268

subjectively, birth is not at all experienced as a separation from
the mother, since the mother, in the rôle of object, is entirely
unknown to the completely narcissistic fœtus. Another con-
sideration that would apply is that affective reactions to a sepa-
ration are known to us, and that we experience them as grief
and mourning, not as anxiety. We recall, to be sure, that in
our discussion of mourning we were also unable to understand
why mourning is so painful.

VIII

It is time to take stock. What we are seeking, it is apparent,
is an insight which shall reveal the nature of anxiety, an
“either–or” which shall distinguish truth from error in regard
to it. But this is difficult of attainment; anxiety is not a simple
thing to grasp. Thus far we have arrived at nothing but con-
tradictions, from among which no unbiased choice was possible.
I now propose to order it otherwise; we will bring together
in unprejudiced manner everything that can be said about
anxiety, while renouncing at the same time the expectation of
achieving an immediate synthesis of the problem.

Anxiety, then, is in the first place something felt. We call
it an affective state, although we are equally ignorant of what
an affect is. As a feeling it is of most obviously unpleasurable
character, but this is not by any means a complete description
of its quality; not every state of unpleasure (Unlust) may we
call anxiety. There are other feelings of unpleasurable char-
acter (mental tension, sorrow, grief), and anxiety must have
other characteristics besides this quality of unpleasure. Shall
we ever succeed, one cannot help asking, in understanding the
differences between these various affects of unpleasure?

Of the feeling of anxiety we can after all learn something.
Its character of unpleasure seems to possess a particular note
of its own–a thing difficult to demonstrate but none the less
probable, nor would it be at all surprising if it were so. But
in addition to this special characteristic so difficult to define, we
perceive more definite physical sensations, which we refer to

269

specific organs, as accompanying anxiety. Since the physiology
of anxiety does not interest us here, it will suffice to draw atten-
tion to specific examples of these sensations, such as those refer-
able to the respiratory organs and the heart, which are the most
common and the most definite of them. They are evidence
that motor innervations, that is, efferent processes, take part in
the total phenomenon of anxiety. The analysis of the anxiety
state gives us, then, as its attributes: (1) a specific unpleasurable
quality, (2) efferent or discharge phenomena, and (3) the per-
ception of these.

The second and third of the foregoing supply in themselves
a distinction from similar affective states, such as for example
grief and sorrow, for of these latter, motor manifestations do
not form an integral part; when such are present, they are defi-
nitely distinguishable as not constituting essential constituents
of the total phenomenon but consequences of or reactions to
the emotional state in question. Anxiety, therefore, is a specific
state of unpleasure accompanied by motor discharge along
definite pathways. In accordance with our general outlook,
we shall believe that an increase of excitation underlies
anxiety, an increase which on the one hand is responsible for
its unpleasurable character and on the other is relieved through
the discharge referred to. This purely physiological summary
will scarcely satisfy us, however; we are tempted to presume
that there is an historical element present which binds the
afferent and the efferent components of anxiety firmly together;
in other words, that the anxiety state is the reproduction of an
experience which contains within itself the requisite conditions
for the increase in stimulation just mentioned, and for its
discharge via given pathways; and it is in virtue of this, there-
fore, that the unpleasure element in anxiety acquires its specific
character. As the prototypic experience of such a sort, we
think, in the case of the human being, of birth, and on this
account we are inclined to see in the anxiety state a reproduc-
tion of the trauma of birth.

In doing so we have claimed nothing which would assign to
anxiety an exceptional position among affective states. For

270

we hold that other affects as well are reproductions of past
experiences of a character vital to the organism, experiences
possibly even antedating the individual; and we draw a com-
parison between these, as universal, specific, congenital hysteri-
cal attacks, and the seizures of the hysterical neurosis, later
and individually acquired, the genesis and significance of which
as memory symbols have been made clearly manifest by analysis.
It would of course be most desirable to be able to demonstrate
the validity of this conception for a number of other affects,
but at the present time we are far from being in a position to
do this.

The tracing back of anxiety to the birth experience needs
justification in the face of certain obvious objections. Anxiety
is a reaction characteristic of probably all organisms, certainly
of all the higher ones, but birth is experienced only by mam-
mals, and it is open to question whether for all of these birth
has a traumatic significance. There is, therefore, such a thing
as anxiety without a prototype in birth. But this objection
takes us from psychology into biology. Precisely because
anxiety, as a reaction to situations of danger, has a biologically
indispensable function to fulfil, it may have been contrived in
different organisms in different ways. We do not know, more-
over, whether in creatures at a further remove from man
anxiety has the same content, afferently and efferently, as in
the human being. All this does not prevent it from being the
case, therefore, that anxiety, in the human being, takes the
birth process as its prototype.

