The Agency of the Letter
The agency of the letter in the unconscious
or reason since Freud
as well as men tightly bound with stout bonds around
their arms and Iegs by folk who will not understand
your language; and you will only be able to give
vent to your griefs and sense of loss of liberty
by making tearful complaints, and sighs, and
lamentations one to another; for those who
bind you will not understand your
language nor will you
LEONARDO DA VINCI
Although the nature of this contribution was determined by the theme
of the third volume of La Psychanalyse, I owe to what will be found
there to insert it at a point somewhere between writing (l'écrit) and
speech - it will be half-way between the two.
Writing is distingiushed by a prevalence of the text in the sense that
this factor of discourse will assume in this essay a factor that makes
possible the kind of tightening up that I like in order to leave the reader
no other way out than the way in, which I prefer to be difficult. In that
sense, then, this will not be writing.
Because I always try to provide my seminars each time with some.
thing new, I have refirained so far from giving such a text, with one exception, which is not particularly outstanding in the context of the series,
and which I refer to at all only for the general level of its argument.
For the urgency that I now take as a pretext for leaving aside such an
aim only masks the difficulty that, in trying to maintain it at the level at
which I ought to present my teaching here, I might push it too far from
speech, whose very different techniques are essential to the formative
effect I seek
That is why I have taken the expedient offered me by the invitation
to lecture to the philosophy group of the Fédération des étudiants dès
lettres to produce an adaptation suitable to what I have to say: its necessary generality matches the exceptional character of the audience, but its
sole object encounters the collusion of their common training, a literary
one, to which my title pays homage.
Indeed, how could we forget that to the end of his days Freud con-
stantly maintained that such a training was the prime requisite in the
formation of analysts, and that he designated the eternal universitas
litterarum as the ideal place for its institution.
Thus my recourse (in rewriting) to the movement of the (spoken)
discourse, restored to its vitality, by showing whom I meant it for,
marks even more clearly those for whom it is not intended.
I mean that it is not intended for those who, for any reason whatever,
in psychoanalysis, allow their discipline to avail itself of some false
identity - a fault of habit, but its effect on the mind is such that the true
identity may appear as simply one alibi among others, a sort of refined
reduplication whose implications will not be lost on the most subtle minds.
So one observes with a certain curiosity the beginnings of a new direc-
tion concerning symbolization and language in the Internationl Journal
of Psychoanalysis, with a great many sticky fingers leafing through the
pages of Sapir and Jespersen. These exercises are still somewhat un-
practised, but it is above all the tone that is lacking. A certain'seriousness'
as one enters the domain of veracity cannot fail to raise a smile.
And how could a psychoanalyst of today not realize that speech is the
key to that truth, when his whole experience must find in speech alone
its instrument, its context, its material, and even the background noise
of its uncertainties.
As my title suggests, beyond this 'speech', what the psychoanalytic ex-
petience discovers in the unconscious is the whole structure of language.
Thus from the outset I have alerted informed minds to the extent to
which the notion that the unconscious is merely the seat of the instincts
will have to be rethought.
But how are we to take this 'letter' here? Quite simply, literally.
By 'letter' I designate that materia1 support that concrete discourse
borrows from language.
This simple definition assumes that language is not to be confused
with the various psychical and somatic functions that serve it in the
speaking subject - primarily because language and its structure exist
prior to the moment at which each subject at a certain point in his mental
development makes his entry into it.
Let us note, then, that aphasias, although caused by purely anatomical
lesions in the cerebral apparatus that supplies the mental centre for these
functions, prove, on the whole, to distribute their deficits between the
two sides of the signifying effect of what we call here 'the letter' in the
creation of signification. A point that will be clarified later.
Thus the subject, too, if he can appear to be the slave of language
is all the more so of a discourse in the universal movement in which
his place is already inscribed at birth, if only byvirtue of his proper
Reference to the experience of the community, or to the substance of
this discourse, settles nothing. For this experience assumes its essential
dimension in the tradition that this discourse itself establishes. This
tradition, long before the drama of history is inscribed in it, lays down
the elementary structures of culture. And these very structures reveal an
ordering of possible exchanges which, even if unconscious, is inconceiv-
able outside the permutations authorized by language.
