FILM

LANGUAGE
A Semiotics of the Cinema
Christian Metz
Translated by Michael Taylor
ffi

The University of Chicago Press 1Il -----

3
1
3 The Cinema:
Language or Language System?
The Era of "Montage-or-Bust"*
In April 1959, in one of the famous interviews in the Cahier:s du cin-
emal Roberto Rossellini-discussing, among other subjects, the
problem of montage-expressed an opinion which was not new but
to which, toward the end of the text, he gave a more personal twist.
First, he made a commonplace observation: In the modern cinema,
montage does not occupy the same place it did in the great period
of 19
2
5-3
0
. Of course it remains an indispensable phase of film crea-
tion: One must select what one films, and what one has filmed must
be combined. And, since one must edit and adjust, should one not
do it in the best possible way and make the cut in the right place?
However, continued the creator of Paisa, montage is no longer con-
ceived of today as all-powerful manipulation. These are not the ac-
tual words of the Italian film-maker, but the formula does summarize
Rossellini's most suggestive remarks.
Montage as supreme ordering-is that not the montage which,
during its great period, lay claim to a persuasive power considered
in some way absolute and which was "scientifically" guaranteed by
Kuleshov's famous experiments? And is it not that montage whose
effectiveness-overestimated perhaps, but nonetheless very real-made
* The French is: montag roi.-Editor.
3
2
PROBLEMS OF FILM SEMIOTICS
such a vivid impression on the young Eisenstein? Startled at first by
the almost dishonoset grossness of the efficacy'" it gave him, Eisen-
stein soon let his OWn mind be conquered by the desire to conquer
other minds, and he became the leading theoretician of the "montage
or bust" approach.
2
What followed was like a fireworks display. With
Pudovkin, Alexandrov, Dziga-Vertov, Kuleshov, Bela Balazs, Renato
May, Rudolf Arnheim, Raymond J. Spouiswoode, Andre Levinson,
Abel Cance, Jean Epstein-and how many others?-montage, through
the enthusiastic and ingenious exploitation of all its combinations,
through the pages and pages of panegyric in books and reviews, be-
came practically synonymous with the cinema itself.
More direct than his fellows, Pudovkin was unwittingly close to
the truth when he declared with aplomb
3
that the notion of montage,
above and beyond all the specific meanings it is sometimes given
(end-to-end joining, accelerated montage, purely rhythmic principle,
etc.), is in reality the sum of filmic creation: The isolated shot is not
eyen a small fragment of cinema: it is onh' ra\\' materiaL a photo-
graph of the real world. Onh' bY montage can one pass from photog-
r.lphy III ,'inClll.1, fTll111 stl\'ish l'l'PY tl) .Ht. Broadly defined, mOntage
is quite simply inseparable from the composition of the \\'ork itself.
4
F,)T llhldnn ..1 LlI1.1ti,'ism. ;15 it must indeed be called, of
man frl'l11 Eisellstei n's : hel,rericJ] \\'orks, Film
1'. ....; .In,) r:, !::": :"':", ,':--5":11.'5 du:. "bsessed by
tkll singk ide'1. the SO\'iet filmm.1ker S.1\\' montage e\'ery\\'here and
extended its boundaries disproportionateh.
5
The histories of litera-
ture and painting. pressed en masse into the sen'ice of Eisenstein's
theon', are used to furnish precursi\'e examples of montage. It was
enough that Dickens, Leonardo da Vinci, or any number of others
combined two themes, two ideas, or two colors for Eisenstein to dis-
cover montage; the most obvious pictorial juxtaposition, the most
properly literary effect of composition, were, to hear him, propheti-
cally precinematographic. All is montage. There is something relent-
The word is used as defined by G. Cohen-Seat (efficience): not as t.he
effectiveness of a particular approach or specific act, but as the power pecuhar
to a means of expression.
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 33
less, almost embarrassing at times, in Eisenstein's refusal to admit
even the smallest place for continuous flows of creation; all he can
see anywhere are prefragmented pieces, which ingenious manipula-
tion will then join together. Furthermore, the manner in which he
describes the creative work of all those he enlists as his forerunners
does not fail, in certain truly improbable passages, to contradict even
the slightest likelihood of any psychogenesis of creation.
6
In the same way Eisenstein categorically refused to admit any kind
of descriptive realism into the cinema. He would not accept that one
could film a scene continuously, and he was full of scorn for what
he called, depending on the passage, "naturalism," "purely objective
representation," or narrative that is simply "informative" (as distin-
guished from "pathetiC" or "organic," that is, in the final analysis,
edited into montage sequence). He would not even consider that the
continuous recording of a short scene that was itself composed and
acted out could be a choice among others. No, one must fragment,
isolate the close-ups, and then reassemble everything. Could not the
filmed spectacle have its own beauty? One dare not say so. As if for-
ever to reassure himself, that great artist whose genius and glory
could have encouraged him a thousand times, manages to have
beauty, which has been pitilessly rejected from every "profilmic"'"
occasion, emerge without possible confusion from the filming, and
only from the filming. Even more: from the montage, and only
from the montage. For, on the level of each shot, there is filming,
therefore composition. But Eisenstein never loses an opportunity to
devalue, in favor of his concern for sequential arrangement, any ele-
ment of art that might have intruded into the forming of the ordered
segments. ". ".
As defined by Etienne Souriau. Whatever is placed in front of the camera,
or whatever one shoots with the camera is "profilmic."
U We know that in his "last period" (Alexander Nevsky and Ivan the Ter-
rible), Eisenstein was governed by a very different aesthetic, an aesthetic of the
image much more than of montage. But this development has left only a faint
trace in his theoretical writings, which concern us here. However, the unpub-
lished manuscripts, which will be gradually published, will have to be seen.
34 PROBLEMS OF FILM SEMIOTICS
THE SPIRIT OF MANIPULATION. A comparison suggests itself-
and deserves more attention than the brief remarks that follow-be-
tween that obsession with breakdown analysis and montage and
certain tendencies of the "modern" spirit and civilization. In its mo-
ments of excess, when inspiration would desert it, montage cinema
(other than in Eisenstein's films) came at times very close to being a
kind of mechanical toy-in a world, it is true, where erector sets are
not the only syntagmatic toys to captivate our children, who acquire
a taste for manipulation in their playing, which, if they later become
engineers, specialists in cybernetics, cven cthnographcrs and linguists,
may be extended into a whole operational attitude, whose excellence
of principle will be more evident here than in film. And, surely, we
know anyway that an age is not defined by the state of mind of a
few, just as "montage or bust" does not define all of the cinema. One
man may be a cybernetics specialist, but another a farmer or a jani-
tor; one film may be all montage, but another may unfold in large
sections. A period, however, is shaped by all of its activities. If one
chooses to emphasize one aspect of it, one is too often criticized for
having at the same time neglected the others; the lack of ubiquity
becomes a sin against intelligence. Nevertheless we must renounce
discussing, along with our subject, all that is not our subject.
At the time of Citizen Kane, Orson Welles, whom the RKO pro-
ducers had given an unusual freedom of means, would go into rap-
tures, according to his biographer,
7
at all the apparatus he had been
made master of: 'That was the finest electric toy ever given to a
boy." Erector set, electric train: both montage toys. Department
stores sell electric trains in separate pieces: A new package of tracks
acquired later on allows a small boy to reassemble his old switching
in a new way; everything fits together. Catalogues list the different
pieces one can buy (classifying them according to their functions in
the over-all scheme): "right switch," "left switch," "ninety-degree
angle crossing," "a twenty-two-degree angle crossing." You would
think they were the parts of discourse according to Kuryglowicz, or
a "text" spewed out by some American fanatic about distributional
analysis. Still, toys are only an amusing example. There is also "pho-
THE CINEMA: LANGUACE OR LANGUAGE SYSTEM? 35
tomontage," collage, paper cut-outs in the animated films of Boro-
wizyk and Lenica, or certain "experimental" shows by the research
teams of ORTF (French national television). Above all there are
cybernetics and the theory of information, which has outdone even
the most structuralist of linguistics: Human language is already
fairly organized, much more in any case than many other code sys-
temS, such as the rules of politeness, art, custom; verbal language is
enriched by a sufficiently rigorous paradigmatic category permitting
the most varied syntagmatic arrangements. But, in the eyes of cer-
tain modern tendencies, it still bears too much "substance" and is not
entirely organizable. Its double substantiality, phonemic and seman-
tic (that is to say, twice human, through the body and the mind),
resists exhaustive pigeon-holing. Thus the language we speak has
become-very paradoxically if one thinks about it-what certain
American logicians call "natural" or "ordinary" language, whereas
no adjective is needed to describe the languages of their machines,
which are more perfectly binary then the best analyses of Roman
Jakobson. The machine has ground up human language and dis-
penses it in clean slices, to which no flesh clings. Those "binary
digits," perfect segments, have only to be assembled (programed) in
the requisite order. The code triumphs and attains its perfection
in the transmission of the message. It is a great feast for the syntag-
matic mentality.
There are other examples. An artificial limb is to the leg as the
cybernetic message is to the human sentence. And why not mention
-for amusement and a change from erector sets-powdered milk and
instant coffee? And robots of all kinds? The linguistic machine, at
the forefront of so many modern preoccupations, remains neverthe-
less the privileged example.
The attitude that conceives and produces all these products is
broadly the same: The natural object (whether human language or
cow's milk) is considered as a simple point of departure. It is ana-
lyzed, literally and figuratively, and its constituent parts are isolated;
this is the moment of breakdown analysis, as in the cinema. Then
the parts are distributed into isofunctional categories: 8 straight tracks
36 PROBLEMS OF FILM SEMIOTICS
to one side, curved to the other. This is the paradigmatic aspect-
and it is only preparatory, as was the filming of individual scenes for
Eisenstein. The grand moment, which one has been waiting for and
thinking about since the beginning, is the syntagmatic moment. One
reassembles a duplicate of the original object, a duplicate which is
perfectly grasped by the mind, since it is a pure product of the mind.
It is the intelligibility of the object that is itself made into an object.
And one never takes into account that the natural object has been
used as a model. Quite the contrary, the assembled object is taken
as the model object-and let the natural object keep still! Thus the
linguist'" will try to apply the conditions of the theory of information
to human language, while the ethnographer will call "model" not
the reality he has studied but the formalization he has derived from
that reality. Claude Levi-Strauss is especially clear on this point.
9
The difference between the natural object and its reconstructed
model is insisted upon, but somehow it is neutralized; the optional or
individual variations of articulation in phonemics, for example, are
"nonrelevant." The goal of the reconstruction, as Roland Barthes
emphasizes, is not to reproduce reality; the reconstruction is not a
reproduction, it does not attempt to imitate the concrete aspect of the
original object; it is neither poiesis nor pseudo-physis, but a simula-
tion, a product of techne.
10
That is to say: the result of a manipula-
tion. As the structural skeleton of the object made into a second
object, it remains a kind of prosthesis.
What Eisenstein wanted to do, what he dreamed of perpetually,
was to make the lesson of events visually apparent, and through
breakdown analysis and montage to make it itself an appreciable
event. From this comes his horror of "naturalism." To Rossellini, who
said 'Things are. Why manipulate them?" the Soviet film-maker
might have replied, "Things are. They must be manipulated." Eisen-
stein never shows us the course of real events, but always, as he says,
the course of real events refracted through an ideological point of
... Who is, incidentally, reticent at times and of two minds about these prob-
lems (consider Andre Martinet's attitude). Others, like P. Guiraud or Roman
Jakobson, are more positive.
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 37
view, entirely thought out, signifying from beginning to end. Mean-
ing is not sufficient; there must also be signification.
Let there be no misunderstanding; this is not a matter of politics.
It is not a question of opposing Eisenstein's political options to some
kind of objectivity; nor is it a question of opposing to his purely nar-
rative (nonpolitical) "prejudices" the possibility of some other read-
ing, direct and mysteriously faithful to the deeper meaning of things,
as Andre Bazin, more subtle than those who find fault with Eisen-
stein for being a Communist, has doneY It is a question only of semi-
otics: What we call the "meaning" of the event narrated by the
film-maker would in any case have a meaning fOT someone (since no
others exist). But from the point of view of the means of expression,
one can distinguish between the "natural" meaning of things and
beings (which is continuous, total, and without distinct signifiers: the
expression of joy on the face of a child) and determined signification.
The latter would be inconceivable if we did not live in a world of
meaning; it is conceivable only as a distinct organizational act by
which meaning is reorganized: Signification tends to make precise
slices of discontinuous significates corresponding to so many discrete
signifiers. By definition it consists in informing an amorphous seman-
ticism. In Potemkin, three different lion statues filmed separately be-
come, when placed in sequence, a magnificent syntagma; the stone
animal seems to be rising and is supposed to yield an unequivocal
symbol of the workers' revolt. It was not enough for Eisenstein to
have composed a splendid sequence; he meant it to be, in addition,
a fact of language (fait de langue).
How far can the taste for manipulation, one of the three forms
of what Roland Barthes calls "sign imagination" ("l'imagination du
signe") gO?12 Does not Moles anticipate a "permutational art" in
which poetry, discarding the chaste mystery of inspiration, will openly
reveal the portion of manipulation it has always contained, and will
finally address itself to computers? The "poet" would program the
machine, giving it a certain number of elements and setting limita-
tions; the machine would then explore all the possible combinations,
13
and the author would, at the end of the process, make his selection.
3
8
PROBLEMS OF FILM SEMIOTICS
A utopia? Or prophecy? Moles does not say it will COme tomorrow,
and there is no reason why one cannot extrapolate from the premises
of today. Prophecies are rarely fulfilled in their predicted forms, but
they are nevertheless indicative. This essay springs from the convic-
tion that the "montage or bust" approach is not a fruitful path for
film (nor, for that matter, poetry)' But it should be seen that that
orientation is entirely consistent with a certain spirit of modernism,
which, when called cybernetics or structural science, yields results
that are much less questionable.
Thus, up to a certain point, "montage or bust" partakes of a state
of mind peculiar to "structural man."14 But no SOoner is it established
than this similarity must be qualified by two reservations, which
might actually be only one. First is the point that the height of mon-
tage came well before the flood tide of the syntagmatic mind. The
latter began really only after World War II, just at the time when
"montage or bust" (1925-30) was being more and more criticized
and rejected by film-makers and theoreticians. IS Second, is it not
paradoxical that the cinema should have been one of the areas in
which the spirit of manipulation began its career? Is not the notion
of a reconstructed reality, which seeks no literal resemblance, con-
trary to film's essential vocation? And is not the camera's role to re-
store to the viewer the object in its perceptual quasi-literalness, even
if what it is made to film is only a preselected fragment of a total sit-
uation? Is not the close-up itself-the absolute weapon of the theo-
reticians of montage in their war against visual naturalism-on a
smaller scale, as respectful of the object's aspect as is the establishing
shot?"" Is not film the triumph of the "pseudo-physis" that the spirit of
manipulation precisely rejects? Is it not entirely based on that famous
"impression of reality," which no one challenges, which many have
studied, and to which it owes its more "realistic" moments as well as
its ability to realize the fantastic?I6
In fact these two reservations are merely one: In the period when
,. In fact, though, not entirely; isolated and magnified, the fragment is
OCcasionally unrecognizable. This has been pointed out already, with reason.