If this is the structure and origin of anxiety, the question
then arises: What is its function? On what occasions is it
reproduced? The answer seems obvious and inescapable.
Anxiety arose as a response to a situation of danger; it will be
regularly reproduced thenceforward whenever such a situation
recurs.

But there is more than this to be said. The motor impulses
accompanying the original anxiety state had probably as much
meaning and utility as the muscular movements of the initial
hysterical attack. If one would explain the hysterical seizure,

271

one needs only, indeed, to look for the situation in which the
movements in question were part of the behavior appropriate
to that situation. Thus, during birth, it is probable that the
directing of nerve impulses to the organs of respiration has
made preparation in advance for the functioning of the lungs,
the acceleration of the heart beat tended to counteract the
accumulation of toxic substances in the blood. This teleology
of function is of course absent from the subsequent reproduc-
tion of the anxiety state as affect, just as it is also lacking in
the recurrent hysterical seizure. If therefore the individual
encounters a danger situation new to him, his responding with
the anxiety that constitutes the reaction to an earlier danger,
instead of with the reaction appropriate to the present one, may
easily become inexpedient. The suitability of the reaction
reappears, however, if the danger is perceived as imminent and
forewarning of it given through the outbreak of anxiety. The
anxiety can then be immediately replaced by more appropriate
measures for dealing with the danger. Two possibilities with
regard to the appearance of anxiety, therefore, may at once be
distinguished: the one, inappropriate and inexpedient, in
response to a new situation of danger; the other, a useful one,
as a means of giving warning of and averting such a situation.
But what is a “danger”? In the act of birth there is an
objective danger to the preservation of life; we know what that
means in the reality sense. But psychologically it has no mean-
ing at all. The danger attending birth has still no psychic
content. For certainly we cannot imagine as existing in the
fœtus anything which in the least approaches any sort of
knowledge of the possibility of death as an outcome. The
fœtus can be aware of nothing beyond a gross disturbance in
the economy of its narcissistic libido. Large amounts of
excitation press upon it, giving rise to novel sensations of
unpleasure; numerous organs enforce increased cathexes in
their behalf, as it were a prelude to the object cathexis soon
to be initiated; what is there in all this that can be regarded as
bearing the stamp of a “danger situation”?
Unfortunately, we know far too little of the mental make-up

272

of the newborn to be able to answer such a question directly.
I cannot even vouch for the usefulness of the description I have
just given. It is easy to say that the newborn infant will repeat
the affect of anxiety in every situation which reminds it of the
birth situation. The real question, however, is by what and
of what it is reminded.

There is left us hardly any other course to pursue than to
study the occasions on which the infant or the slightly older
child gives evidence of a readiness to develop anxiety. In his
book, The Trauma of Birth, Rank has made a very vigorous
attempt to demonstrate a relationship between the earliest
phobias of the child and the impression which the birth experi-
ence has made upon it, but I cannot consider the attempt a
very happy one. Two criticisms can be brought against it,
of which the first is that it makes the assumption that in the
process of birth the child has been the recipient of sense impres-
sions, particularly visual ones, the renewal of which may evoke
the memory of the birth trauma and therewith a reaction of
anxiety. This assumption is entirely unproved and very
improbable; it is not credible that the child has preserved any
other than tactile and general sensations from the act of birth.
If, then, the child later shows a fear of small animals which
disappear into holes or come out of them, Rank explains this
reaction as its perception of an analogy which, however, would
not strike the child. Secondly, in appraising these later anxiety
situations Rank holds responsible the memory either of the
happy existence within the uterus or of its traumatic disturb-
ance, entirely according to the necessities of the case, thus
throwing the door wide open to arbitrariness of interpretation.
Individual instances of this childhood anxiety flatly contradict
the Rankian principle. If the child is brought into darkness
and solitude, we should expect that it would welcome this
restoration of the intrauterine situation; and if the fact that in
precisely these circumstances the child reacts with anxiety is
ascribed to the memory of the interruption of that happy state
through birth, one may be pardoned for failing to appreciate
the appositeness of such reasoning.