With the result that the ethnographic duality of nature and culture is
giving way to a ternary conception of the human condition - nature,
society, and culture - the last term of which could well be reduced to
language, or that which essentially distinguishes human society from
But I shall not make of this distinction either a point or a point of de-
parture, leaving to its own obscurity the question of the original rela-
tions between the signifier and labour. I shall be content, for my little
jab at the general function of praxis in the genesis of history, to point out
that the very society that wished to restore, along with the privileges of
the producer, the causal hierarchy of the relations between production
and the ideological superstructure to their full political rights, has none
the less failed to give birth to an ^esperanto in which the relations of lan-
guage to socialist realities would have rendered any literary formalism
For my part, I shall trust only those assumptions that have already
proven their value by virtue of the fact that language through them has
attained the status of an object of scientific investigation.
For it is by virtue of this fact that linguistics is seen to occupy the key
position in this domain, and the reclassification of the sciences and a re-
grouping of them around it signals, as is usually the case, a revolution in
knowledge; only the necessities of communication made me inscribe it
at the head of this volume under the title 'the sciences of man' - despite
the confusion that is thereby covered over.
To pinpoint the emergence of linguistic science we may say that, as
in the case of all sciences in the modern sense, it is contained in the
constitutive moment of an algorithm that is its foundation. This algorithm
is the following:
which is read as: the signifier over the signified, 'over' corresponding to
the bar separating the two stages.
This sign should be attributed to Ferdinand de Saussure although it
is not found in exactly this form in any of the numerous schemas, which
none the less express it, to be found in the printed version of his lectures
of the years I906-7, I908‹9, and I9I0-11, which the piety of a group of
his disciples caused to be published under the title, Cours de linguistique
génerale, a work of prime importance for the transmission of a teaching
worthy of the name, that is, that one can come to terms with only in its
That is why it is legitimate for us to give him credit for the formula-
tion S/s by which, in spite of the differences among schools, the begin-
ning of modern linguistics can be recognized.
The thematics of this science is henceforth suspended, in effect, at the
primordial position of the signifier and the signified as being distinct
orders separated initially by a barrier resisting signification. And that is
what was to make possible an exact study of the connections proper to
the signifier, and of the extent of their function in the genesis of the
For this primordial distinction goes well beyond the discussion con-
cerning the arbitrariness of the sign, as it has been elaborated since the
earliest reflections of the ancients, and even beyond the impasse which,
through the same period, has been encountered in every discussion of the
bi-univocal correspondence between the word and the thing, if only in
the mere act of naming. All this, of course, is quite contrary to the
appearances suggested by the importance often imputed to the role of the
index finger pointing to an object in the learning pro¦ss of the infans
subject learning his mother tongue, or the use in foreign language
teaching of so-called 'concrete' methods.
One cannot go further along this line of thought than to demonstrate
that no signification can be sustained other than by reference to another
signification: in its extreme form this amounts to the proposition that
there is no language (langue) in existence for which there is any question
of its inability to cover the whole field of the signified, it being an effect
of its existence as a language (langue) that it necessarily answers all needs.
If we try to grasp in language the constitution of the object, we cannot
fail to notice that this constitution is to be found only at the level of
concept, a very different thing from a simple nominative, and that the
thing, when reduced to the noun, breaks up into the double, divergent
beam of the 'cause' (causa) in which it has taken shelter in the French
word chose, and the nothing (rien) to which it has abandoned its Latin
These considerations, important as their existence is for the philoso-
pher, turn us away from the locus in which language questions us as to
its very nature. And we will fail to pursue the question further as long
as we cling to the illusion that the signifier answers to the function of
representing the signified, or better, that the signifier has to answer for
its existence in the name of any signification whatever.