But this nuance, which will have to be studied separately, may be temporarily
neglected in a global study.
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 39
a certain form of intellect agent becomes more aware and sure of it-
self, it is natural that it should tend to abandon areas, such as the
cinema, that restricted it, and should gather its forces elsewhere.
Conversely, in order for it to exert a strong influence on the cinema,
this mentality had to be as yet in its beginnings and not yet estab-
lished.""
From "Cine-Langue" to Cinema Language
The preceding, somewhat cavalierlf.lf. view was not meant to explain
everything. It was put forward as a hypothesis. The observation con-
cluding it is negative and doubly so: Film did not lend itself well to
manipulation, and the spirit of manipulation itself was not under-
stood well. There remains a positive fact to account for: The cinema
-which is not the only field in which manipulation may be, how-
ever imperfectly, conceived-the cinema, in preference to other fields,
was chosen (and with what fanaticism!) by certain theoreticians of
manipulation.
During the same period, to be sure, montage was being affirmed
elsewhere than on the screen: in mechanical arts, in engineering tech-
niques, and in the constructivist theater. Eisenstein was educated at
the School for Public Works of St. Petersburg before I9Ij. He him-
self declared that his theory of the "montage of attractions" had been
suggested to him by the assembling of tubular parts in engineering,
as well as by the techniques of juxtaposition used in circuses and
music hallsY He was actively engaged in the constructivist move-
ment of the young Soviet theater. He was an admirer of the Kabuki
,. It goes without saying that I am speaking here about cinematographic
creation. For the author of this book is obviously in no position to maintain
that syntagmatic methods are not suited to the analysis of film. This is some-
what like the problem of the "creator" and the "theoretician"-despite their
closeness in the modern period, which, though it has often been pointed out,
is still insufficient-who necessarily approach the same object by means that
are so different that they are not always able to resist the pessimistic and over-
simplifying illusion that it is not the same object.
,.,. And which strikes me as being even more cavalier today than when this
article was written.
40 PROBLEMS OF FILM SEMIOTICS
theater, which he considered to be a pure product of montage. He
wrote for Lef, Mayakovsky's review, staged Tretyakov, worked for
the Prolecul t (popular theater), for the Free Experimental Theater,
for Meyerhold's theater, etc. But none of these influences and conta-
gions
18
can exempt a truly cinematographic study from examining
the factors that, in the nature of film itself, could produce-even if
by some quasi-misunderstanding-the specifically filmic designs of
the manipulative mind.
For the error was tempting: Seen from a certain angle, the cinema
has all the appearances of what it is not. JI. It is apparently a kind of
language (une sorte de langage), but it was seen as something less, a
specific language system (une langue). H It allows, it even necessi-
tates, a certain amount of cutting and montage; its organization,
which is so manifestly syntagmatic, could only be derived, one be-
lieved, from some embedded paradigmatic category, even if this para-
digmatic category was hardly known. Film is too obviously a message
for one not to assume that it is coded.
For that matter, any message, provided it is repeated often enough
and with a sufficient number of variations-as is the case with film
-becomes in time like a great river whose channels are forever shift-
~ Despite the clumsy formulations of a man who was partly self-taught,
which are scattered throughout his books (though not in his films), Eisenstein
remains, to my mind, one of the greatest film theoreticians. His writings are
crammed full of ideas. His thoughts on language systems (in spite of an ex-
uberant and somewhat unfocused vocabulary), however, will have to be re-
stated in terms of language.
U A language system (langue) is a highly organized code. Language itself
covers a much broader area: de Saussure said that language is the sum of
langue plus speech (parole). The concept of a "fact of language" (fait de
langage) in Charles Bally or ~ m l Benveniste points in the same direction.
If we were to define things and not words, we might say that language, in its
broadest reality, is manifest every time something is said with the intention of
saying it (see Charles Bally, "Qu'est-ce-qu'un signe?" in Journal de psycholo
gie normale et pathologique, Paris, vol. 36, 1939, nos. 3 and 4, pp. 161-74,
and especially p. 165). Of course, the distinction between verbal language
(language proper) and other "sign systems" (semes) (sometimes called "lan-
guages in the figurative sense") comes to mind, but it must not confuse the
issue. It is natural that semiotics be concerned with all "languages," without
pre-establishing the extent and the limits of the semic domain. Semiotics can
and must depend heavily on linguistics, but it must not be confused with
linguistics.
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 41
ing, depositing here and there along its course a string of islands:
the disjointed elements of at least a partial code. Perhaps these is-
lands, barely distinguishable from the surrounding flood, are too
fragile and scattered to resist the sweep of the current that gave
them birth and to which they will always be vulnerable. Neverthe-
less there are certain "syntactical procedures" that, after frequent use
as speech, come to appear in later films as a language system: They
have become conventional to a degree. Many people, misled by a
kind of reverse anticipation, have antedated the language system;
they believed they could understand the film because of its syntax,
whereas one understands the syntax because one has understood,
and only because one has understood, the film. The inherent intelli-
gibility of a dissolve or a double exposure cannot clarify the plot of a
film unless the spectator has already seen other films in which dis-
solves and double exposures were used intelligibly. On the other
hand, the narrative force of a plot, which will always be understood
only too well-since it communicates with us in images of the world
and of ourselves-will automatically lead us to understand the dou-
ble exposure and the dissolve, if not in the first film we see them,
at least by the third or fourth. As Gilbert Cohen-Seat has said, the
language of film will always have the advantage of being "already
entirely written out in actions and in passions important to US."19 All
experiments on filmic intellection tend to prove this. The works of
B. and R. Zazzo, Ombredane, Maddison, Van Bever, Mialaret, Me-
lies, Lajeunesse and Rossi, Rebeillard, etc. all share the idea that it
is only those syntactical procedures that have become too convention-
alized that cause difficulties of understanding among children or
primitive subjects, unless the film's plot and its world of diegesis,
which remain understandable in the absence of such procedures, are
able to make the procedures themselves understandable.
After these digressions let us now return to Rossellini's interview.
''Things are," he had said. 'Why manipulate them?" Obviously he
was not referring to the techniques of organization in their broadest
sense, but only and explicitly to the actual cinematographic theory
of "montage or bust." Thus he was echoing (to the utmost joy of
4
2
PROBLEMS OF FILM SEMIOTICS
Cahiers du Cinema, which was not interviewing him innocently) a
whole tendency that the Cahiers had incubated and of which it was
practically the incarnation. While the Italian director spoke, one
thought of the French film-makers and writers. Was it not in Con-
nection with Rossellini's films'" that Andre Bazin had developed his
famous theories on the sequence shot, depth of field, and continuity
shooting?20 As for all those comrades and friends, were they not all
working toward the same goal, the death of a certain concept of the
cinema: the erector-set film? If the cinema wan ts to be a true lan-
guage, they thought, it must cease to be a caricature. Film has to say
something? Well then, let it say it! But let it say it without feeling
obliged to manipulate images "like words," arranging them according
to the rules of a pseudo-syntax whose necessity seemed less and less
evident to the mature minds of what is called-beyond the narrow
sense of the "new wave"-the "modern" cinema? The days of Dziga
Vertov's "cine phrase" ("film sentence") and "cine langue" ("film
language-system") were gone!2I
Thus Andre Bazin was not alone. There was Roger Leenhardt,22
there was Jean Renoir with his many statements in favor of the se-
quence shot,23 there was-to limit ourselves to film-makers who were
also theoreticians-Alexandre Astruc whose famous "camera stylo"24
was, despite appearances, the exact opposite of the old notion of
cine langue. A pen (stylo) only writes what it is made to write. What
Astruc wanted was a cinema as free, personal, and sharp as some
novels are, but he was careful to explain
25
that his "vocabulary"
would be made up of the very aspect of things, "the impasto of the
world." The "montage or bust" approach, on the contrary, consisted
in dismantling the immanent perception of things in order to reel it
off in slices, which would become simple signs to be used wherever
one pleased. Around the same time, in a work whose title itself made
cinema a language,26 Marcel Martin observed in passing27 that the
reader would not find a strict system of signs. Then, following Mer-
.. And also in relation to Italian neorealism in general, and to certain as-
pects of Orson Welles, William Wyler, Jean Renoir, Erich von Stroheim,
Friederich Murnau, etc.
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 43
leau-Ponty's lecture on "Le Cinema et la nouvelle psychologie,"28 film
began to be defined here and there, or at least approached, from what
one called the "phenomenological" angle: A sequence of film, like a
spectacle from life, carries its meaning within itself. The signifier is
not easily distinguished from the significate. "It is the felicity of art
to show how a thing begins to signify, not by reference to ideas that
are already formed or acquired, but by the temporal and spatial ar-
rangement of elements."29 This is an entirely new concept of order-
ing. The cinema is the "phenomenological" art par excellence, the
signifier is coextensive with the whole of the significate, the spectacle
its own signification, thus short-circuiting the sign itself: This is
what was said, in substance, by Souriau, Soriano, Blanchard, Marcel,
Cohen-Seat, Bazin, Martin, Ayfre, Astre, Cauliez, Dort, Vailland,
Marion, Robbe-Grillet, B. and R. Zazzo, and many others in the
course of one article or another. It is possible, even probable, that
they went too far in this direction: For the cinema is after all not
life; it is a created spectacle.'" But let us put these reservations aside
for the moment, and simply record what was in fact a convergence in
the historical evolution of ideas about film.
Rossellini's remarks, though they may not be very philosophical,
nevertheless point in the same direction. Let us listen to him some
more: The cinema, he says, is a language, if one means by that a
.. In Esthetique et psychologie du cinema (Paris: Editions Universitaries,
vol. I, 1963), Jean Mitry states matters more vigorously: After having been
everything, montage now tends to be nothing, at least in some theories. But
the cinema is inconceivable without a minimum of montage, which is itself
contained within a larger field of language phenomena (pp. IO-II). The
analogy .and the quasi-fusion of the signifier and the significate do not define
all of film, but only one of its constituents-the photographic material-which
is no more than a point of departure. A film is made up of many images,
which derive their meaning in relation to each other in a whole interplay of
reciprocal implications, symbols, ellipses, etc. Thus the signifier and the sig-
nificate are given a greater distance, and so there is indeed a "cinematographic
language" (see especially pp. 119-23). I mention this in order to insist on the
difference between such a "language" and a "film language-system." The parti-
sans of what is called "non-montage" (the Bazin tendency), even if they
have occasionally made statements on the aesthetics of film that are too ex-
clusive, at least can be credited-on the level of a sort of intuitive and spon-
taneous semiotics-with rejecting any concept of the cinema as langue and
affirming the existence of a cinematographic language.
44 PROBLEMS OF FILM SEMIOTICS
"poetic language." But the theoreticians of silent film saw in it a real,
specific vehicle (Rossellini's word) about which we are much more
skeptical today. For the creator of Open City, who is not normally
concerned with semiotics, this was a kind of conclusion. It was ut-
tered somewhat haphazardly (in his choice of terms at any rate),
spontaneously, but in fact with great felicity: It is unusual for a man
of the profession, other than in his films, to suggest so many things
in so few words.
A Non-System Language:
Film NaTTativity
When approaching the cinema from the linguistic point of view, it
is difficult to avoid shuttling back and forth between two positions:
the cinema as a language; the cinema as infinitely different from
verbal language. Perhaps it is impossible to extricate oneself from
this dilemma with impunity.
After analyzing the "logomorphic" nature
30
of film, Gilbert Cohen-
Seat came to the provisional conclusion that one must at least over-
come the temptation to consider the cinema as a language.
31
Film
tells us continuous stories; it "says" things that could be conveyed
also in the language of words; yet it says them differently: There is
a reason for the possibility as well as for the necessity of adaptations.
To be sure, it has often been justly remarked
32
that, since film has
taken the narrative road-or, what F. Ricci calls the "novel's way"
(voie romanesque) 33-since the feature-length fiction film, which
was only one of many conceivable genres, has taken over the greater
part of total cinematic production, it could only be the result of a
positive development in the history of film, and particularly in the
evolution from Lumiere to Melies, from "cinematography" to cin-
ema.
34
There was nothing unavoidable, or particularly natural, in this.
Yet even those \\' ho emphasize the historical aspect of this growth
never conclude that it \vas meaningless or haphazard. It had to hap-
pen. but it had to happen for a reaSlm: it had to be that the wrv na-
ture of the cinema rendered such an evolution if not certain at least
probable.
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 45
There was the demand, what the spectator wanted. This is the
main idea in Edgar Morin's analyses (which it would be superfluous
to repeat with less talent). Although the spectatorial demand cannot
mould the particular content of each film, Morin recently explained,
it is perfectly capable of determining what one might call the spec-
tacle's formula. 35 The ninety-minute show with its digressions (doc-
umentary, etc.) of lesser narrativity is one formula. It will perhaps
not endure, but for the moment it is sufficiently pleasing; it is ac-
cepted. There have been other formulas, for example two "great
movies" in a single showing. But these are all variations. The basic
formula, which has never changed, is the one that consists in making
a large continuous unit that tells a story and calling it a "movie."
"Going to the movies" is going to see this type of story.
Indeed the cinema is eminently apt to assume this role; even the
greatest demand could not have diverted it in any lasting way along
a path that its inner semiological mechanism would have made im-
probable. Things could never have occurred so fast nor remained in
the state we still find them, had the film not been a supreme story-teller
and had its narrativity not been endowed with the nine lives of a cat.
The total invasion of the cinema by novelesque fiction is a peculiar,
striking phenomenon, when one thinks that film could have found
so many other possible uses which have hardly been explored by a
society that is nevertheless forever in pursuit of technographic nOV-
elty.
The rule of the "story" is so powerful that the image, which is said
to be the major constituent of film, vanishes behind the plot it has
woven-if we are to believe some analyses
36
-s
o
that the cinema is
only in theory the art of images. Film, which by nature one would
think adapted to a transversal reading, through the leisurely investi-
gation of the visual content of each shot, becomes almost immediately
the subject of a longitudinal reading, which is precipitous, "anxious,"
and concerned only with "what's next." The sequence does not string
the individual shots; it suppresses them. Experiments on the memory
retention of film-whether by Bruce, Fraisse and de Monmollin, Re-
beillard, or Romano and Botson-all come to the same conclusion,
7. r ... .,
46 PROBLEMS OF FILM SEMIOTICS
though by different means: All one retains of a film is its plot and a
few images. Daily experience confirms this, except of course for those
who, seeing very few films, retain them entirely (the child's first
movie; the farmer's one film of the year; and even then). On an en-
tirely different level, Dreyfus
37
observes that the attempts to create
a certain "language" in a type of modern film (Antonioni, Godard,
etc.) occasionally produce a disturbing, however talented, overabun-
dance, since the film track itself is already and always saying some-
thing.