I am forced to the conclusion that the earliest phobias of

273

childhood do not permit of being directly traced to the impres-
sion made upon the child by the act of birth, and that they have
thus far, in fact, defied all explanation. A certain predispo-
sition to anxiety on the part of the infant is indubitable. It is
not at its maximum immediately after birth, to diminish
gradually thereafter, but first makes its appearance later on
with the progress of psychic development, and persists over a
certain period of childhood. When early phobias of this sort
continue beyond such a period, they give rise to the suspicion
of a neurotic disturbance, although their relationship to the
definite neuroses of later childhood is in no wise clear.

Only a few instances of the expression of anxiety in infancy
are intelligible to us; we shall have to keep to these. Thus,
the three situations of being left alone, being in the dark, and
finding a strange person in place of the one in whom the child
has confidence (the mother), are all reducible to a single situa-
tion, that of feeling the loss of the loved (longed for) person.
From this point forward the way is clear to an understanding
of anxiety and to the reconciling of the contradictions which
seem to be connected with it.

The memory picture of the person longed for is certainly
cathected in very intense degree, probably at first in hallucina-
tory fashion. But this is without result, and now it appears
as if this longing were transformed into anxiety. It decidedly
seems as if this anxiety were an expression of helplessness, as if
the still very undeveloped creature did not know what else to
do with his longing. Anxiety thus seems to be a reaction to the
perception of the absence of the object, and there at once spring
to mind the analogies that castration anxiety has also separation
from a highly valued object as its content and that the most
basic anxiety of all, the “primal anxiety” of birth, arises in con-
nection with separation from the mother.

The next consideration takes us beyond this emphasis upon
loss of the object. If the infant longs for the sight of the
mother, it does so, surely, only because it already knows from
experience that she gratifies all its needs without delay. The
situation which the infant appraises as “danger”, and against
which it desires reassurance, is therefore one of not being grati-

274

fied, of an increase of tension arising from non-gratification of
its needs–a situation against which it is powerless. I believe
that from this standpoint everything falls into place; the situ-
ation of privation, in which stimuli reach an unpleasurable
magnitude and intensity without an ability to cope with them
psychically and thus provide for their discharge, must represent
to the infant a situation analogous to the birth experience, a
repetition of the danger situation; what the two situations have
in common is the economic disturbance brought about by an
increase in stimuli demanding some disposition made of them,
this common factor hence being the very essence of the
“danger”. In both cases the reaction of anxiety appears, a
reaction which still in the infant proves to the purpose since
the discharge of the anxiety via the respiratory and vocal
musculature now calls the mother to the infant’s side, just as
earlier it aroused respiratory activity to get rid of internal
stimuli. More than this sign of danger the child does not need
to have preserved from birth.

Along with the experiencing of the fact that an external and
perceptible object may put an end to the danger situation
reminiscent of birth, there takes place a displacement of the
content of the danger from the economic situation to that
which occasions it, namely, object loss. The perception of the
absence of the mother now becomes the danger at the appear-
ance of which the infant gives the signal of anxiety, even before
the economic situation which is feared has arisen. This change
represents a first great step in advance in the economy of self-
preservation, and includes at the same time the transition from
the automatically unpurposed creation de novo of anxiety to
its purposeful reproduction as a signal of danger.

In both respects, alike as an automatic phenomenon and as
a safety signal, anxiety proves to be a product of the psychic
helplessness of the infant which is the obvious counterpart of
its biological helplessness. The striking coincidence that both
birth anxiety and the anxiety of the infant alike claim separa-
tion from the mother as their prerequisite needs no psycho-
logical interpretation; it is simply enough explicable biologi-
cally by the fact that the mother, who in the beginning had

275

satisfied all the needs of the fœtus through her body mecha-
nisms, continues after birth as well to exercise in some measure
this same function, although by other means. Intrauterine
life and early infancy form a continuum to a far greater extent
than the striking cæsura of the act of birth would lead us to
believe. The psychic mother object replaces for the child the
biological fœtal situation. Hence we should not forget that
during intrauterine life the mother was not an object, and that
there were no objects at all at that period.

