Downloaded from https://academic.oup.

com/screen/article/16/3/19/1603276 by King's College London user on 09 March 2023

The Unattainable Text

Raymond Bellour

That the film is a text, in the sense in which Barthes uses the word,
is obvious enough. That as such it might, or should, receive the
same kind of attention as has been devoted to the literary text is
also obvious. But already not quite so obvious. We shall soon see
why.
The text of the film is indeed an unattainable text. In saying
this, despite the temptation of a play on words, I do not mean to
evoke the special difficulties which very often make it impossible
to obtain the film in the material sense or the proper conditions to
constitute it into a text, ie the editing table or the projector with
freeze-frame facility. These difficulties are still enormous: they are
very often discouraging, and go a long way to explaining .the com-
parative backwardness of film studies. However, one can imagine,
if still only hypothetically, that one day, at the price of a few
changes, the film will find something that is hard to express, a
status analogous to that of the book or rather that of the gramo-
phone record with respect to the concert. If film studies are still
done then, they will undoubtedly be more numerous, more imagina-
tive, more accurate and above all more enjoyable than the ones we
carry out in fear and trembling, threatened continually with dis-
possession of the object. And yet, curious as it might seem, the
situation of the film analyst, even when he does possess the film,
any film, will not change in every particular.
I shall not linger over the indisputable fact that one does not
have the text, the ' methodological field ', the ' production ', the
1
traversal', as Barthes puts it, when one has the work, the ' frag-
ment of substance V But without going into the theoretical laby-
rinths opened up by the notion of the text, I shall stress two

1. ' De 1'ceuvre au texte,' Revue d'Esthetique n 3, 1971. The quotations

that follow arc taken from this same article.
20 things. On the one hand the material possession of the work alone
permits one full access to the textual fiction, since it alone allows
one a full experience of the multiplicity of operations carried out
in the work and makes it precisely into a text. On the other, as

Downloaded from https://academic.oup.com/screen/article/16/3/19/1603276 by King's College London user on 09 March 2023

soon as one studies a work, quotes a fragment of it, one has
implicitly taken up a textual perspective, even if feebly and one-
dimensionally, even if in a restrictive and regressive fashion, even
if one continues to dose the text back onto itself although it is,
as Barthes has insisted, and before him, Blanchot, the locus of an
unbounded openness. That is why it is possible, in a slide which
is both justified and somewhat abusive, like all slides, to speak of
quoting the text when by text one means work, even if at a later
stage one may be driven, as Barthes has been, to think the literary
experience from the starting-point of an opposition between the
work and the text. In connection with these terms, but without
evading them, I should just like to emphasise here an elementary
fatality: the text of the film is unattainable because it is an
unquotable text. To this extent, and to this extent only, the word
text as applied to film is metaphorical; it clearly pin-points the
paradox which inflicts the filmic text and to such a degree only the
filmic text.
When one chooses to read, to study a work, to recognise in it
the pressure of the text, so close in a sense to what Blanchot has
conceptualised as literature, nothing is more immediate, simpler
than to quote a word, a phrase, a few lines, a sentence, a page.
Omit the quotation marks that signal it and the quotation is
invisible, it is quite naturally absorbed into the page. Despite the
change of regime it introduces, it does not really break up the
reading; it even helps to make description, analysis a special form
of discourse, in the best of cases a new text, by a reduplication
whose fascination has been fully felt by modern thought. This
effect is obviously peculiar to the literary work, more generally
to the written work, and to it alone. It lies in the undivided con-
formity of the object of study and the means of study, in the
absolute material coincidence between language and language. That
is why only the written work was able to provide, so to speak, a
pretext for a theory of the text, or at least for the first effects of
its practice. That is why Barthes so strongly distrusts everything
that escapes the written, for the meta-language effect is more
tangible there, by definition. Indeed, one speaks the more ' about'
an object the less one can draw it into the material body of the
commentary. At the same time this is obviously to emphasise the
absolute privilege of written expression in this conversion of the
work into a text. The material reality of a commentary which in
its turn comes to have more or less the function of a text con-
stitutes the necessary mediation for this transmutation which in
the last instance would like to appear in the absolute guise of a
play. That is to say that in fact it aims for an integral reconciliation
 between language and language, and between the subject and the 21
subject, receiving from the exteriority of language the absolution
that would restore it to its desire. For clarity's sake, one thing
should be remembered. This idea arises with the joint emergence