Narrativity and logomorphism. It is as if a kind of induction cur-
rent
38
were linking images among themselves, whatever one did, as
if the human mind (the spectator's as well as the film-maker's) were
incapable of not making a connection between two successive images.
Still photography-a close relative to film, or else some very old
and very distant second cousin-was never intended to tell stories.
Whenever it does, it is imitating the cinema, by spatially deploying
the successivity that film unfolds in time. The eye proceeds down
the page of the "picture romance" magazine in the prescribed order
of the photographs-which is the order they would have unreeled in
on the screen. "Picture romances" are frequently used to retell the
story of an existing film. This is the consequence of a deep similarity,
which is in turn the result of a fundamental dissimilarity: Photog-
raphy is so ill-suited to story-telling that when it wants to do so it
becomes cinema. "Picture romances" are derived not from photogra-
phy but from film. An isolated photograph can of course tell nothing!
Yet why must it be that, by some strange correlation, two juxtaposed
photographs must tell something? Going from one image to two im-
ages, is to go from image to language.
Kuleshov's experiments, as I said earlier, were considered for many
years the "scientific" basis for the supremacy of montage. No one,
however, has paid sufficient attention to the fact that, in the midst of
the age of "montage or bust," there existed another interpretation of
those famous experiments. An interpretation that seemed to provide
yet more ammunition for the partisans of manipulation, but in fact
implied a covertly discordant position (more modestly than it should
. $ ---- -
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 47
have), which only the future was to bring to light. It was contained
in Bela Balazs's book Der Geist des Films (1930).39 With a kind of
shrewdness peculiar to him, the Hungarian theoretician remarked
that, if montage was indeed sovereign, it was so by necessity, for,
when two images were juxtaposed purely by chance, the viewer
would discover a "connection." That, and nothing else, is what Kule-
shov's experiments demonstrated. Film-makers obviously understood
this and decided that this "connection" would be their tool, to manip-
ulate according to their will. Yet right from the beginning their hand
was forced by the viewer, or rather, by a certain structure of the hu-
man mind, that obdurate diachronist. Listen to Balazs: "One presup-
poses an intention. . . . The viewer understands what he thinks
montage wants him to understand. Images . . . are . . . linked to-
gether . . . internally through the inevitable induction of a current
of signification.... The power [of montage] exists and is exerted,
whether one wants it or not. It must be conscientiously used." In his
Esthetique et psychologie du cinema Jean Mitry elaborates in much
greater detail an interpretation of the "Kuleshov effect" (pp. 283-
85). He concludes that the famous experiments in no way authorize
the theory of "montage or bust" (according to which the diegesis is
marginal to the development of montage effects, which tend to pro-
duce an abstract logic, or piece of eloquence, independent of the film
itself). They simply demonstrate the existence of a "logic of implica-
tion," thanks to which the image becomes language, and which is in-
separable from the film's narrativity.
Thus, film montage-whether its role is the triumphant one of
yesterday or the more modest one of today-and film narrativity-as
triumphant now as it was in the past-are only the consequences of
that current of induction that refuses not to flow whenever two poles
are brought sufficiently close together, and occasionally even when
they are quite far apart. The cinema is language, above and beyond
any particular effect of montage. It is not because the cinema is lan-
guage that it can tell such fine stories, but rather it has become lan-
guage because it has told such fine stories.
Among the theoreticians and film-makers who have moved the
- - -
48 PROBLEMS OF FILM SEMIOTICS
cinema away from the spectacle to bring it closer to a novelistic
"writing" capable of expressing everything-its author as well as the
world-of repeating and sometimes replacing the novel in the multi-
ple task it had assumed since the nineteenth century,40 we find pre-
cisely, and by no accident, many of those who are the least concerned
with "cinematographic syntax" and who have said so, not without
talent at times, in their articles (Bazin, Leenhardt, Astruc, Truffaut)
or have shown so in their films (Antonioni, Visconti, Godard, Truf-
faut). There are individual cases, of course: With Alain Resnais,
with Jean-Luc Godard, there is a body of montage which reappears,
but with a new meaning; the genius Orson Welles makes films be-
yond any restriction-a master of visual surprise, he can use, when
necessary, continuous camera shots as involved and involving as a
sentence of Marcel Proust's. But, whereas individual style is one
thing, the evolution of cinematographic language is another, differ-
ing from it, not in substance (for it is the film-makers who make up
the cinema), but in scale and in one's approach to it. One would
need forty chapters to do justice to the first, whereas, for the latter,
two suffice, at least for the present: cine langue (cinema language
system) and cinema language. I have mentioned Antonioni, Visconti,
Godard, and Truffaut because, of the directors having a style, they
seem to me to belong among those, furthermore, in whom one can
most clearly see the change from the will to system to the desire for
language." They frequently use the sequence shot where the parti-
sans of montage would have cut and reassembled; they fall back on
what is called, for better or worse, the "tracking shot" (and which
implies nothing other than a noncodified mobility of the camera, a
movement that is truly free)"" where traditional film syntaxes distin-
guish between rear and forward "dolly," "pan" and "tilt," etc.
41
Thus language is enriched by whatever is lost to system. The two
phenomena are one. It is as if the code's signifying abundance were
linked to that of the message in the cinema-or rather separated from
.. This text was written in the beginning of 1964: today there would be
man; names to add to it.
.. A very successful example: the sequence of the travel agency in Breathless
by Jean-Luc Godard.
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 49
it-by an obscurely rigorous relationship of inverse proportions: Code,
when it is present, is crude-the great film-makers who believed in it
were great despite it; the message, as it becomes refined, circumvents
the code. At any given moment, the code could change or disappear
entirely, whereas the message will simply find the means to express
itself differently."
"CineLangue"andVerbalLanguages:
TheParadoxof theTalkingMovies
In the period when the cinema considered itself a veritable language
system, its attitude toward verbal languages was one of utmost dis-
dain. It feared their competition, which had come about only by its
intrusion into their midst. One might think that before 193
0
the
very silence of film would have given it automatic protection against
the detested verbal element. Like the deaf who sleep peacefully un-
disturbed by any noise, the silent cinema, deriving strength from its
weakness, would lead, one might think, a still and tranquil life. But
not at all! No period was ever more noisy than that of the silent
movies. Manifestoes, vociferations, invectives, proclamations, and
vaticinations succeeded each other in denouncing always the same,
ghostlike enemy: speech-speech, which was radically absent-neces-
sarily so-from film itself and which existed finally (as victim and
executioner) only in the speeches delivered against it. The young
Jean Epstein; the young Rene Clair; Louis Delluc (who died young);
the cohorts of the "pure cinema" with their impetuous Egeria, Ger-
maine Dulac; Bela Balazs; Charlie Chaplin; and naturally the legion
in close formation of Soviet pioneers: They were all full of contempt
for the word. And I have mentioned only the loudest.
.. Today I would no longer state the relationship between the code and
the message in such strictly antagonistic termS. It now appears to me that the
realities codes possess are more complex, more various, subtler-and therefore
more compatible, so to speak, with the richness of messages. On these points,
see the whole of Chapter 5, "Problems of Denotation in the Fiction Film," liS
well as passages from "The Modern Cinema and Narrativity." See also "Prob-
Iemes actuels de theorie du cinema," (not reprinted here), Revue d'esthetique
(special issue: "Le cinema") vol. xx, no. 2-3 (especially p. 221).
50 PROBLEMS OF FILM SEMIOTICS
Of course it is easy to smile at them. Yet there was more truth in
their paradoxical anathemas than at first appears. The old verbal
structures, although officially absent from film, were nevertheless a
haunting presence. The attack was not without an object: Obviously
there were the explanatory titles; above all there was the gesticulation
in acting, whose true reason for being-we will have to return to this
-was not, as has been wrongly said, in the infirmity of the silent
image, or in acting habits mechanically inherited from the theater
(how to explain that in some silent pictures there is no gesticula-
tion?), but in a subconscious attempt to speak without words, and
to say without verbal language not only what one would have said
with'it (which is never entirely impossible), but in the same way it
would have been said. Thus there came into being a kind of silent
gibberish, simultaneously overexcited and petrified, an exuberant
gabbling whose every gesture, every bit of mimicry, stood with scru-
pulous and clumsy literalness for a linguistic unit, almost always a
sentence whose absence, which would not otherwise have been cata-
strophic, became abundantly obvious when the gesticulated imitation
so clearly emphasized it. Yet it was sufficient for a Stroheim""-re-
duced like his peers to silent pictures, and as anxious as they to ex-
press a great deal despite this limitation-to circumvent speech (in-
stead of attacking it head on, while at the same time shamefully
copying it) for film to be enriched and to become quieter, for the pre-
viously clumsily localized significations to become more continuous
and to reveal a full, complex meaning.
Those for whom the silent film still spoke too much were thus not
entirely wrong. They anticipated many truths; however, it was in the
steps of a much broader development of thought-and one that was
more obscure, and more deeply motivated. They were almost afraid
of verbal language, for even as they were defining the cinema as a
nonverbal language, they were still obscurely thinking of some
pseudo-verbal system within their films. Obscurely, yet clearly
"" Think of the magnificent seduction scene in Wedding March (1927),
which was entirely constructed from the imperceptible facial play of the
actor-auteur. Not a gesture, but what expressiveness!
~
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 51
enough for them to see the language of words as a powerful rival
forever on the point of overstepping its bounds. An analysis of the the-
oretical writings of the period would easily show a surprising conver-
gence of concepts: The image is like a word, the sequence like a sen-
tence, for a sequence is made up of images like a sentence of words,
etc. By assuming this position, the cinema, for all its proclamations of
superiority, was condemning itself to perpetual inferiority. In con-
trast to a e ~ n e language (verbal language), it defined itself unwit-
tingly as a coarser duplication. It only remained for it to sport its
plebeian condition (many of Marcel L'Herbier's articles seem to have
no other aim), in secret fear of a more distinguished older brother.
One can see how the paradox of the talking cinema was already
rooted at the heart of the silent movies. But the greatest paradox was
yet to corne: The advent of the talking movies, which should have
changed not only the films but the theories about them, in fact modi-
fied the latter in no way, at least during the first several years. Films
talked, and yet one spoke about them as if they were silent. But there
was an exception: A profoundly new tendency, unjustly scorned
until Bazin began to rehabilitate it,42 was being developed in the
writings of Marcel Pagnol. This tendency carne from outside the
cinema, and its roots were not buried in the problems of the silent
film (which is what aroused the fury of its enemies). Very signifi-
cantly, it began to appear only after the arrival of the talking
movies.
43
It was able to avoid what I have called the paradox of the
talking cinema.
The introduction of speech into the cinema did not substantially
modify the attendant theoretical positions. It is known that not a few
lovers of cinematographic purity hesitated before accepting this new-
comer to the filmic world. They gave their word here and there that
they would not use it, or that they would use it as little as possible,
and in any case never realistically; furthermore, the fashion would
soon pass. There was also an attempt-both a diversion and a regres-
sion-which consisted in opposing "sound" to "talking." Noise and
music were accepted but not speech, which, of all the sounds in the
world, in whose midst it nevertheless exists, was kept-in theory-
52 PROBLEMS OF FILM SEMIOTICS
under a mysterious and specific taboo. This produced the "wordless
talking film, the shy film, the film garnished with the creakings of
doors and the tinkling of spoons; the moaning, shouting, laughing,
sighing, crying, but never the talking film," as the Provenc;al dram-
atist expressed with such verve.
44
But all this was only episodic.
The great debate on the statutory admission of speech into film was
largely platonic, and one was hardly aware of it other than as some-
thing that occupied a large place in the theory of film. The films
themselves were already talking: They began to do so very rapidly,
almost all of them did so, and continued to do so.
Not the least part of the paradox is, in fact, the ease with which
speech did in fact find its way into the films of all those whose pro-
nouncements had indissolubly linked the survival of the art of film
to the permanence of its silence.
45
Accepted in practice, speech was
not accepted in theory. One persisted in explaining that nothing es-
sential had changed-hardly a reasonable position!-and that the laws
of cinematographic language remained what they had been in the
past. Arnoux summarizes an opinion that was very common at the time
when he affirrns
46
that the good talking films were good for the
same reason that the good silent films were, and that the "talkie" was
after all only one of many technical improvements, less important all
in all than the close-up shot, which had been "invented" long before.
From our greater distance in time, we can only be amazed at this
obstinate refusal to see the arrival of speech as an occurrence of capi-
tal importance-at any rate an occurrence deserving to be given its
place in theory, and consequently to displace older, accepted ele-
ments from their respective positions. This is the condition for any
true change, as we know-linguists have shown this to be the case in
diachronics, and Proust has demonstrated it with human feelings.
But speech was simply added (if even that) to the theory of cinema,
as if it were in excess and there were no more room for it-and this
at the very time when the silent cinema (not to mention the "sound"
cinema, that voluntary invalid, that stillborn child) was entirely dis-
appearing from the screens.
This unwillingness to see, or rather to hear, is even to be found,
THE CINP.MA: LANGUAGE OR LANGUAGE SYSTEM? 53
though in a less sterile and caricatural form, among those who had
the richest and most fruitful reaction to the birth of the talking film
(Pagnol excepted). In the "Manifesto for Orchestral Counterpoint,"47
Eisenstein, Alexandrov, and Pudovkin generously welcome the
sound-track, in the absence of speech itself. Their attitude is positive.
For them it is a matter of giving visual counterpoint an auditory di-
mension, of multiplying the old cinema by the new. However-and
precisely because a healthy and intelligent reaction makes us regret
its omission, since it provides them with such a rich frame-one no-
tices that nowhere do the three Soviet directors take speech into con-
sideration. For them the sound cinema is a cinema that is squared-
multiplied by itself, and only by itself. The authors of the "Mani-
festo" were thinking of noise and music; for them a film remained an
uttered discourse. They would not consider, they rejected, the idea
that an uttering element, speech, could be inserted into film.
We should not blame them: It was more difficult to approach these
subjects in 1928 than in 1964. But we can take advantage of our
greater distance to note that the appearance of speech in film in-
evitably brought the cinema closer to the theater-contrary to an
opinion too often held, and in spite of the numerous analyses (often
correct in their own terms) that since 1930 have underlined the dif-
ferences between cinematographic and theatrical speech. For the
most part these analyses point to the same conclusion: In their dif-
ferent ways they all suggest that in the theater the word is sovereign,
a constituent of the representation, whereas, in film, speech is gov-
erned and constituted by the diegesis. This is an important difference
one can hardly challenge. In a deeper sense, nevertheless, any utter-
ance, whether governing or governed, by nature tells us something
first, whereas an image, or noise or music, even when it is "telling"
us a great deal, must first be produced.