It is easy to see that in this continuum there is no room for
an abreacting of the birth trauma, and that any other function
of anxiety than that of a signal for avoiding a situation of
danger is not discoverable. Object loss as the precondition of
anxiety now has some further implications. For the next
transformation of anxiety, the castration anxiety which makes
its appearance in the phallic phase, is a separation anxiety also,
and is similarly conditioned. The danger here is separation
from the genital. A seemingly entirely legitimate line of
thought of Ferenczi’s enables us to recognize clearly here the
point of connection with the earlier content of the danger
situation. The high narcissistic value attaching to the penis
may be referable to the fact that the possession of this organ
contains a guarantee of reunion with the mother (or mother
substitute) in the act of coitus. Deprivation of this member is
tantamount to a second separation from the mother, and thus
has again the significance (as in the case of birth) of being
delivered over helpless to the unpleasurable tension arising
from the non-gratification of a need. This need, of which the
increase is feared, is now, however, a specialized one, a need of
the genital libido, and no longer an undifferentiated one, as
in infancy. I would add here that the fantasy of returning to
the uterus is the substitute for coitus which we find in impotent
men (those inhibited by the threat of castration). In the spirit
of Ferenczi’s formulation one may say that the individual who
wished to have his genital organ act as a proxy in his return to
the uterus in fact regressively substitutes for this organ his
whole body.

The various steps in the development of the child, its

276

increased independence, the sharper differentiation of its
mental apparatus into various agencies, the appearance of its
new needs—all these cannot remain without their effect upon
the content of the danger situation. We have followed the
change in the content of the latter from loss of the maternal
object to castration, and we now see the next step therein as
caused by the power of the superego. With the impersonaliza-
tion of the parental authority at whose hands castration was
feared, the danger becomes more indefinite. Fear of castration
develops into dread of conscience, into social anxiety. It is
now no longer easy to state what it is that there is fear of. The
formula, “separation, exclusion from the horde”, applies only
to that more lately developed portion of the superego which
was patterned after social models, not to the nucleus thereof
which corresponds to the introjected parental authority.
Expressed in more general terms, it is the anger, the punish-
ment, of the superego, the loss of its love, which the ego appre-
hends as a danger and to which it responds with the signal of
anxiety. The final transformation undergone by this fear of
the superego has appeared to me to consist of death- (life-)
anxiety, fear felt for the projection of the superego upon the
powers of destiny.

Formerly I attached a certain value to the proposition that
the cathexis withdrawn in repression finds employment as a
discharge in the form of anxiety. This seems to me today of
very little interest. The difference consists in the fact that
formerly I believed anxiety to originate in every instance auto-
matically through an economic process, whereas the present
conception of anxiety as a signal intended by the ego for the
purpose of influencing the pleasure-pain mechanism renders us
independent of this economic restriction. It does not contra-
dict this supposition, of course, that for the arousing of affect
the ego employs precisely the energy set free by the withdrawal
of cathexis in repression, but it has become unimportant to
distinguish with which moiety of energy this is accomplished.

Another assertion I once made now demands reëxamination
in the light of our new conception. I refer to the statement
that the ego is the real seat of anxiety; I think that this state-

277

ment will prove to be correct. That is to say, we have no
reason to ascribe any expression of anxiety to the superego.
But when it is a matter of an “anxiety of the id”, one does not
have so much to contradict this as to emend an infelicitous
expression. Anxiety is an affective state which can of course be
experienced only by the ego. The id cannot be afraid, as the
ego can; it is not an organization, and cannot estimate situ-
ations of danger. On the contrary, it is of extremely frequent
occurrence that processes are initiated or executed in the id
which give the ego occasion to develop anxiety; as a matter of
fact, the repressions which are probably the earliest are moti-
vated, like the majority of all later ones, by such fear on the
part of the ego of this or that process in the id. We have good
grounds here for once again distinguishing the two cases: that
in which something happens in the id which activates one of
the danger situations to which the ego is sensitive, causing the
latter to give the anxiety signal for inhibition; and that in
which there develops in the id a situation analogous to the
birth trauma, which automatically brings about a reaction of
anxiety. The two cases are brought into closer approximation
to each other if it is emphasized that the second corresponds to
the initial and original situation of danger, whereas the first
corresponds to one of the anxiety–occasioning situations subse-
quently derived from it. Or, to relate the matter to actually
existing disorders: the second case is that which is operative
in the ætiology of the “actual” neuroses, the first is character-
istic of the psychoneuroses.