Downloaded from https://academic.oup.com/screen/article/16/3/19/1603276 by King's College London user on 09 March 2023

of the two concepts literature and science of literature. It arose
for the first time, in a still uncertain fashion, with romanticism
and the beginnings of literary criticism; a second time at the turn
of the century in the first great mutual concussion of literature
and the human sciences, in Nietzsche and Mallarm£, Freud and
Saussure; a third time today under the internal and external pres-
sure exercised on literature by what Barthes has called ' the con-
joint action of Marxism, psycho-analysis and structuralism'. To
sum up, let us say that the science of literature has enabled us to
recognise in the work the reality and the Utopia of the text, but
this movement has no meaning unless it dissolves the science into
the body of its object, to the extent, in the ideal case, of abolish-
ing any divergence between science and literature, analysis and
the work.2
It is from the starting point of this both real and mythical level
that the apparently quite secondary fact of the possibility of quota-
tion turns out to assign a paradoxical specificity to the cinematic
text. The written text is the only one that can be quoted un-
impededly and unreservedly. But the filmic text does not have the
same differential relations with the written text as the pictorial
text, the musical text, the theatrical text (and all the intermediate
mixed texts they give rise to). The pictorial text is in fact a quot-
able text. No doubt the quotation stands out in its heterogeneity,
its difference; no doubt there are many material difficulties in its
way, difficulties expressing the specifically material loss undergone
by the work from the very fact of its reproduction. The format of
the book in particular, always reductive, obviously produces an
inevitable distortion through the disproportion between the original
and its reproduction. But the quotation is on the other hand per-
fectly satisfactory, allowing a remarkable play on the detail with
respect to the whole. From the critical point of view it has one
2. ' In its own way the text shares in a social Utopia; before History
(always supposing the latter does not choose barbarism), the Text
achieves, if not the transparency of social relations, at least that
of relations of language: it is the space in which no language has
an edge on any other, in which languages circulate (retaining the
circular sense of the term). . . . A theory of the Text cannot be
satisfied with a metalinguistic exposition; to destroy metalanguage,
or at least (for it may be necessary to resort to it for the time being)
to cast suspicion on it, is part of the aim of the theory itself: dis-
course about the Text should never be anything but text itself,
textual research and travail, since the Text is that social space that
allows no language any shelter outside it, nor any subject of
enunciation in the position of judge, teacher, analyst, confessor,
decipherer: the theory of the Text cannot but coincide with a
practice of writing.'
22 advantage that only painting possesses: one can see and take in
the work at one glance. Which literary analysis cannot do, except
when it has as its object short poems in which vision and reading
are superimposed (eg Ruwet's, L6vi-Strauss's and Jakobson's analy-
ses of Baudelaire sonnets). Beyond these, even when it chooses to

Downloaded from https://academic.oup.com/screen/article/16/3/19/1603276 by King's College London user on 09 March 2023

quote ' the whole text' in limit-case experiments like Barthes's in
S/Z, it can only rediscover the inevitable linearity of the written.
The musical text, conversely, sets two obstacles in the way of
quotation. First, at the level of the score. This is certainly quotable,
in whole or in part, like the literary text. But it opposes an
.infinitely greater heterogeneity to language than that of the pic-
ture; that of a specific codification whose extreme technicality
marks a break. On the other hand, and much more profoundly (for
a society in which everyone could read music is conceivable - was
this not the case in the micro-societies of the aristocracy and the
bourgeoisie?), the musical text is divided, since the score is not
the performance. But sound cannot be quoted. It cannot be
described or evoked. In this the musical text is irreducible to the
text, even if it is, metaphorically, and in reality thanks to the
plurality of its operations, just as textual as the literary text. With
the one difference that it cannot really be experienced except by
hearing it, and never by analysing it, subjecting it to a reading,
since then one is no longer hearing it, or only hearing it virtually.
Finally, one last problem and not the least: the score is fixed but
performance changes. Some more or less aleatoric types of modern
music which increase this gap between score and performance take
the phenomenon to an extreme, but do not change its terms. The
work is unstable. In a sense this mobility increases even further
the degree of textuality of the musical work, since the text, as
Barthes has said again and again, is mobility itself. But by a kind
of paradox, this mobility cannot be reduced to the language which
attempts to grasp it in order to bring it out by duplication. In
this the musical text is less textual than are the pictorial text and
above all the literary text, whose mobility is in some sense inversely
proportional to the fixity of the work. The possibility of keeping to
the letter of the text is in fact the condition of its possibility.
The theatrical text demonstrates the same paradox and the same
division, although in a different way. On the one hand, the work,
the text in the ordinary sense of the term, can be reduced un-
equivocally to the problematic of the literary text, except that the
play more or less inevitably brings with it the absence of its per-
formance. On the other, the performance creates a mobile text, as
open and aleatoric as that of the musical text. A mise-en-scene
can be discussed, its principles stated, its novelty, its uniqueness
felt, but it cannot really be described or quoted. Its textuality,
though indisputable, escapes the text once again through its
infinite mobility, the too radical divergence between the text which
provides it with a pretext and a material and vocal figurability
without any real delimitation. At most, just as the gramophone 23
record has become the fixed memory of the concert, making an
end if not of the variety of interpretations, at least of the internal
variability of each performance, one might imagine fixing some