For even reduced to this latter formulation, the heresy is the same -
the heresy that leads logical positivism in search of the 'meaning of
meaning', as its objective is called in the language of its devotees. As a
result, we can observe that even a text highly charged with meaning can
be reduced, through this sort of analysis, to insignificant bagatelles, all
that survives being mathematical algorithms that are, of course, without
To return to our formula S/s: if we could infer nothing from it but the
notion of the parallelism of its upper and lower terms, each one taken in
its globality, it would remain the enigmatic sign of a total mystery. Which
of course is not the case.
In order to grasp its function I shall begin by reproducing the classic,
yet faulty illustration by which its usage is normally introduced, and one can see how it opens the way to the kind of error referred to above.
<img src="agency1.jpeg" align="middle">
In my lecture, I replaced this illustration with another, which has no
greater claim to correctness than that it has been transplanted into that
incongruous dimension that the psychoanalyst has not yet altogether
renounced because of his quite justified feeling that his conformism
takes its value entirely from it. Here is the other diagram:
<img src="agency2.jpeg" align="middle">
where we see that, without greatly extending the scope of the signifier
concerned in the experiment, that is, by doubling a noun through the
mere juxtaposition of two terms whose complementary meanings ought
apparently to reinforce each other, a surprise is produced by an unexpec-
ted precipitation of an unexpected meaning: the image of twin doors
symbolizing, through the solitary confinement offered Western Man for
the satisfaction of his natural needs away from home, the imperative
that he seems to share with the great majority of primitive communities
by which his public life is subjected to the laws of urinary segregation.
It is not only with the idea of silencing the nominalist debate with a
low blow that I use this example, but rather to show how in fact the
signifier enters the signified, namely, in a form which, not being im-
material, raises the question of its place in reality. For the blinking gaze
of a short sighted person might be justified in wondering whether this
was indeed the signifier as he peered closely at the little enamel signs that
bore it, a signifier whose signified would in this call receive its final
honours from the double and solemn procession from the upper nave.
But no contrived example can be as telling as the actual experience of
truth. So I am happy to have invented the above, since it awoke in the
person whose word I most trust a memory of childhood, which having
thus happily come to my attention is best placed here.
A train arrives at a station. A little boy and a little girl, brother and
sister, are seated in a compartment face to face next to the window
through which the buildings along the station platform can be seen
passing as the train pulls to a stop. 'Look', says the brother, 'we're at
Ladies!'; 'Idiot!' replies his sister, 'Can't you see we're at Gentlemen'.
Besides the fact that the rails in this story materialize the bar in the
Saussurian algorithm (and in a form designed to suggest that its resis-
tance may be other than dialectical), we should add that only someone
who didn't have his eyes in front of the holes (it's the appropriate image
here) could possibly confuse the place of the signifier and the signified
in this story, or not see from what radiating centre the signifier sends
forth its light into the shadow of incomplete significations.
For this signifier will now carry a purely animal Dissension, destined
for the usual oblivion of natural mists, to the unbridled power of ideolo-
gical warfare, relentless for families, a torment to the Gods. For these
children, Ladies and Gentlemen will be hen¦forth two countries to-
wards which each of their souls will strive on divergent wings, and
between which a truce will be the more impossible since they are actually
the same country and neither can compromise on its own superiority
without detracting from the glory of the other.
But enough. It is beginning to sound like the history of France. Which
it is more human, as it ought to be, to evoke here than that of England,
destined to tumble from the Large to the Small End of Dean Swift's egg.
It remains to be conceived what steps, what corridor, the S of the
signifier, visible here in the plurals in which it focuses its welcome
beyond the window, must take in order to rest its elbows on the ventila-
tors through which, like warm and cold air, indignation and scorn come
hissing out below.
One thing is certain: if the algorithm S/s with its bar is appropriate,
access from one to the other cannot in any case have a signification. For
in so far as it is itself only pure function of the signifier, the algorithm
can reveal only the structure of a signifier in this transfer.
Now the structure of the signifier is, as it is commonly said of language
itself, that it should be articulated.
This means that no matter where one starts to designate their recipro-
cal encroachments and increasing inclusions, these units are subjected to
the double condition of being reducible to ukimate differential elements
and of combining them according to the laws of a closed order.