There is no way in which the division into active and passive as-
pects can occur in the same way for verbal language, which has al-
ways been bound to the human agent and to determined significa-
tion, and for the nonverbal "languages" of the image, of noise or even
of music, which are for their part linked to the impassivity of the
54 PROBLEMS OF FILM SEMIO'I'ICS
world and to the malleability of things; whatever one may say about
it, a film dialogue is never entirely diegetic. Even if we set aside "film
commentaries" (it is enough to note their existence), which would
support us too easily, we observe that the verbal element is never en-
tirely integrated into the film. It sticks out, necessarily. Speech is al-
ways something of a spokesman. It is never altogether in the film,
but always a little ahead of it. The most brilliantly emphatic musical
or pictorial compositions, on the contrary, never intervene between
the film and the spectator; they are experienced as an integral part of
the film: However rich they may be, they remain media.
Despite all that was questionable in his ideas about filmed theater,
Marcel Pagnol was undoubtedly the least mistaken person during the
years from 19
2
7 to 1933, when it was indeed difficult not to be
wrong. There were those who refused to admit sound. Others ac-
cepted it grudgingly. Still others had taken to it enthusiastically.
Some-like the hostess who, wanting to have a great musician to
dinner, invites his chattering wife as well, in the improbable hope
that her dreaded manners might not after all be so dreadful-some,
in a grand gesture of courageous acceptance, even considered letting
a few words be added to the background noise they so highly prized.
Pagnol was almost the only one to accept the talking cinema-that is
to say, a cinema that talked.
After these few words of history, let us now try to define formally
the paradox of the talking cinema, which Pagnol alone was able to
avoid. When the cinema was silent, it was accused of "talking" too
much. When it began to talk, one declared that it was still essen-
tially silent and should remain so. Is there anything surprising, then,
in the affirmation that, with the advent of the "talkie," nothing had
changed? And, in fact, for a certain kind of cinema, nothing had
changed. Before 1930, films chattered Silently (pseudo-verbal gesticu-
lation). After 1930, they were loquaciously silent: A torrent of words
was superimposed over a structure of images which remained faith-
ful to its old laws. Cine-langue never was, and could not become, the
talking cinema. Not in 1930 did films begin talking, but in 1940, Or
thereabouts, when little by little they changed in order to admit
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 55
speech, which had remained waiting at the door, so to speak.'"
But, one will say, it is well known that the first talking films talked
too much. Certainly. But if this is so noticeable, it is because they
did not really talk. Today, on the contrary, we are not shocked by
speech in a film. Perhaps it is used more sparingly? Not always. But,
say it is. In any event, the crux of the matter lies elsewhere: Today
films talk better, and speech no longer surprises us, at least as a gen-
eral rule. Let us make it clear that, as (zlms, they talk better. The
text has not necessarily improved, but it harmonizes better with the
film.
For a cinema that claimed to be language but conceived of itself
as a language system (a universal and "nonconventional" one, to be
sure, but a language system, nevertheless, since it wanted to create
a system that was fairly strict and that logically would precede any
message), the verbal languages could only offer it an unwanted in-
crease and an unseasonable rivalry: No one could think seriously of
inserting, even less of fusing, them into the texture of images-one
could hardly conceive of making them agree with the images. The
cinema began to speak only after it had begun to conceive of itself
as a language that was flexible, never predetermined, sufficiently sure
of itself to do away with a permanent, ill-tempered guard in front of
its own doors, and bountiful enough to be enriched by the wealth
of others. The "sequence shot" did more for talking films than the
advent of the talking cinema itself. As Etienne Souriau remarks in
another context,48 a technical invention can never resolve a problem
in art; it can only state it, so that it can be resolved by a second, prop-
erly aesthetic, invention. This is the well-known dialectic of long-
term progress and short-term regression.
In order to get a better understanding of the talking cinema, one
should study a certain type of "modem" film,49 particularly the work
of the inseparable trio, Alain Resnais, Chris Marker, and Agnes
Varda. In their films, the verbal, even openly "literary," element, is
.. Today I would qualify this statement. There were good talking films right
from the beginning of the talkies; the period 1930-33 was fairly rich. A little
later there were also certain American comedies. Etc. . . .
56 PROBLEMS OF FILM SEMIOTICS
given great weight in the overall composition, which is nevertheless
moreauthentically "filmic" than ever. In Last Yearat Marienbad, the
image and the text play a sort of game of hide-and-seek in which
they give each other passing caresses. The sides are equal: Text be-
comes image, and image turns into text. This interplay of contexts
givesthefilm itspeculiarcontexture.
One is reminded of the famous problem of "cinematographic
specicity" about which Andre Bazin, in almost everyone of his ar-
ticles, makes at least one or two very enlightcning statements. The
overtself-affirmation of a strongpersonality is not always the trait of
those who possess the strongest personalities: And thus it was for a
certain typeofcinemainthepast.
A StateofMindandaStage:
CriticismofCineLangue
We forgive the cinema of the past for its excesses, because it has
given us Eisenstein andafew others. But one always excuses genius.
The concept of cine langue constituted a whole theoretical corpus
and must be evaluated as such. The point is important, for the per-
spectiveofthecritic or the historian does notcoincide here with that
ofthe theoretician. Thetheory is full ofsplendid deadends, concepts
that have not survived, but while they were alive, gave us some of
thegreatestmasterpiecesofthescreen.
Moreover, although a kind of "erector-set cinema" did exist, there
neverwas an "erector-set lm." Thecommon trend in many Ims of
the period was enshrined mainly in writings and manifestoes. It was
neverentirely investedin a particular Im, unless perhaps in certain
avant-garde productions on the borderline between normal cinema
andpureexperimentation (seeing these films today onebegins to feel
that the others were indeed normal. Let the critic evaluate the con-
tributionofthese experimental Ims.)
As for the historian, he will rightly observe that only through its
excesses-theoretical as well as practical-could the cinema begin to
gain a consciousness of itself. The cinema as a language system is
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 57
also the birth of the cinema as an art, some time after the entirely
technical invention of the Cinematograph (as Andre Bazin said of
the avant-garde in its strict sense).50 This can be extended to in-
clude a good partof all the cinema of that period. Let the historian
studyeverythingpositive-andthatis alot-in thiscrisisof adolescent
originality.
Inthebroadsenseusedhere, thelanguage-systemcinema produced
asizable portion ofthebest cinema of its period; through it there oc-
curred something that affected both art and language. This is why
I have spoken only about it. But one must not losc sight of the pro-
portions: For each film of the language-system tendency, there were,
as there is always, ten undistinguished films that managed to avoid
the problems of the silent cinema and the paradox of the talking
cinema. Before 1930, the grade-B movie director would go out and
photograph African elephants; after 1930, he would record music-
hall numbers, music andwords included. Neither the presenceor ab-
senceofspeechcouldcramphisstyle.
Therewas anothercinema, too, which was neither the cine langue
nor the ordinary dud. At the height of the "montage or bust" era,
Stroheim and Murnau were heralding the modern cinema. It was a
matter of talent and personality, because their cinema had no theory
and did not at rst get a following. Ideologically, manipulation was
supreme. And this is normal: Stages are made to be skipped by a
talentedminority.""
".. I have said nothing about the period prior to I920, which is neverthe
less of capital importance for the genesis of cinematographic language (es-
pecially D. W. Griffith). But the problems this period poses are foreign to the
intent of these pages, which do not pretend to give a historical account. It is
quite apparent that a director like Feuillade, to take one example out of
many, is exempt from the kind of excesses for which I am blaming the
cine-langue attitude. The question that interests me (language or language
system?) was conceivable only after the lirst theories of film made their ap-
pearance, roughly in I920. Before then, the cinema was too busy being
created. Lumiere invented the Cinematograph; he did not invent "film" as we
know it today (a complex narrative body of considerable magnitude). The
great pioneers before I920 invented the cinema (see Mitry, Esthitique et
psychologie du cinema, op. cit., pp. 267-85 ofvol. I). Before the problems of
semiotics raised here could even have meaning, let alone an object, the cinema
had first to exist, and it had to begin thinking ofitself in terms oftheory.
58 PROBLEMS OF FILM SEMIOTICS
A Seminal Concept: Cinematographic Specificity
As Rossellini said, the cinema is a language of art rather than a spe-
cific vehicle. Born of the fusion of several pre-existing forms of ex
pression, whichretain some of theirown laws (image, speech, music,
andnoise), thecinemawas immediately obliged to compose, inevery
sense of the word. From the very beginning, threatened with extinc-
tion, itbecame an art. Its strength, or its weakness, is that it encom-
passes earliermodes of expression: Some, trulylanguages (theverbal
element), and some, languages only in more or less figurative sense
(music,images, noise).
Nevertheless, these "languages" are not all found on the same
planewithrespect to the cinema: Speech, noise, andmusic were an-
nexed at a latertime, butfilm was born with image discourse. A true
definition of "cinematographic specificity" can therefore only be
made on two levels: that of filmic discourse and that of image
discourse.
As a totality, filmic discourse is specific through its composition.
Resembling truelanguages as it does, film, withitssuperior instancy,
isof necessityprojected"upward"into the sphere ofart-where it re-
verts to a specific language. Thefilm total can only be a languageif
it is already an art.
But within this totality there is an even more specific core, which,
contrary to the other constituent elements of the filmic universe, is
not found in a separate state in otherarts: It is the image discourse.
Theproportionsin its casearereversed, becauseasequenceofimages
is a language first. Perhaps, being too far removed from thelanguage
wespeak, it isonly a figurative language? So beit. Itis nevertheless
a language, in such a way that Rossellini's characterization (which
was obviously directed at film as a whole) cannot be applied to it.
Image discourse is a specific vehicle. It did not exist before the cin-
ema; until 1930 it alone was sufficient to define film. In technical or
medical films it functions exclusively as a vehicle and is not linked
to any attemptatartistic realization. On the other hand, in films of
THE CINEMA: LANGUAGE. OR LANGUAGE. SYSTEM? 59
fiction the language of theimage tends to become an art (within a
vaster sphere of art)-just as verbal language, which has a thousand
utilitarian applications, is abletobecome incantation, poetry, theater,
ornovel.
Thespecific natureoffilm is defined by the presence of a langue
tendingtowardart,withinanartthattendstowardlanguage.
We have twO things, therefore. Not three. There is indeed lan-
guage system, but neither the image discourse nor filmic discourse
are languagesystems. Whether language or art, the image discourse
is an open system, and it is not easily codified, with its nondiscrete
basic units (theimages), its intelligibility (which is too natural), its
lack of distance between the significate and the signifier. Whether
artorlanguage,thecomposedfilm is an even moreopen system,with
its whole sections of meaning directly conveyed to the audience.""
Filmas we know itis not an unstable compound, because its ele-
ments are not incompatible. And they are not incompatiblebecause
none of them are language systems. It is hardly possible to use two
language systems simultaneously: IfI amspoken to in English I am
not spoken to in Gennan. Languages on the other hand, are more
tolerant of superimpositions, atleast within certain limits: Whoever
speaks to meby means of verbal language (Englishor Gennan) can
at the same time make use of gestures. As for the arts, they can be
superimposed within evenbroaderlimits: Witnessopera,ballet, and
chanted poetry. Cinema gives us the impression-occasionally erro-
neous, by the way-of rendering everything compatible with every-
thingelse, because its field of action lies for the most part between
language andart,andnotwithin langue. Thecinema as we know it
-itmay assumeothershapes, someofwhich are alreadybeginning to
emergeincertainCineramaspectacles-isa"fonnula" with many ad-
vantages: It joins consentingarts and languages in a durable union
in which their individual faculties tend to become interchangeable.
Itisacommonwealth,aswellasamarriageoflove.
If Today I would not use the words "natural" and "directly." Or at least I
8
wouldusethemmore circumspectly.See secondfootnote p. 7 .
60 PROBLEMS OF FILM SEMIOTICS
Film and Linguistics
Does this mean, then, that the study of film-at a time when linguis-
tics itself, faithful on the whole to the Saussurian teachings, is mainly
concerned with language systems-cannot have a linguistic dimension?
I am persuaded on the contrary that the "filmolinguistic" venture
is entirely justified, and that it must be fully "linguistic"-that is to
say, solidly based on linguistics itself. But how is one to interpret this
statement, since the cinema is not a language system? That is what
I would now like to clarify.
The study of film is concerned with linguistics at two points, two
separate moments of its procedure, and in the second, not quite with
the same linguistics as in the first.
It was de Saussure, we know, who made the study of language the
subject of linguistics.
51
But de Saussure also laid the foundation for
a much broader science-semiotics-of which linguistics was to be
only one branch, although an especially important one.
52
Conversely,
there were those here and there who began to study the inner mecha-
nism of nonverbal systems (traffic signals, cartographic conventions,
numbers, the gestures of politeness), or of transverbal systems (the
verbal formulas of politeness, poetry, folktales, myths), or even of
certain systems bridging the verbal and the nonverbal (kinship as
seen by Claude Levi-Strauss, with its dual organization into "appella-
tions" and "attitudes,")53 and almost all of them were fervent readers,
admirers, or direct disciples of the Swiss scholar. We will return to
the point. But already we can notice a rather striking point: In the-
ory, linguistics is only a branch of semiotics, but in fact semiotics was
created from linguistics. In a way it is very normal: For the most part
semiotics remains to be done, whereas linguistics is already well ad-
vanced. Nevertheless there is a slight reversal. The post-Saussurians
are more Saussurian than de Saussure himself: They have taken the
semiotics he foresaw and are squarely making it into a translinguistic
discipline. And this is very good, for the older brother must help the
younger, and not the other way around. Moreover, de Saussure him-
self hints at the possibility of such a cross-influence: Linguistics
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 61
could be of great help to semiotics, he writes, if it became more
semiological.54
By focusing its attention on human language, linguistics proper
has been able to obtain a knowledge of its subject with an often en-
viable rigor. It has cast a bright light, which has (not paradoxically)
illuminated neighboring topics. Thus broad aspects of the image dis-
course that a film weaves become comprehensible, or at least more
comprehensibly, when, in a first stage, they are as entities examined
distinct from langue. To understand what film is not is to gain time,
rather than to lose it, in the attempt to grasp what film is. The latter
aim defines the second stage of film study. In practice, the two stages
are not separable, for one always opens onto the other. I call one of
them the "first stage" because it benefits from the capital of linguis-
tics, which encourages one to begin with it. The "second stage" is
properly semiological and translinguistic; it is less able to depend on
previously acquired knowledge, so that, far from being helped, it
must, on the contrary, participate-if it is able to-in work that is
new. Thus it is condemned to suffer the present discomfort of
semiotics.