Downloaded from https://academic.oup.com/screen/article/16/3/19/1603276 by King's College London user on 09 March 2023

mise-en-scene, as has been done on all too few occasions, by the
only means apt to reproduce it: the film. Which, pushing aside the
problem of the theatre, automatically reinforces the paradoxical
uniqueness of the cinema.
Indeed, the film presents the remarkable speciality, for a spec-
tacle, of being a fixed work. The scenario, the initial technical cut,
are indeed not absolutely comparable with the score or the
theatrical play. They are pro-texts, as, without being similar, plans
and drafts are for the written work, sketches for a picture. Per-
formance, in the film, is annihilated in the same way, to the
advantage of the immutability of the work. This immutability, as
we have seen, is a paradoxical precondition for the conversion
of the work into a text, insofar as, if only by the abuttment it
constitutes, it favours the possibility of a voyage through language
which unties and reties the many operations by which the work is
made into a text. But this movement, which brings the film closer
to the picture and the book, is at the same time a broadly con-
tradictory one: indeed, the text of the film never fails to escape
the language that constitutes it. In a sense one can no more quote
a film than one can a musical work or a theatrical production.
However, this is not quite true. The analysis of the film suffers
the force of this paradox, which derives from the perfect delimit-
ation of the work, but equally from the mixture of materials whose
location is the cinema.
Once it is a talking cinema, it conjoins five matters of expression,
as Christian Metz has shown: phonetic sound, written titles,
musical sound, noises, the moving photographic image. The first
two of these pose no apparent problems for quotation. Nothing is
more easily reproduced than the dialogue of a film: publishers
know what they are doing when they imply, as they often do, that
they are recreating the film for us by printing its dialogue and
playing a dubious game with the image to recreate that absolutely
illusory thing known as its story. But it is quite obvious that
something is lost thereby: written titles belong fully to the written,
dialogue both to sound and to the written (it was written before
being spoken, and even if it is improvised, it can be transcribed,
since it does not change). Thus it undergoes a considerable reduc-
tion as soon as it is quoted: it loses tone, intensities, timbres,
pitches, everything that constitutes the profound solidity of the
voice. The same is true of noises, except that it is much less easy
to reduce them to the signified, since this reduction can only be
a translation, a kind of paraphrastic evocation. In this respect,
what might be called motivated noise, which can always be evoked
more or less since it indicates the real, should always be dis-
24 tdnguished from arbitrary noise, which can go so far as to serve
as a score, then escaping all translatability since it is not even
codified as the musical score is (confining ourselves for simplifi-
cation's sake to music in which the score is still truly determinant).
Note that these are only two extremes, extremes which can be