These elements, one of the decisive discoveries of linguistics, are
phonemes; but we must not expect to find any phonetic constancy in the
modulatory variability to which this term applies, but rather the syn-
chronic system of differential couplings necessary for the discernment of
sounds in a given language. Through this, one sees that an essential
element of the spoken word itself was predestined to flow into the
mobile characters which, in a jumble of lower-case Didots or Gara-
monds, render validly present what we call the 'letter', namely, the
essentially localized structure of the signifier.
With the second property of the signifier, that of combining according
to the laws of a closed order, is affirmed the necessity of the topological
substratum of which the term I ordinarily use, namely, the signifying
chain, gives an approximate idea: rings of a neckla¦ that is a ring in
another neckla¦ made of rings.
Such are the structural conditions that define grammar as the order of
constitutive encroachments of the signifier up to the level of the unit
immediately superior to the sentence, and lexicology as the order of con-
stitutive inclusions of the signifier to the level of the verbal locution.
In examining the limits by which these two exerciseS in the under-
standing of linguistic usage are determined, it is easy to see that only the
correlations between signifier and signifier provide the standard for all
research into signification, as is indicated by the notion of 'usage' of a
taxeme or semanteme which in fact refers to the context just above that
of the units con¦rned.
But it is not because the undertakings of grammar and lexicology are
exhausted within ¦rtain limits that we must think that beyond those
limits signification reigns supreme. That would be an error.
For the signifier, by its very nature, always anticipates meaning by
unfolding its dimension before it. As is seen at the level of the sentence
when it is interrupted before the significant term: 'I shall never...',
'All the same it is...', 'And yet there may be. ..'. Such sentences are
not without meaning, a meaning all the more oppressive in that it is
content to make us wait for it.
We are forced, then, to accept the notion of an incessant sliding of the
signified under the signifier - which Ferdinand de Saussure illustrates
with an image resembling the wavy lines of the upper and lower Waters
in miniatures from manuscripts of Genesis; a double flux marked by fine
streaks of rain, vertical dotted lines supposedly confining segments of
All our experience runs counter to this linearity, which made me speak
once, in one of my seminars on psychosis, of something more like
'anchoring points' ('points de capiton') as a schema for taking into account
the dominance of the letter in the dramatic transformation that dialogue
can effect in the subject.
The linearity that Saussure holds to be constitutive of the chain of dis-
course, in conformity with its emission by a single voice and with its
horizontal posidon in our writing - if this linearity is necessary, in fact,
it is not suflicient. It applies to the chain of discourse only in the direc-
tion in which it is orientated in time, being taken as a signifying factor
in all languages in which 'Peter hits Paul' reverses its dme when the terms
But one has only to listen to poetry, which Saussure was no doubt in
the habit of doing, for a polyphony to be heard, for it to become clear
that all discourse is aligned along the several staves of a score.
There is in effect no signifying chain that does not have, as if attached
to the punctuation of each of its units, a whole articulation of relevant
contexts suspended 'vertically', as it were, from that point.
Let us take our word 'tree' again, this time not as an isolated noun,
but at the point of one of these punctuations, and see how it crosses the
bar of the Saussurian algorithm. (The anagram of 'arbre' and 'barre'
should be noted.)
For even broken down into the double spectre of its vowels and con-
sonants, it can still call up with the robur and the plane tree the significa-
tions it takes on, in the context of our flora, of strength and majesty.
Drawing on all the symbolic contexts suggested in the Hebrew of the
Bible, it erects on a barren hill the shadow of the cross. Then reduces to
the capital Y, the sign of dichotomy which, except for the illustration
used by heraldry, would owe nothing to the tree however genealogical
we may think it. Circulatory tree, tree of life of the cerebellum, tree of
Saturn, tree of Diana, crystals formed in a tree struck by lightning, is it
your figure that traces our desdny for us in the tortoise-shell cracked
by the fire, or your lightning that causes that slow shift in the axis of
being to surge up from an unnamable night into the '`