Image Discourse in Relation to Langue:
The Problem of Cinematographic "Syntax"
THE SECOND ARTICULATION.
55
There is nothing in the cinema that
corresponds, even metaphorically, to the second articulation.'" This
". Rather than to the cinema, this affirmation should in fact be applied to
"cinematographic language," i.e., to the s p i ~ level of codification that is
constituted by the signifying organizations proper to film and common to all
films. Therefore it would be more correct to say that the cinema as such has no
second articulation (and, as we will see further, no first articulation either).
But the "cinema" as a totality-the sum of all that is said in films, as well 3S
of all the signifying organizations (perceptual, intellectual, iconological, ideo-
logical, "symbolic," etc.) that affect the understanding of a whole film-the
cinema as a totality represent's a much vaster phenomenon, within which cine-
matographic language constitutes only one among many signifying levels. To
that extent, it is not impossible that certain cinematographiC significations are
ruled by systems that, in one way or another, contain several articulations. The
concept of cinematographic language is a methodological abstraction: This
language is never present alone in lilms but is always in combination with
62 PROBLEMS OF FILM SEMIOTICS
articulation is operative on the level of the signifier, but not on that
of the significate: Phonemes and a fortiori "features" are distinctive
units without proper signification. Their existence alone implies a
great distance between "content" and "expression." In the cinema
the distance is too short. The signifier is an image, the significate is
what the image represents. Furthermore, the fidelity of the photo-
graphic process, which gives the image particular verisimilitude, and
various other systems of signification: cultural, social, stylistic, perceptual, etc.
Secondly, it is appropriate to note that, in the linguistic sense of the term,
the articulations-i.e., what is called the "double articulation"-are not the
only conceivable types of articulation.
For either one of these two reasons, one must make a careful distinction be-
tween two affinnations: The first, which is advanced in this text, consists in
saying that cinematographic language in itself exhibits nothing resembling the
double linguistic articulation. The second, for which I assume no responsi
bility, would consist in saying that the cinema has no articulations.
Indeed, there would be nothing absurd in supposing-and this is only an
example-that the total cinematographic message brings five main levels of
codification into play, each one of which would be a kind of articulation: (I)
perception itself (systems for structuring space, "figures," and "backgrounds,"
etc.), to the degree that it already constitutes a system of acquired intelligi-
bility, which varies according to different "cultures"; (2) recognition and
identification of visual or auditive objects appearing on the screen-that is to
say, the ability (which is also culturally acqUired) to manipulate correctly the
denoted material of the film; (3) all the "symbolisms" and connotations of
various kinds attach themselves to objects (or to relationships between objects)
outside of films-that is to say, in culture; (4) all the great narrative structures
(in Claude sense) which obtain outside of films (but in them as
well) within each culture; and, finally, (5) the set of properly cinemato-
graphic systems that, in a specific type of discourse, organize the diverse ele-
ments furnished to the spectator by the four preceding instances.
We know that Umberto Eco has recently formulated an interesting hypoth-
esis according to which the cinematographic message taken as a whole would
involve only three main levels of articulation (Appunti per une semiologica
delle comunicazioni visivi, University of Florence, Bompiani, 1967, pp.
139"-52). In the preface of his work, Le cru et Ie cuit (PIon, 1964, p. 31,
notably; The Raw and the Cooked [New York: Harper & Row, 1969)), one
recalls, as well, Claude Levi-Strauss distinguished two main levels of organi-
zation in pictorial art-and it is easy to apply this to cinematographic art: On
one hand the objects that are represented on the canvas, and on the other hand
the properly pictorial composition into which they enter. For his rart, Pier
Paolo Pasolini sees two principal levels of articulation in the pictoria message,
somewhat like those discerned by Levi-Strauss (La lingua scritta dell'azione,
paper contributed to the Second Festival of New Cinema, Pesaro II, Italy,
June 1966, reprinted in Nuovi Argomenti, new series, no. 2, April-June
1966, pp. 67-103).
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 63
the psychological mechanisms of participation, which ensure the fa-
mOUS "impression of reality," shorten the distance even more-so that
it is impossible to break up the signifier without getting isomorphic
segments of the significate. Thus the impossibility of a second articu-
lation: Film constitutes an entirely too "intrinsic seme," to use Eric
Buyssens's terminology. If, in an image representing three dogs, I
isolate the third dog, I am necessarily isolating both the signifying
and the signified "third dog." The English logician and linguist
Ryle makes fun of a certain naive concept of language (which de
Saussure already condemned), which he calls ironically the "FlDO-
Fido theory": The name FlDO corresponds exactly to the dog, Fido;
words stand in direct ratio to an equal number of pre-existing things.
This point of view, very backward in linguistics, is less so in the
cinema, where 'there (ore as many "things" in the filmic image as there
were in the filmed spectacle.
The theoreticians of the silent film liked to speak of the cinema as
a kind of Esperanto. Nothing is further from the truth. Certainly
Esperanto does differ from ordinary languages, but that is because it
accomplishes to perfection what they strive for but never attain: a
system that is totally conventional, specific, and organized. Film also
differs from true languages but in the opposite way. It would be more
correct to say that the true languages are caught between two Es-
Without going into the details of these several analyses, I will remark simply
that, in principle, they do not contradict the ideas here expressed. For it is
obvious that the authors I have mentioned (I) bring into their account aside
from cinematographic language itself one or another system of signification
that is mainly cultural and extends beyond the cinema, although it does come
into play in the deciphering of the film as totality; (2) they have in mind
levels of articulation of which they ask only that they be authentic-and I
agree 'with them that they are-but of which they demand no equivalence to
linguistic articulations (whereas, I want to insist, precisely on the absence of
these equivalences, without, however, maintaining that cinema has no articu-
lation at all).
Similarly, it can be remarked that my "large syntagrnatic category of the
image-track," which will be outlined further on (Chapter 5) by its very ex-
istence constitutes a specific articulation: It resembles neither the first, nor the
second articulation of verbal language-since it does not divide the film into
units comparable to phonemes or monemes-but it undoubtedly has the effect
of articulating (in another way-that is to say, on the level of the discourse)
the cinematographic message.
64 PROBLEMS OF FILM SEMIOTICS
perantos: the true Esperanto (or the ido, or novial, or whatever
which is reached through anexcess oflinguisticity, andtheother,
cinema,whichhasadearthoflinguisticity.
Inshort, the universality ofthe cinema is a two-fold phenomenon:
Positive aspect: The cinema is universal because visual perceptioo
varies less throughout the world than languages do. Negative aspect:
The cinema is universal because it lacks the second articulation.
There is a solitary relationship between the two observations that
must be emphasized: A visual spectacle entails a joining of the sig-
nifier to the significates, which in turn renders impossible their dis-
junction at any given moment and, therefore, the existence of a sec-
ond articulation.
Strictlyspeaking, Esperantois manufactured; itis aproductoflan-
guage. For the most part "visual Esperanto" is a raw material that
precedeslanguage. In thisconceptoffilmic Esperanto thereis, all the
same, Some truth: Itisin thesecond articulation thatlanguagesdiffer
mostradically, onefrom another,andthatmenfail to understandeach
56
other. As Roman ]akobso
n
observes, thesentenceis always moreor
less translatable. 57 Thatis because itcorresponds to a real mental im-
pulse, and not to a code unit. Theword can still yield interlingUistic
equivalences,imperfect to be sure, butsufficient to make dictionaries
possible.Thephonemeiscompletelyuntranslatable,sinceitisentirely
defined by its position in the phonemic grid of each language. The
absence of a meaning cannot be translated. Thus we return to the
idea that image discourse needs no translation, and that is because,
having no second articulation, it is already translated into all lan-
guages: The height of the translatable is the universal.
Andre Martinet believes that, strictlyspeaking, one cannot talk of
language,exceptwherethereisdoublearticulation.58Inpointoffact,
the cinema is not a language but a language of art. The word, lan-
guage, has numerous meanings, more or less strict, and each is in its
Own way justified. Thispolysemic multiplication tendS-in my judg-
ment-to branch out in two directions: Certain systems (even the
leasthumanones)are called "languages" iftheir formal structure re-
sembles that ofOur spoken languages: This is the case with the lan-
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 65
guageof chess (which de Saussure found so interesting) or with the
,binarylanguages ofcomputers. At theother pole, everything that ex-
!presses man to himself (even in the least organized and least linguis-
I tic way) is felt to be a "language": the language of flowers, the lan-
.guage of painting, even the language of silence. The semantic field
of the word, language, seems to be organized along these two axes.
Now it is in "language"in its most proper meaning (human phonic
language) that these two vectors of metaphorical expansion are
rooted: For verbal language is used by men to communicate among
themselves, and it is highly organized. The two nodes of figurative
meaning are already there. Keeping these conditions of usage in
mind, which does not always allow us to abide by meanings one
would like to be strict about, it seems appropriate to look at the
cinemaas alanguagewithoutasystem.
FIRST ARTICULATION. The cinema has no phonemes; nor does it,
whatever onemay say, have words. Except on occasion, and more or
less by chance, it is not subject to the first articulation. It should be
shown that the almost insurmountable difficulties that film "syn-
taxes" launch themselves into are derived for the most part from an
initial confusion: Theimage is defined as a word, the sequence as a
sentence. The case is, however, that the image (at least in the cin-
ema) corresponds to one or more sentences, and the sequence is a
complexsegment of discourse.
The word, sentence, of course, here and on the following pages,
designates the oral and not the written sentence of grammarians (a
complex statement with multiple assertions, contained between two
marked punctuations). I am speaking of the linguists' sentence.'" In
.. This distinction between two kinds of sentence seems less important to me
today than it didwhenI wrote this article. First ofall, from a purely linguistic
point of view, Vendrys's analysis, which I used to support my argument, is
subject to a number of reservations, particularly since Noam Chomsky's work
has progressively shed new light on the problem of the sentence. Secondly,
from a properly cinematographic point of view, it is in any case impossible to
say whetherthe"shot" corresponds to one or to severalsentences: Thequestion
of knowing whether these sentences would be of the "written" or "oral" type,
66 PROBLEMS OF FILM SEMIOTICS
his famous example, designed precisely to distinguish between the
two types of sentence, Joseph Vendrys59 maintains that there are
five sentences (in the sense that interests us) in the following state-
ment: "You see that man/ over there/ he's sitting on the sandi well,
I met him yesterday/ he was at the station." I do not want to suggest
that a film sequence with the same contents would have exactly these
five sentences (shots). I mean simply that the image in the cinema
is a kind of "equivalent" to the spoken sentence, not to the written
sentence. It is not impossible that certain shots or groups of shots
might, in addition, correspond to the "written" type of sentence-but
that is another problem. In many respects,60 film recalls written ex-
pression a great deal more than spoken language. But at a certain
point in the division into units, the shot, a "completed assertive state-
ment," as Benveniste would call it,61 is equivalent to an oral sentence.
Roman Jakobson writes
62
that Shimkin, in his work on proverbs,
was brought to propose that, in the proverb, "the highest coded lin-
guistic unit functions simultaneously with the smallest poetic unit."
The image discourse of the cinema represents a nonverbal area (while
the proverb is transverbal). Nevertheless the shot, a "sentence" and
not a word (like the proverb), is indeed the smallest "poetic" entity.
How is one to understand this correspondence between the filmic
image and the sentence? First of all, the shot, through its semantic
content, through what Eric Buyssens would call its "substance"63 is
closer, all things considered, to a sentence than to a word. An image
"simple" or "complex," is therefore quite secondary. One can say simply that
a film "shot" is very different from a word, that it always constitutes an ac-
tualized unit of discourse, and that consequently it is to be situated on the level
of the sentence. And that is indeed the observation I made in this text; but it
should have exempted me from looking for more precise equivalences between
the shot and one or another type of internal phrastic structure observable in
languages. Since the shot is not made of words, it can "correspond" only ex-
ternally to the sentence, i.e., in relation to discourse (see Chapter 5, part 3).
As long as one seeks internal equivalences, one will be led into an impasse:
Let us indeed suppose that in certain circumstances a shot can appear to be
equivalent to several sentences (a thing that will not fail to occur): How are
we to know if these sentences, in a written text, have always been separate
sentences or if, for example, they had at one time been different "clauses" of a
single complex sentence?
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 67
shows a man walking down a street: It is equivalent to the sentence
"A man is walking down the street." The equivalence is rough, to be
sure, and there would be much to say about it; however the same
filmic image corresponds even less to the word "man," or the word
"walk," or the word "street," and less still to the article "the" or to the
zero-degree morpheme of the verb "walks."
The image is "sentence" less by its quantity of meaning (a con-
cept too difficult to handle, especially in film) than by its assertive
status. The image is always actualized. Moreover, even the image-
fairly rare, incidentally-that might, because of its content, correspond
to a "word" is still a sentence: This is a particular case, and a particu-
larly revealing one. A close-up of a revolver does not mean "revolver"
(a purely virtual lexical unit), but at the very least, and without
speaking of the connotations, it signifies "Here is a revolver!" It car-
ries with it a kind of here (a word which Andre Martinet rightly
considers to be a pure index of actualization.
64
Even when the shot is
a "word," it remains a kind of "sentence-word," such as one finds in
certain languages.
CINEMA AND SYNTAX. The image is therefore always speech,
never a unit of language. It is not surprising that the authors of
"cinematographic grammars" have found themselves at an impasse.
They claimed to have written a syntax of film, but, in fact, with their
image-word, they had been thinking of something half way between
a lexicon and a morphology, something that does not exist in any lan-
guage. The cinema is something else.
There is a syntax of the cinema, but it remains to be made and
could be done only on a syntactical, and not a morphological, basis.
66
De Saussure observed
66
that syntax is only an aspect of the syntag-
matic: A thought that should be meditated on by anyone concerned
with film. The shot is the smallest unit of the filmic chain (one
might perhaps call it a "taxeme," as Hjelmslev uses the term
67
); the
sequence is a great syntagmatic whole. One should examine the rich-
ness, exuberance even, of the syntagmatic arrangements possible in
film (which will bring one to see the problem of montage under a
68 PROBLEMS OF FILM SEMIOTICS
new light), and contrast it to the surprising poverty of the paradig-
matic resourcesofthe cinema. '"
THE PARADIGMATIC CATEGORY OF FILM. In thewritings of theore-
ticians, the word, montage, in its broad sense includes cutting, but
the opposite never Occurs. The moment of ordering (montage) in
film is somehow more important-"linguistically" at least-than the
choosing ofthe images (cutting), no doubt because the latter, being
too open, is nota choice, butratheran act of decision, a kind ofcre-
ation. Thatis why, on theartistic level, the contentofeach shot is of
greatimportance (although the organization is itself an art). On the
level ofthe visual subject, there is art, if anything. Art continueson
the level of the sequence or of the composed shot, and "cinemato-
graphic language" begins. Hence the condemnation of "beautiful
photography" in the cinema.