Downloaded from https://academic.oup.com/screen/article/16/3/19/1603276 by King's College London user on 09 March 2023

inverted: an arbitrary, but simple noise can be delimited, while
a motivated but overcomplex one cannot. How in an analysis is
one to deal with the noise band of a film like La mort en ce jardin.
for example, made up solely of the noises of the Amazon forest,
but so rich that it substitutes more or less for music? The bird
calls in The Birds can be thought of in the same way; orchestrated
by Robert Burks, thanks to the possibilities of electronic sound,
they constitute a true score in this film from which music is
apparently absent. In short, noise constitutes a greater obstacle to
the textuality of the film the more it is one of the major instru-
ments of its textual materiality. Musical sound obviously takes this
divergence between text and text to the extreme: given the specifi-
cations implied by the phenomenon of combination which makes
film music not a work in itself but an internal dimension of the
work, we have here again the problems posed in this respect by
musical works. With one difference, and by no means a negligible
one. If the division between score and performance, code and sound,
remains an integral one, here the musical text is received, thanks
to a petrifaction seemingly opposed to its very virtuality, in that
immutability of the work which defines the film.
There remains the image. And with it, rightly or wrongly, the
essential. First for a historical reason: for thirty years, with the
indispensable support of written titles (and not counting the inter-
mittent assistance of a music outside the material specificity of
the work), it represented the film, all films: the cinema. To the
extent that even today it is too often confused with it, by an
excessive simplification the a priori assumptions of which have been
unravelled by Christian Metz. The unique situation of the image
among the cinema's matters of expression will perhaps allow us,
if not to excuse this excess, at least to understand it. The image
is indeed located, with respect to the echo it might receive from
language, half-way between the semi-transparency of written titles
and dialogue and the more or less complete opacity of music and
noise. Moreover, it is this which quite logically gives the image as
such, as a moving image, the highest degree of cinematic specificity
among the matters of expression whose combination, on the other
hand, creates many more or less specifically cinematic co-ordina-
tions. Until very recently, no doubt, this insistence on the speci-
ficity of the image was usually a convenient pretext to subtract the
film from any true critical undertaking and to negotiate, as it were,
the image in terms of the scenario, ie of contents, themes. But
over and above its distortions, its inadequacies, which are as
negative as they are idealist, this contradiction did confusedly
express something absolutely essential: a highly paradoxical 25
relationship between the moving image and the language which
seeks to reveal in the film the filmic text itself. This has been
clearly seen since the area was turned upside down by the semio-

Downloaded from https://academic.oup.com/screen/article/16/3/19/1603276 by King's College London user on 09 March 2023

logy of the cinema and the first true textual analyses. It is no
accident that the only code constituted by Christian Metz has
been a syntagmatics of the image band, and if most analyses have
concentrated, with a kind of quite explicable impatience and
fascination, on the textual workings of the image as it were,
expressing a voluntarily agreed restriction that clearly never ceases
to transgress its limit, since that limit is illusory.
This restriction and fascination derive from the paradox intro-
duced by the moving image. On the one hand it spreads in space
like a picture; on the other it plunges into time, like a story which
its serialisation into units approximates more or less to the musical
work. In this it is peculiarly unquotable, since the written text
cannot restore to it what only the projector can produce; a move-
ment, the illusion of which guarantees the reality. That is why the
reproduction even of many stills only ever reveals a kind of radical
inability to assume the textuality of the film. However, stills are
essential. Indeed they represent an equivalent, arranged each time
according to the needs of the reading, to freeze-frames on the
editing table, with the absolutely contradictory function of open-
ing up the textuality of the film just at the moment they inter-
rupt its unfolding. In a sense it is really what is done when stopping
at a sentence in a book to re-read it and reflect on it. But then it
is not the same movement that is frozen. Continuity is suspended,
meaning fragmented; but the material specificity of a means of ex-
pression is not interfered with in the same way. The cinema,
through the moving image, is the only art of time which, when
we go against the principle on which it is based, still turns out to
give us something to see, and moreover something which alone
allows us to feel its textuality fully: a theatrical play cannot be
stopped, unless it has been filmed, nor can a concert, and if a
gramophone record is stopped there is simply nothing left to hear.
That is why it turns out that despite what it does allow, the gramo-
phone record (or the recording tape), which might seem the
magical instrument of musical analysis, only apparently resolves a
basic contradiction, that of sound. The frozen frame and the still
that reproduces it are simulacra; obviously they never prevent the
film from escaping, but paradoxically they allow it to escape as a
text. Obviously the language of the analysis is responsible for the
rest. It attempts to link together the multiplicity of textual opera-
tions between the simulacra of the frozen images like any other
analysis. But the analysis of the film thus receives its portion of an
inevitability known to no other: not to literary analysis, which
constantly makes language return freely to language; nor to the
analysis of the picture, which can partly or wholly re-establish
26 its object in the space of the commentary; nor to musical analysis,
irreducibly divided between the accuracy of a score and the other-
ness of a performance; nor to that of theatrical representation,
where the same division is at once less complete and less precise.
In fact, filmic analysis, if it is to take place at all, must take upon

