es,
que debe haber continuidad identificable de caracteres. 3) Debe
haber un principio que ligue caracteres nuevos al antiguo yo. Con
respecto a 10 anterior adoptamos el principio de potencialidad y
aetualizacion de poteneialidades de Aristoteles que nos es 'litH.
(4 ) La afirmacion mas importante en relaci6n al yo es que el yo
debe poseer conciencia, la cual es temporal en dos sentidos: la con-
ciencia existe en ciertos tiempos y existe durante ciertos period os,
Solamente cuando un ser posee conciencia puede ser concebido como
un yo empirico. Ya que solo en tanto un ser tiene conciencia se Ie
puede asignar un principio de individuacion, Un ser que posee
mismidad posee necesariamente conciencia de si mismo. Considero
que el concepto del yo empirico es formalmente el concepto de un
principio de asignaeion de ciertos fenomenos --de cierta seleccion
de fenomenos relevantes- a un foco.
(1) La pregunta sobre la continuidad del yo se torna en la pre-
gunta sobre la 'continuidad' de cierto tipo de principio unificador y
explicador del comportamiento personal.
(2) La pregunta por la identidad del yo se torna en la pregunta

91
para proveer los criterios para 1a identificacien de tal principio
explicatorio y unificador.
(3) La afirmaci6n sobre la evolucion temporal del yo ahora se
torna en 1a afirmaci6n de que un cierto principio de explicacidn es
esencialmente capaz de tal evolucion.
(4 ) Si el concepto del yo es un principio por· el cual se asigna
y se explica el comportamiento de un ser esencialmente consciente,
se torna claro el por que de la tentacion de identificar el concepto
del yo con la conciencia, ya que en muchos casos (pero no en to-
dos) 1a conciencia nos sirve para explicar los fenomenoe corporales.

92