.. I am no longer of the opinion, as I reread this, that the two aspects of
the problem can be strictly speaking "contrasted." For, as I have said-butnot
sufficiently clearly-in the text reprinted here, they are not situated on the
same level: One can speak of "paradigmatic poverty" in relation to the image
(see pp. 65-67) and of "syntagmatic richness" in relation to the structuring
of images (see p. 67). At the same time, however, it should be remarked
that the existence of several types of image-ordering has the effect of creat-
ing (on the level of discourse) a specific paradigmatic category, which is
constituted preciselyby the total system of the different syntagmas. One cannot
indeed conceive of a syntagmatic category with no corresponding paradigmatic
category on the same level, (thatis, a paradigmatic category related to units of
the same magnitude), nor of a paradigmatic category with no corresponding
syntagmatic category on the same level: by definition the syntagmatic cate-
gories and paradigmatic categories are strict correlatives. I will return to these
problemsinChapter5.
Second observation: One must not exclude the possibility that between the
images themselves therearedifferent kinds ofparadigmatic associations, since in
all human groups one finds various cultural "symbolisms" that relate to iconog-
raphy. Theseparadigms, however, are notpeculiar to cinematographiclanguage.
Third and final observation: One must keep in mind when distinguishing
between the "image" and the "structuring of images" that the first term can
designate either the shot (as opposed to the sequence) or the filmed subject (as
opposed to the shot, which is already the product of an initial composing or ar-
rangement). Ineithercase, what we call "image"is really the photographic fact
(or phonographic fact, if we are referring to the sound-track), and what it is
contrasted to is the filmic fact. The latter indeed unfolds on two levels: within
theshot, from "subject" to "subject"; within thesequence, from shotto shot. For
moreonthispoint,seeChapter5, particularlypart6.
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 69
The image paradigm is fragile in film; often still-born, it is ap-
proximate, easily modified, and it can always be circumvented. Only
to aslightdegree does thefilmic image assume meaning in relation to
the other images that could have occurred at the same point along
the chain. Norcan the latter be inventoried; their number is, if not
limitless, at least more "open" than the "most open" linguistic inven-
tory. Thereis no equivalent here to the "peribolic" unraveling whose
importance in verbal language has been underlined by Guillaume.
68
Charles Bally69 observed thatcertain units that are opposed to an un-
limited and undefinable number of terms (dependent only on con-
text, thespeakers, and the association of ideas) are often in the long
run unopposed, really, to any term: This is somewhat the case with
the filmic 'image.
Everything is present in film: hence the obviousness of film, and
hence also its opacity. The clarification of present by absent units
occurs much less than in verbal language. The relationships in prre-
sentia areso rich that theyrenderthestrictorganization ofin-absentia
relationships superfluous and difficult. A film is difficult to explain
because it is easy to understand. The image impresses itself on us,
blockingeverythingthat is notitself.
A rich message with a poor code, or a rich text with a poor sys-
tem, the cinematographic image is primarily speech. It is all asser-
tion. The word, which is the unit of language, is missing; the sen-
tence, which is theunitofspeech, is supreme. Thecinema can speak
only in neologisms. Every image is a hapax.'" It would be fruitless to
.. It is becoming less and less certain, in fact-especially since one has begun
to understandChomsky'swork-thatthe sentenceis a unitofspeech. Inoneway
the sentence is even the unit of language par excellence, since a language is a
systemthatallowsonetomakesentences.
As for thefilmic image, which is onlya "sentence"because ofits function in
discourse, and not because of its internal structure, it does indeed remain a
"hapax," but it also is contained by larger units that are not hapaxes. I was
therefore only partially correct in saying that the cinema can only talk by
neologism. I should have said that, in order to speak, the cinema is able to use
only neologism as its basic material, but that, in speaking it integrates these
neologisms (without however altering them in their details) intoa second order
not governed bythe single law ofproliferation. On these problems, see Chapter
5, parts3and4
- - - - - - - - .. ----_. . . . ~ ~ ~
70 PROBLEMS OF FILM SEMIOTICS
search among images for true associative series or strict semantic
fields. Even the cautious, flexible structuralism of a Stephen Ull-
mann,70 for example, is out of place here, for it is lexicological, and
a "filmolinguistic" structuralism can only be syntactic.
There is a paradigmatic category of film. But the commutable units
are large signifying units. Thus, in the scholarly work of J. L. Rieu-
peyrout on the history of the Western, we are told that there was a
period when the "good" cowboy was indicated by his white clothes
and the "bad" by his black costumes. The audience, apparently, al-
ways knew which was which. This allows us to establish a rudi-
mentary commutation as much on the level of the signifier (white/
black) as on that of the significates (good/bad). The two colors are
already predicated (since attributed to present clothes), and so are
the two qualities (since it is the cowboy in the image who is either
good or bad), prior to the commutation, and this is the essential dif-
ference from a lexical and a fortiori phonological commutation. But
that is not all: The paradigm, perhaps precisely because it is engaged
too much in "speech," is unstable and fragile; the convention of the
cowboy in white, or in black, did not last long. It was inevitable that
one fine day a film-maker, bored with the routine, should get the
idea of dressing his rider in gray, or in a white shirt and black pants,
and so much for the paradigm! The poverty of the paradigm is the
counterpart of a wealth distributed elsewhere: The film-maker can
express himself by showing us directly the diversity of the world,
and in this he differs from the reciter of tales. Thus the paradigm is
very rapidly overwhelmed: This is another aspect of the kind of
struggle, which occurs at certain points in the cinema, between code
and message. The great directors (and is it not puerile to repeat al-
ways that the cinema is not them, for who else could it be?) have
avoided the paradigm.
Or at least have avoided certain paradigms. For the "type" cowboy
in black/cowboy in white defines only one kind of filmic paradigm.
"Syntactic" by the syntagmatic extent of its commutable segments
and by their asserti\'e status, such an opposition, howe\'er, bears, by
its contents, on aftecti\'e impressions ("the cowboy is good") that re-
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 71
tain something of a "levical" quality. Other filmic oppositions, also
more or less commutable are more rooted in syntax and bear more on
kinds of "morphemes."'" A large number of camera movements (rear
and forward dolly) or techniques of punctuation (dissolve or cut)
can be considered in this light....... It is a case of one relationship op-
posed to another relationship. In addition to the commutable ele-
ments, there is always a kind of support
71
that is ideally invariable.
Rear and forward dolly shots correspond to two intentions of a "see-
ing"-but that "seeing" always has an object; that which the camera
moves away from or approaches. It is, therefore, to the theory of syn-
categorematic tenus that one should look here: Just as the word "but"
never expresses the idea of the adversative as such, but always an ad-
versative relationship between two realized units, a forward dolly ex-
presses a concentration of attention, not on itself, but always on an
object.
The duality of support and relationship, in a language that per-
mits the simultaneity of several visual perceptions, explains why such
procedures have something supersegmental to them: The support
and the relationship are often perceived at the same time. Further-
more, in the cinema the "relationship" is often one and the same as
the camera's (and the spectator's) "seeing" of the support object. A
forward dolly on a face is a way of seeing that face. That is why so
.. The word is taken here in the sense in which it is opposed to the term
"semanteme," or the more current '1exeme," and not in the sense: minimum
unit having its own signification. It seems to me that I did not sufficiently insist,
in the passage above, on the fact that paradigms of this second variety-pre-
cisely because they come closer to those of true grammar-are much less easily
judged than the others in terms of "originality" or "banality" when one en-
counters them in films. It is these paradigms that constitute "cinematographic
language" itself, whereas the systems like that of the cowboy in black or white
only affect a few film subjects, and only for a restricted period. I return to this
problem later, in Chapter 6 (see especially pp. 221-23, including note on 223).
.... And, even more so, many o d i ~ e d montage orderings (see "Problems of
Denotation in the Fiction Film," especially pp. 119-33). That is why I am
indeed less skeptical today about the paradigmatic category of film than I was
when I wrote this article; what I had not seen is that a major part of the para-
digmatic category of film must be sought for in the syntagmatic category itself-
that is to say in the interplay between various different image orderings. On this
point see also "Problems of Denotation. . . ."
72 PROBLEMS OF FILM SEMIOTICS
many materially unreal filmic procedures are psychologically con-
vincing, as has been occasionally noted.
72
Consider, for example, the
rapid forward dolly, which makes the object grow larger to our eyes,
or oblique framing, or certain extreme close-ups-they are all in-
stances in which the appearance of the object is hardly plausible.
But the super segemental aspect of the support-relationship pair dis-
places filmic "plausibility" toward the level of the living, constructive
dynamism of perception, and away from that of the objective circum-
stances of the perceived situation, for in the same segment the film
contains a perceived and a perceiving instance. There are many cam-
era movements that work by bringing an implausible object to a
"seeing" that makes it plausible.
Filming Intellection
73
A film is always more or less understandable. If by chance it is not at
all understood, that is as a result of peculiar circumstances, and not
of the semiological process proper to the cinema. Naturally the cryp-
tic film, like cryptic utterance, the extraordinary film, like the extraor-
dinary book, the film that is too rich or too new, like an explanation
that is too rich or too new, can very well become unintelligible. But
as "language" a film is always grasped-except by abnormal persons
who would not understand any other form of discourse any better,
and often not as well; except by the blind, suffering from a selective
impairment blocking reception of the signifier (like the deaf with
speech); except, finally, in those cases where the actual substance of
the signifier is materially damaged (the old film, scratched, yellowed,
and undecipherable; the speaker whose voice is so hoarse he cannot
be understood).
Aside from these cases, a film is always understood, but always
more or less so, and this "more or less" is not easily quantifiable, for
there are no discernible degrees, no units of signification that can
be immediately counted. With two persons speaking different lan-
guages, one should-in principle, at least-be able to enumerate the
quantitative degrees of their mutual understanding: A knows three
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 73
words of the language spoken by B, and B six words of the language
spoken by A. Within a given sentence, it is a particular word, and
not its neighbor, that has not been understood; or it can be estab-
lished that one certain word, by a kind of ricochet, has rendered the
whole sentence unintelligible to the hearer. A linguistic unit is either
recognized or not by the hearer, since it already exists in the lan-
guage. Marcel Cohen's suggestion (to study the degrees of under-
standing between languages) can be successfully undertaken, despite
great difficulties. But in the cinema the units-or rather, the elements
-of signification that are present together in the image are too nu-
merous and too continuous: Even the most intelligent viewer can-
not discern them all. On the other hand it is sufficient to have gen-
erally understand the main elements to grasp the approximate, overall,
and yet relevant meaning of the whole: Even the dimmest spectator
will have roughly understood. There have been some rather in-
teresting experiments,74 thanks to which one has been able to isolate
the character of what is easy or difficult to understand in film. But
it cannot be inferred from this that the degree of understanding of a
normal commercial film can easily be established for a particular
viewer or category of viewers.
One must also set aside clearly all those cases-very numerous in
the cinema, as well as in verbal language, literature, and even every-
day life-in which the message is unintelligible because of the very
nature of what is being said, without the semiological process being
affected. Many films are incomprehensible (either entirely or in part
and depending on the audience) because their diegesis contains reali-
ties or concepts that are too subtle, too exotic, or mistakenly thought
to be familiar. The fact has not been sufficiently emphaSized that, in
these cases, it is not the film itself, but on the contrary what the film
does not make clear, that is incomprehensible. And the reason this
has not been emphasized is because the current fashion is to insist
that everything is language, to such an extent that what is said is
overpowered and reduced to nothing by how one says it. This is a
very common illusion. The angry lover shouts to his faithless mis-
tress, "You don't understand me!" But she understands him only too
- --- - ... ~ - - ~ ~ _ .
74 PROBLEMS OF FILM SEMIOTICS
well; the case is simply that she no longer loves him. Whether filmic
or verbal, language cannot suppress reality; on the contrary, it is
rooted in reality. If men do not "understand" each other, it is not
only because of words, but also because of what the words contain.
How many "misunderstandings" are actually the result of too great an
understanding! People always see a lack of understanding where
there is real disagreement. A whole army of Korzybskis and "general
semanticists"(!) cannot put a stop to antagonism, stupidity, and in-
difference. The audience of local shopkeepers* who booed Anto-
nioni's L'Avventura at the Cannes Film Festival had understood the
film, but either they had not grasped, or were indifferent to, its mes-
sage. Filmic intellection had nothing to do with their attitude; what
bothered them was simply "life" itself. It is normal that the problems
of the couple as stated by Antonioni should leave a large section of
the audience indifferent, puzzled, or derisive. **
~ They are given free tickets by the municipality of Cannes and constitute
what one refers to as the Festival audience.
~ ~ In this passage reprinted here, the material has been considered too ex-
clusively from the point of view of the cinema, and I have not paid sufficient at-
tention to the possibilities of a general semiotics of culture. To be sure, it was
not the cinematographic language in L'Avventura (which was utilized in a
particularly clear fashion) that put off the Cannes "shopkeepers"; in this sense
one can say that it was the subject of the film, and therefore "life:' that had
annoyed them. If all that is wanted is to show that the particular problem of
filmic intellection is not relevant in such cases, the argument suffices. But if one
wants to go further, one must then indeed observe that the "subject" of the film
(as well as "life" itself) is, in turn, liable to be more or less understood-again
depending on the audience, and on the form in which it is presented-according
to a set of cultural systems that, though they are foreign to the cinema, never-
theless do constitute organized systems of signifiers.
In other words, the distinction referred to between what is said and how it is
said should be made relative: One can identify what pertains to the "saying"
and what belongs to the "said" only in relation to the instance of "saying" with
which one is occupied in each particular analysis. When analyzing another set
of signifiers, one might well find that what was part of the "saying" now comes
under the "said." In every human phenomenon of some magnitude-the cinema
included-various cultural systems intervene together and overlap in complex
ways. A "content" determined by one of these systems can be annexed by an-
other (which encloses it although it did not determine it) within the same
overall "message." What we call "the cinema" is not only cinematographiC
language itself; it is also a thousand social and human significations that have
been wrought elsewhere in culture but that occur also in films. Furthermore,
the "cinema" is also each individual film as a unique composition, with signifying
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 75
Cinema and Literature:
The Problem of Filmic Expressiveness
The cinema is not a language system, because it contradicts three im-
portant characteristics of the linguistic fact: a language is a system of
signs used for intercommunication. Three elements to the defini-
tion.
75
Now, like all the arts, and because it is itself an art, the cinema
is one-way communication. As we have already seen, it is only partly
a system. Finally, it uses only very few true signs. Some film images,
which, through long previous use in speech, have been solidified so
that they acquire stable and conventional meanings, become kinds
of signs. But really vital films avoid them and are still understood.