Downloaded from https://academic.oup.com/screen/article/16/3/19/1603276 by King's College London user on 09 March 2023

itself this rhythmical as well as figurative and actantial narrative
component for which the stills are the simulacra, indispensable but
already derisory in comparison with what they represent. Thus it
constantly mimics, evokes, describes; in a kind of principled despair
it can but try frantically to compete with the object it is attempting
to understand. By dint of seeking to capture it and recapture it, it
ends up always occupying a point at which its object is per-
petually out of reach. That is why filmic analyses, once they begin
to be precise, and while, for the reasons I have just suggested,
they remain strangely incomplete, are always so long, according
to the extent of their coverage, even if analysis is, as we know,
always in a sense interminable. That is why they are so difficult, or
more accurately so ungrateful to read, repetitive, complicated, I
shall not say needlessly so, but necessarily so, as the price of their
strange perversity. That is why they always seem a little fictional:
playing on an absent object, never able, since their aim is to make
it present, to adopt the instruments of fiction even though they
have to borrow them. The analysis of film never stops filling up a
film that never stops running out: it is the Danaids' cask par
excellence. This is what makes the text of the film an unattainable
text: but it is so surely only at this price.
Although it would already be to go much further, we might
change our point of view completely and ask if the filmic text
should really be approached in writing at all. I think a contrario
of the wonderful impression I received on two occasions, to cite
only these two, when confronted with two quotations in which
film was taken as the medium of its own criticism. This was in
two broadcasts in the series * Cineastes de notre temps', on Max
Ophuls and Samuel Fuller. One saw, and then resaw while a voice
off emphasised certain features, two of the most extraordinary
camera movements in the history of the cinema, in which such
movements are by no means uncommon. The first in the ball in Le
Plaisir, just as the masked figure more and more unsteadily crosses
the length of the ball room, then collapses in a box where, beneath
the mask of a young man an old one is revealed; the second, in
Forty Guns, follows the hero from the hotel he is leaving to the
post office to which he goes to send a telegram, and saves for the
end of a long dialogue his meeting in a single continuous field
with the ' forty guns ' who race past on horseback on the left side
of the frame. Here there is no longer any divergence, no need of
narration. A true quotation, in all its obviousness. But this sudden
quotability which film allows to film (and in the same way sound
to sound) obviously has its other side: will oral language ever be
able to say what written language says? And if not, at the price of 27
what changes? Beneath the appearances of an answer a contrario,
this is a serious question, economic, social, political, profoundly
historical, since it touches on the formidable collusion of writing

Downloaded from https://academic.oup.com/screen/article/16/3/19/1603276 by King's College London user on 09 March 2023

and Western history in which the written alternately or even simul-
taneously performs a liberating and repressive function. Can or
should the work, be it image or sound, in its efforts to accede to
the text, ie to the social Utopia of a language without separation,
do without the text, free itself from the text?

FILM FORM
Theory and Practice in Film Analysis
The first issue of a new magazine whose aim is to apply
current development in film theory to the study of specific
films and film-making practice

Number One, now available, includes:

PETER WOLLEN on North-by-North-West
ANDREW TUDOR on Major Dundee
ANTHONY HARRILD on Red Psalm
STEPHEN CROFTS on Chapayev
CELIA BRITTON on La Grande Illusion
Subscription for one year (three issues):
UK £1.50 USA $5.00
This issue (including postage):
UK 60p USA $2.00
Orders, subscriptions and contributions should be
addressed to FILM FORM, Room 116, Squires Building,
Sandyford Road, Newcastle upon Tyne, NE1 8ST
 BRITISH FILM INSTITUTE

Downloaded from https://academic.oup.com/screen/article/16/3/19/1603276 by King's College London user on 09 March 2023

HDUCAT.ONALADV.SORYSERV.CE SEMINAR

The Imaginary - The Insufficient Signifier

by Jacqueline Rose

Monday 10 November 1975, 7.00 p

m
Boardroom, 4th Floor, BFI

Simone Simon in Co* i>«>pfe~

3WS

'round; Cat People.

Because we

recurrent study.
Further information from-
A^gonsandDistribution

'British Film Institute,

81 Dean Street,
London W1V6A A