We now see that we need not dismiss earlier formulations
as without value but have merely to bring them into line with
our newer understanding. It is undeniable that in abstinence,
in perverted interference with the normal discharge of sexual
excitation or in the diverting of the latter from its psychic
elaboration, anxiety arises directly out of libido; that is to say,
there is brought about that state of helplessness of the ego in
the face of excessive tension arising from ungratified need
which results, as in birth, in the development of anxiety, so
that there is again a possibility, which although obvious is of
no great consequence, that it is precisely the excess of unuti-
 

278

lized libido that finds its dis-
charge in the form of anxiety. We
know that psychoneuroses develop with particular readiness on
the basis of these “actual” neuroses; and this may mean that
the ego makes attempts to minimize and to fix by means of
symptoms the anxiety which it has learned to hold temporarily
in suspension. Probably analysis of the traumatic war neuroses
(although this term includes a wide variety of disorders, cer-
tainly) would have shown that a certain proportion of them
share the characteristics of “actual” neuroses.

When we represented the various danger situations as
developing out of the original prototype of birth, we were far
from maintaining that every later anxiety-occasioning situation
simply renders inoperative those which were earlier effective
in giving rise to anxiety. The progressive development of the
ego contributes, it is true, to depriving of value and relegating
to unimportance the earlier danger situation, so that it may be
said that to a given period of development is assigned the
anxiety-occasioning situation which is, so to speak, appropriate
to it. Psychic helplessness is the danger which is consonant
with the period of immaturity of the ego, as object loss is the
danger appertaining to the state of dependence of early child-
hood, the danger of castration to the phallic phase, and dread
of the superego to the latency period. And yet all these danger
situations and anxiety determinants may persist alongside one
another and cause the ego to react with anxiety at a later period
also than the appropriate one; or several of them may become
operative simultaneously. Possibly there also exists a close
relationship between the danger situation which is effective in
the given case and the form of the neurosis which develops in
consequence.¹⁾

¹⁾Since the differentiation between the ego and the id was made, our interest
in the problems of repression has necessarily undergone a revival. Until then
we were satisfied to dwell upon those of its elements which are referable to the
ego–namely, the keeping of the repressed material out of consciousness and its
withholding from motor discharge, and the creating of substitute- (symptom-)
formations; of the repressed instinctual impulse itself we assumed that it per-
sisted unchanged for an indefinite period in the unconscious. Now our interest
shifts to the fate of the repressed, and we begin to feel that this persistence,

279

When in an earlier chapter of this inquiry we encountered
the significance of the danger of castration in more than one
neurotic disorder, we warned ourselves against overestimating
this factor, since it assuredly could not be the crucial one in the
female sex, the sex certainly more predisposed to neurosis. We
see now that we are in no danger of taking castration anxiety
to be the sole motive force behind the defense processes result-
ing in neurosis. I have explained elsewhere how the develop-
ment of the little girl is guided to tender object cathexis
through the castration complex. It is precisely in the female
that object loss seems to remain the most effective situation of
danger. As to that which gives rise to her anxiety, we may
introduce the slight modification that it is no longer a matter
of feeling the absence, or of the loss of reality, of the object,
but rather of the loss of the object’s love. Since it is certainly
true that hysteria has a greater affinity with femininity, just as
compulsion neurosis has with masculinity, the idea suggests
itself that as a determinant of anxiety, loss of love plays a rôle
in hysteria similar to that of the threat of castration in the
phobias and of dread of the superego in compulsion neurosis.

Translated by HENRY ALDEN BUNKER

unchanged and unchanging, is not a matter of course, is perhaps not even the
rule. The original impulse has in any case been inhibited and deflected from
its aim. But has its root persisted in the unconscious, having proved resistant
to the modifying and depreciatory influence of life? Do there therefore still
exist the old desires, of the earlier existence of which analysis informs us? The
answer appears obvious and certain: The old repressed desires must still persist
in the unconscious, since we find their lineal descendants, the symptoms, still
active. But this answer is inadequate; it does not make it possible to distinguish
between the two possibilities that, on the one hand, the old desire now operates
only through its descendants, to which it has transferred all its cathectic energy,
or, on the other hand, that the desire itself persists in addition. If it was its
destiny to be expended in the cathexis of its descendants, there remains the
third possibility that in the course of the neurosis the wish was reactivated
through regression, so out of accord with the present may it be. One need not
regard these considerations as otiose; there is much in the phenomena of both
the morbid and the normal life of the psyche which seems to demand the
raising of such questions. In my study of the breakdown of the œdipus complex
I became mindful of the distinction between the mere repression and the true
disappearance of an old desire or impulse.