Therefore the nerve center of the semiological process lies elsewhere.
The image is first and always an image. In its perceptual literal-
and signified elements distinct from those of cinematographic language in general.
In the case of L'Avventura, one can isolate at least three autonomous factors of
signification (sets of signifiers) that are present at the level of the concrete
message: (I) cinematographic language (a much larger category but one that
does not exclude the film), (2.) L'Avventura as a work of art (which, on both
the levels of expression and content, adds to cinematographic language many
particular structures that do not belong to the general "writing" of the cinema),
and (3) a peculiar ideology (that of the "modern couple," of the "exhaustion
of feeling:' etc.) derived from a historical and sociocultural situation foreign to
the cinema, but liable to be reflected in films. The insufficiently educated audio
ence in my example did not understand the film, not because it could not grasp
system I, but surely because it was unable to decipher correctly systems 2. and 3.
The cinematographic "saying" was not responsible. But nor was the "said" en-
tirely so either-or, if it was, it was so only in relation to the previous "saying"-
since it comprised on the one hand a particular "Antonionian" content and on
the other a certain "socia-ideological" content. So that, in a way, it is still true
that the cause lay with the "saying" and not with what was "said"; yet in an-
other way the responsibility was in the "said" and not in the "saying." In the
study of a determined signifying system many things that appear to be pure
substance actually correspond to significations derived from elsewhere, where
they existed as forms and not as substances. Unless one entirely abandons the
endeavour to speak about Man, one cannot avoid the enclosing of meaning. The
reason that the cinema as a totality gives a first impression of being a collective
body lacking in any kind of strict organization is in large part becasue it is one
of the locations where a very large number of signifying systems, each having its
relative autonomy, come together from the four corners of culture: cinemato-
graphic language itself is only one of those systems. Whatever cinematographic
language does not account for is not necessarily condemned to formlessness;
simply, it has been formed elsewhere.
76 PROBLEMS OF FILM SEMIOTICS
ness it reproduces the signified spectacle whose signifier it is; and thus
it becomes what it shows, to the extent that it does not have to sig-
nify it (if we take the word in the sense of signum facere, the special
making of a sign). There are many characteristics to the filmic image
that distinguish it from the preferred form of signs-which is arbi-
trary, conventional, and codified. These are the consequences of the
fact that from the very first an image is not the indication of some-
thing other than itself, but the pseudopresence of the thing it
contains.
The spectacle recorded by the film-maker may be natural ("real-
istic" films, scenes shot in the street, cinema verite, etc.) or arranged
(the film-operas of Eisenstein's last period, Orson Welles's films, and,
in general, the cinema of the unreal, or of the fantastic, expressionist
cinema, etc.). But it is baSically all one thing. The subject of the
film is either "realistic" or not; but, whatever the case, the {lIm itself
only shows whatever it shows. So we have a {lIm-maker, realistic or
not realistic, who {llms something. What happens? Whether natural
or arranged, the filmed spectacle already had its own expressiveness,
since it was after all a piece of the world, which always has a mean-
ing. The words a novelist uses also have pre-existing since
they are segments of language, which is always significant. Music
and architecture have the advantage of being able to develop imme-
diately their properly aesthetic expressiveness-their style-in mate-
rials (sound or stone) that are purely impressionable and do not
designate.
76
But literature and the cinema are by their nature con-
demned to connotation,77 since denotation always precedes their ar-
tistic enterprise. *
Film, like verbal language, can be used merely as a vehicle, with-
out any artistic intention, with designation (denotation) governing
alone. Consequently, the art of the cinema, like verbal art, is, so to
speak, driven one notch upward: 78 In the final analysis it is by the
wealth of its connotations that Proust's great novel can be distin-
"Prior" to literature, denotation is secured through idiom. "Prior" to the art
film, it is secured: (1) through perceptual analogy, (2) through the cinema
language that contains a partial denotative code (derived incidentally from the
earlier search for connotations). On this point see Chapter 5, pp. 117-19.
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 77
gUished-in semiological terms-from a cookbook, or a hlm of Vis-
conti's from a medical documentary.
Mikel Dufrenne believes that in any work of art the world that is
represented (denoted) never constitutes the major part of what the
author has to say. It is merely a threshold. In the nonrepresentational
arts it is even missing: The art of stone and the art of sound do not
designate anything. When it is present its function is only to intro-
duce the expressed world;79 the artist's style, the relationship of
themes and values, a recognizable "accent"-in short, the connotative
universe.
In this respect, however, there is an important difference between
literature and the cinema. In the cinema, aesthetic expressiveness is
grafted onto natural expressiveness-that of the landscape or face the
film shows us. In the verbal arts, it is grafted, not onto any genuine
prior expressiveness, but onto a conventional of
language-which is generally inexpressive. Consequently the intro-
duction of the aesthetic dimension-expressiveness added to expres-
siveness-into the cinema is made with ease: An easy art, the cinema
is in constant danger of falling victim to this easiness. It is so easy to
create an effect when one has available the natural expression of
things, of beings, of the world! Too easy. The cinema is also a dif-
ficult art: For, Sisyphus-like, it is trapped under the burden of its fa-
cility. There are very few films which do not have a little art in them;
fewer still contain a great deal of art. Literature-especially poetry-
is a so much more improbable art! How can that insane craft ever
succeed?-To bestow an aesthetic expressiveness (that is, in a natural
way) upon those "words of the tribe" Mallarme railed against-where
linguists agree in recognizing only a small portion of expressiveness
and a very large portion of arbitrary signification, even when one
considers the modifications brought to the famous theory of the "ar-
bitrary" since de Saussure (the presence in language of partial moti-
vation-whether phonic, morphological, or semantic-brought to light
principally by Ullmann; the motivations by the signifier and other
"impliCit associations" analyzed by Charles Bally; and, in general, the
various studies on the "motivated" areas of language). But when the
7
8
PROBLEMS OF FILM SEMIOTICS
poet has succeeded in his initial alchemy and has made words ex-
pressive, the greater part of his task is done: In this respect, litera-
ture, which is a difficult art, enjoys at least that advantage. Its en-
deavor is so arduous that the weight it bears is hardly a danger. There
are a great number of books entirely lacking in art; there are a few
books possessing enormous art.
The concept of expression is used here as defined by Mikel Du-
frenne. There is expression when a "meaning" is somehow immanent
to a thing, is directly released from it, and merges with its very form..y.
Some of Eric Buyssens's "intrinsic semes" perhaps fall under this defi-
nition. Signification, on the contrary, links from the outside an iso-
lable signifier to a significate that is itself-this has been known since
de Saussure
8
0-a concept and not a thing. These are the "extrinsic
semes" Buyssens writes about.
81
A concept is signified; a thing is ex-
pressed. Being extrinsic, signification can only derive from a conven-
tion; it is of necessity obligatory, since one would deprive it of its
only support-consensus-by rendering it optional. This, one recog-
nizes, is the famous "thesis" of the Greek philosophers. There is more
than one difference between expression and signification: One is
natural, the other conventional; one is global and continuous, the
other divided into discrete units; one is derived from beings and
things, the other from ideas..y..y.
.y. Gestalt, and not graphic contour.
.y..y. Today, I would say rather that expressiveness is a meaning established
without recourse to a special and explicit code. But not without recourse to vast
and complex sociocultural organizations, which are represented by other forms of
codification. On this point, see Chapter ), pp. 110-14 and 14
0
-4
2
In general,
if the sum of the effects of meaning we call expressive, or motivated, 01 sym-
bolic, etc., appears to be "natural"-and is indeed so in a certain way, for ex-
ample to a phenomenology or a psychology of meaning-it is mainly because the
effects are very deeply rooted in cultures, and because they are rooted at a
level that, in these cultures, lies far beyond the various explicit, specialized, and
properly informative codes. One Can of course argue whether these deep signi6-
catory organizations existing at these distant levels can rightly be considered as
proper codes. But, whatever the case, they are more or less organized systems,
which can convey meaning and vary from one human group to another. If as a
general rule the system-user experiences them not as codes but as effects of
natural meaning, that is because he has suffiCiently "assimilated" them to the
extent that he does not possess them in a separate state. Thus, as a paradoxical
consequence, the deepest cultural codifications are experienced as the most
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 79
The expressiveness of the world (of the landscape or face) and the
expressiveness of art (the melancholy sound of the Wagnerian oboe)
are ruled essentially by the same semiological mechanism: "Mean-
ing" is naturally derived from the signifier as a whole, without re-
sorting to a code. It is at the level of the signifier, and only there, that
the difference occurs: In the first case the author is nature (expres-
siveness of the world) and in the second it is man (expressiveness of
art).
That is why literature is an art of heterogeneous connotation (ex-
pressive connotation added to nonexpressive denotation), while the
cinema is an art of homogeneous connotation (expressive connotation
added to expressive denotation). The problem of cinematographic
expressiveness should be studied from the point of view I have just
outlined, for it would bring one to consider style, and therefore the
author. In Eisenstein's Que Viva Mexico, there is a famous shot of
the tortured, yet peaceful faces of three peons buried to their shoul-
ders being trampled by the horses of their oppressors. It is a beauti-
ful triangular composition, a well-known trademark of the great di-
rector. The denotative relationship yields a signifier (three faces)
and a significate (they have suffered, they are dead). This is the "sub-
ject," the "story." There is natural expressiveness: Suffering is read
on the peons' faces, death in their motionlessness. Over this is super-
imposed the connotative relationship, which is the beginning of art:
The nobility of the landscape as it is structured by the triangle of the
natural. Other codifications-which are cultural too, but are more superficial
or more specialized-are, on the contrary, much more easily identified by the
user as conventional and separate systems.
In the text above, I gave, among other examples of expressiveness, what we
quite rightly call "facial expression." Certainly it is not through the effect of
"cinematographic language" (nor of any other explicitly infonnative code)
that the film spectator is able to decipher the expressions he reads on the hero's
face. However, it is not through the effect of nature itself either, for the ex-
pressions of the face have meanings that vary considerably from one civilization
to another (think of the difficulty one experiences in trying to understand the
facial expressions in a Japanese film). Nevertheless it remains true that in films
of our own culture we understand them quite naturally-that is to say, through
the effect of a knowledge that is very old and very deep in us, that functions by
itself, and that-for us-is henceforth merged with perception itself.

H
Hi
I
80 PROBLEMS OF FILM SEMIOTICS
faces (form of the image) expresses what the author, by means of
his style, wanted it to "say": The greatness of the Mexican people,
his certainty of their eventual victory, a kind of passion in that man
from the North for all that sunny splendor. Therefore, aesthetic ex-
pressiveness. And yet still "natural": The strong and savage grandeur
rises very directly out of the plastic composition that turns suffering
into beauty. Nevertheless, two languages exist side by side in this
image, since one can identify two signiflers: (I) three faces in a bar-
ren stretch of land; (2) the landscape given a triangular shape by
the faces-and two significates-( I) suffering and death; (2) gran-
deur and triumph. One notices that, as usual, the connoted expres-
sion is much "vaster" than the denoted expression, and is also discon-
nected from it.
82
One finds the denotative material (signifier and
significate) functioning as the signifier of the connotation: The
solemn, sorrowful victory that the image connotes is expressed by the
three faces themselves (signifiers of denotation) as well as by the mar-
tyrdom they exhibit (significate of denotation). The signifier of the
aesthetic language is the sum of the signifier-significate of a prior
language (the anecdote, the subject) with which it is interlocked.
This is precisely Hjelmslev's definition of connotation; the linguist,
we know, does not use the terms "signifier" and "significate," but
expression and content ("centimatique" and "plerematique")' How-
ever, for the student of the cinema the word expression (as distinct
from is too useful to be given the meaning of "Sig-
nifier," for the result would be a very annoying polysemic collision.
From our point of view, therefore, "expression" does not designate the
signifier, but rather the relationship between a signifier and a signifi-
cate, when that relationship is "intrinsic."'" It would even be possible,
in the case of expressive semes, to use expresser and expressed, re-
.. This terminological problem does not seem as serious to me today; it suf-
fices to indicate clearly in each case what one is talking about. Moreover the
relationship between expression and signification no longer seems to me to be as
adversative; the distinction retains all of its value for a phenomenology of mean-
ing, but for a semiological analysis it may be a matter of codified organizations
in both cases, although each has a different character and is situated at a differ-
ent level. See Chapter 5, pp. 110-14.
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 81
serving "signifier" and "significate" for nonexpressive relationships
(signification proper). But one hesitates dropping such established
terms, which, since de Saussure, have been linked to so many impor-
tant analyses, as the words and
Comparisons are frequently made between the cinema and "lan-
guage," where the identity of the latter is uncertain and variable. At
times it is literature (the art of language), and at other times it is
ordinary language that one contrasts to film. In such a muddle it is
impossible to see clearly. The art of words and the art of images, as
we have seen, are located along the same semiological horizon; on the
connotative level they are neighbors. If the art of the cinema is com-
pared to ordinary language, however, everything is changed; the two
members are" no longer on the same plane. The cinema begins where
ordinary language ends: at the level of the "sentence"-the film-
maker's minimum unit and the highest properly linguistic unit of
language. We then no longer have two arts; what we have is one art
and one language (in this particular case, language itself). The
strictly linguistic laws cease when nothing is any longer obligatory,
when ordering becomes "free."'" But that is the point where film be-
gins; it is immediately and automatically situated on the plane of rhet-
oric and poetics.
How then is one to explain such a curious lack of symmetry, a lack
that insidiously confuses scholars and renders books obscure? On the
.. The quotation marks were placed around the word "free" to indicate that
the freedom I am talking about is never total-since, in the next sentence I
mention "rhetoric and poetics" (to which one might add, incidentally, the
various Barthesian "writings"). In many ways "Cinematographic language" is
one of these "writings"-on this point, see note, Chapter 6, p. 223 Simply, it is
true, as Roman Jakobson noted, that, as one considers syntagmatic units of in-
creasing magnitude, the portion of freedom available to the speaker becomes
increasingly important. In this respect, the level of the sentence is a kind of
threshold, below which the speaker is ruled (for the most part anyway) by the
law of idiom, beyond which he falls under various laws of "composition,"
"rhetoric," etc., which are less restrictive-or perhaps have other restrictions?-
than those of idiom. One can, if not ignore them, at least circumvent them,
play with them, bend them, etc. That is indeed why the most authentic crea-
tivity (or "originality") is by no means inseparable from a total freedom: The
French classical writers of the seventeenth century understood this perfectly,
and it was bourgeois romanticism that made us forget it.
82 PROBLEMS OF FILM SEMIOTICS
verbal side two levels-verbal language and literature-are readily dis-
tinguished. On the filmic side there is only and always the "cinema."
To be sure, one can make a distinction between films that are purely
utilitarian (educational films, for example) and films that are artistic.
Nevertheless one senses that this is not altogether satisfactory and
that the distinction is not as clear cut as that between the poetic or
dramatic word and a conversation in the street. There are, of course,
borderline cases that obscure the division: The films of Flaherty,
Murnau, or Painleve, which are both documentaries (biological and
ethnographic) and works of art. In the verbal order, however, one
may find many equivalent examples. The crux of the matter must,
therefore, lie elsewhere. In truth there is no totally aesthetic use of
the cinema, for even the most connotative image cannot avoid being
also a photographic representation. Even in the period when film-
makers like Germaine Dulac dreamed of a "pure cinema," the most
nonrealistic avant-garde films, the films that were the most resolutely
devoted to the exclusive concerns of rhythmic composition, still rep-
resented something-whether it was the variations of changing cloud
patterns, the play of light on water, or the ballet of pistons and con-
necting rods. Nor is there a totally "utilitarian" cinema, for even the
most denotative image has some connotations. The most literal educa-
tional documentary cannot prevent itself from framing its images and
organizing their sequence with at least something like an artistic con-
cern; when a "language" does not already exist, one must be some-
thing of an artist to speak it, however poorly. For to speak it is partly
to invent it, whereas to speak the language of everyday is Simply to
use it.
All this is derived from the fact that in the cinema connotation is
homogeneous with denotation, and like it, is expressive. One is for-
ever shifting from art to non-art, and vice-versa. The beauty of the
film is governed to some extent by the same laws as the beauty of the
filmed spectacle; in some cases it is impossible to tell which of the two
is beautiful and which of the two is ugly. A film by Fellini differs
from an American Navy film'"-made to teach the art of tying knots
.. Let us not forget that there are thousands of films of this type.
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 83
to new recruits-through its talent and through its intention and not
through its most intimate semiological workings. Purely vehicular
films'" are made in the same way others are, whereas a poem by Vic-
tor Hugo is not shaped in the same way as a conversation between
two office workers: First of all, one is written while the other is oral;
a film, however, is always filmed. But that is not the main point. It is
because of its heterogeneous connotation (that is, the fact that it im-
parts a value to words that are in themselves nonexpressive) that the
gap between the functional use of the verb and its aesthetic use came
into being. **
Thus the impression of having on the one hand two realities (or-
dinary language and literature) and on the other only one (the "cin-
ema").t n ~ thus-finally-the truth of that impression. Verbal lan-
guage is used at every hour, at every moment. In order to exist,
literature assumes that a book must first be written by a man-a
special, costly act that cannot be diluted in ordinary activity. Whether
"utilitarian" or "artistic," a film is always like a book and not like a
conversation. It must always be created. Like a book still, and unlike
the spoken sentence, a film does not automatically entail a direct an-
swer from an interlocutor present to give an immediate reply in the
same language; and in this sense film is expression rather than sig-
nification. There is a somewhat obscure but perhaps essential solidary
.. Like those of the American Navy I have just mentioned, or even like the
technical fihns of the French Institut de Filmologie, or technological fihns in
general. The documentaries one sees in movie theaters are something else; they
are, in purpose at least, already art films.
.... In fact, even in the verbal order, pure denotation is very rare. Everyday
language carries strong connotations. In Le Langageet la vie (Geneva, 19:2.6),
Charles Bally analyzes the spontaneous expressiveness of everyday or "popular"
language at length and shows that in essence it is no different from literary or
poetic expressiveness. But this is another problem. The "gap" I am referring to
still exists-in the verbal order, not in the cinema-between expressive conno-
tation (whether it is "literary" or "ordinary") and pure denotation (i.e., the
inexpressive code of langue).
tIn Esthetique et psychologie du cinema, vol. I, Jean Mitry notes quite
rightly (p. 48) that the word "cinema" designates three different things: A
means of mechanical recording, which lies this side of art (animated pho-
tography); cinematographic art, which is also language (filmic fact); and, finally,
a means of broadcasting (cinematographic fact).
84 PROBLEMS OF FILM SEMIOTICS
relationship between communication (bilateral relation) and "arbi-
trary" signification; conversely, unilateral messages often depend on
(nonarbitrary) expression-a relationship that is easier to grasp. A
thingora being yields its singularity through expression in a message
that implies no answer. Even the most harmonious love is not a dia-
logueso much as itis a kind of duet. Jacques tells Nicole of his love
for her; Nicole tells Jacques of her love for him. They are therefore
not speaking about the same thing, and one says rightly that their
love is "shared" (divided). They do not answer each other-indeed
howcanonereallyansweraperson expressinghimself?
Shared, their love is divided into two loves, which yield two ex-
pressions. For Jacques and Nicole, expressing as they did two differ-
entsentiments, to evolve, rather than the give-and-take of a dialogue,
the agreement of a true encounter tending toward a fusion that abol-
ishes all dialogue, there had to be a kind of coincidence-hence the
rareness of the occurrence-rather than that interplay of influences
and after-the-fact adjustments by which a dialogue is characterized.
Like Jacques (without Nicole) or like Nicole (without Jacques),
films and books express themselves and are not really answered. But
if, using ordinary language, I ask, "What time is it?" and someone
answers, "Eight o'clock," I have not been expressing myself; I have
signified, I havecommunicated,andI havebeenanswered.
It is therefore true that we identify only one cinema, unlike the
double term literature-language, and furthermore that that one cin-
emaresemblesliteratureratherthan language.
CinemaandTranslinguistics:
TheLArge Signifying Units
By initiallycasting light on what the cinemais not, and thanks to its
analysis of language systems, linguistics-and especially that part of
it which leads to translinguistics (semiotics)-gradually allows us to
glimpsewhatthecinemais. Thesmallestfilmic unitis the"sentence,"
theassertion, theactualizedunit.Thissuggestscertaincomparisons.
A whole tendency of modern research, mostly in the straight line
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 85
of the Saussurianenterprise, hascome to beconcerned precisely with
the sentence. Joseph Vendryes
83
notes that a gesture of the hand is
equivalent to a sentence rather than to a word. Eric Buyssens makes
asimilar observation
84
about traffic signs and, more generally, about
all semes thatcannot bebroken down intosigns. ClaudeLevi-Strauss
defines the smallest unit of the myth-the"mytheme"-as the assign-
ing ofa predicate to a subject-thatis to say, as an assertion; he even
adds
85
that each mytheme can perhaps, when it is first transcribed
onto cards, be relevantly summarized by a sentence; and, at a later
stage in formalization, the "grand mytheme" is still, he says, a pack-
age of predicative relationships-in short, a set of sentences having a
recurrent theme. In his study of proverbs, already mentioned, Shim-
kin sees the smallest poetic unit in the largest coded linguistic unit.
Vladimir Propp's analysis of Russian folktales is undertaken in a sim-
ilarspirit.
86
Roland Barthes hasdefined the modern myth as a unitof
speech,87 and has emphasized-precisely in relation to the cinema-
the "largesignifying units."88 Georges Mounin
89
believes that certain
"nonlinguistic systems of communication" have become so important
in modernsociety that the time has arrived to undertakeseriously the
semiotic de Saussure had dreamt of (and this is what Eric
hadbeensayingalreadyin thevery first linesof his book), insteadof
dispatching it in a few sentences clapped on to the end of textbooks
of linguistics. Roman Jakobson believes that poetry could be studied
in a more linguisticspirit, on thecondition that linguistics in turn be
concerned with units larger than those of the sentence.
90
These are
all convergingperspectives. >I-
.. Since this articlewas written (February 1964) this trend has become even
more pronounced. One would now have to add to the list-to mention only
those contributions of a general, theoretical nature, and more especially those
applicable to significatory bodies other than verbal languages-the work of
Algirdas Julien Greimas and Luis J. Prieto, whose precise significance has be-
come more clearly apparent now that each writer has outlined his thought ina
coherent, over-all exp<>se: Semantique structurale (1966) for Greimas, and
Messages etsignaux(also 1966)for Prieto.Similarly,thegatheringofseveral of
Benveniste's articles into a single volume, Prohlemes de linguistique
generale (again 1966) has contributed to clarifying the notion of discourse.
Also, since 1964a numberofsemiological investigations havebeen undertaken,
which, on questions such as narration and discourse, pardy overlap on the
$ e......... . ....... . ----
86 PROBLEMS OF FILM SEMIOTICS
The "nonlinguistic systems" Georges Mounin refers t0
91
are those
of numbers (telephone, social security, etc.), traffic signals, cartog-
raphy, the symbology of tourist guidebooks, and advertising images.
Hedoes not speak of the cinema. Nevertheless he does observe that,
in the modern world, the image tends gradually to lose its original
decorative role and to acquire an informative function. Above all he
underscores the fact that many of these nonlinguistic systems are
ruledby a singlearticulation. "Semanteme by semanteme," he says,92
"never phoneme by phoneme." However it seems that many non-
linguistic systems are broken down into"sentences," rather than into
semantemes. Many, butnot all. ThoseMounin mentions, which can
beanalyzedintowords (like the symbols of international tourism sig-
nifying "restaurant," "hotel," or "garage"), doindeed exhibit a single
articulation. Itis these systems, as the authorsays, that justify Andre
Martinet'squestion: Cana perfect"ideographic system" exist, "alan-
guage that would no longer be spoken but that one would continue
to write," a "system in which the units of content would merge with
those of expression"-whereas the second articulation divides dis-
course into expressive units without corresponding content? In the
cinema,asinothernonIinguisticsystems, theunitsof contentarealso
"merged" with those of expression, but in a different sense-on the
level ofthe "sentence."* Any traffic sign is a sentence in the impera-
semiotics of film: literary studies, mythological studies, studies of narrativity,
etc., in France (see, for example, Communications, no. 8, special issue on the
structural analysis of the narrative, 1966), as well as elsewhere (Italy, the
United States, Poland, the Soviet Union, Czechoslovakia, etc....with
various resulting conferences). As for the cinema itself, the successive round-
table discussions (1965, 1966, 1967) conducted within the framework of the
NewCinmea Festival (PesllIO, Italy)haveallowedvariouscontributionstocome
to light, like those of Pier Paolo Pasolini or of Umberto Eco, which are men-
tioned elsewhere in this book. There was also the second volume of Jean
Mitry'sEsthetiqueetpsychologieducinema(1965).Etc.
.. I am not suggesting that each shot equals a single sentence. Thatis why I
have placed the word sentence between quotation marks throughout this pas-
sage. The "correspondence" between shot and sentence is on a global scale and
is derived from the fact that a shot is an actualized unit, a unit of discourse,
and is inherently dissimilar to the word. Thefilmic shot is of the magnitude of
the sentence, so to speak.
This is not the case-itis a notable difference within a deeper resemblance
87
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM?
tive mode, rather than a semanteme. The jussive may actualize as
clearlyas theindicative."DoNotPass!" Twoelements are identified:
the lexeme (concept of "passing") and the imperative morpheme,
which simultaneously actualizes and constitutes the sentence. This
double function of "verbs" and more generally of predicates (which
provide a lexemic content, and also constitute the statement as such)
has been studied by linguists like Jean Fourquet, Louis Hjelmslev,
Emile Benveniste, and AndreMartinet,93 especially inrelation to the
-withtraffic signs, which are discussedalittle further on in the text: They are
also of the magnitude of the sentence, but, in addition, it is possible to find
more precise equivalences between a highway sign and a sentence (like my
example, "Do not Pass"). Amongother things, this obviously derives from the
fact that a traffic sign constitutes a signifying unit that is poorer and easier to
analyze than cinematographiclanguage. Thatiswhy I hesitate to use the word
5eme in connection with the cinema (and particularly with the shot). This
term-in one of its acceptations at least, for authors like Bernard Pottier and
Greimas use it differently-isthe one that Eric Buyssens andLuis Prieto use to
designate precisely the units of signification ofthe magnitude of the statement,
such as one finds in various signifying systems; the statement proper thus be-
comes the specifically linguistic form of the seme. (On this problem, see my
article "Seme" in Supplement scienti!ique a1a Grande Encycfopedie Larousse,
\
19 , and "Remarque sur Ie mot et Ie chiffre. A Propos des conceptions de
68
Luis J. Prieto," in La Linguistique, 1967, vol. 2-texts that are not reprinted
in this book.) In many nonlinguistic signifying bodies one finds units clearly
t
different from the "word," and clearly on the magnitude ofthe sentence-upto
this point, exactly as in the cinema-butwhich are, moreover, finite innumber,
l
relatively easy to enumerate, and each one of which is equivalent, through its
semantic substance, to a sentence that can be more or less reconstructed (as is
the case with sign boards). The concept of the "seme," particularly since it
has been remarkably developed by Luis Prieto, seems to me to be henceforth
"ready" for theanalyses ofallinstances of this kind (and there are many). On
the other hand theconceptof the seme inits presentform could not be applied
to signifying bodies like cinematographic language where one encounters units
which, although they are of the magnitude of the statement, are infinite in
number, impossible to enumerate, and none of which permits an exact equiva-
lence with a sentence, but only very vague "equivalences" with a large seg-
ment of linguistic discourse comprising an indeterminate number of successive
sentences.I have, incidentally, nothingbetterto suggest for "replacing" thecon-
cept of the seme in such cases; one can only note that the general history of
semiotics has nothing to offer on this point so far (at least to my knowledge);
thatis whyI use paraphrases like "units of the magnitude of the sentence, but
which . . . etc." For thatmatter it may be that this lacuna is permanent (that
it is not really a lacuna), and that the difficulty derives Simply from the fact
that "cinematographiclanguage" hasnospecific unitson the level of the image,
butonlyontheleveloftheorderingof images.
88 PROBLEMS OF FILM SEMIOTICS
famous problem of nominal sentences, which can be compared to the
problem of close-ups in the cinema, or to the problem of sign boards
and signals: A telephone sign does not just mean "telephone" (a
purely lexical unit), but "telephone HERE." It is a self-sufficient state-
ment, which implies the existence of something in reality; it is there-
fore not a word.
One can be misled, however, by the fact that in systems that are by
nature poor, the breakdown of the "sememe" into sentences (and con-
seguently the absence, really, of the nrst articulation itself) is not
necessarily accompanied by a numerical multiplication of the units,
which remain small in number and more or less stable. This makes it
appear that there is an articulation-and in a certain sense there is
one indeed, since it is true that the units are fairly stable and they
can be at least approximately enumerated. But the discretion of dis-
crete units does not prevent them from being "sentences." It is the
natural poverty of the things signined that, in this case, guarantees a
kind of automatic economy rendering the nrst articulation superflu-
ous, and performing the same function as the articulation, yields the
impression that the poverty has been derived from the articulation-
for one lends only to the rich; furthermore, as Andre Martinet in-
sists, the nrst articulation provides verbal language with a function of
economy.94 As for the sememes of sign boards, they benefit from a
prior retrenchment, since the number of institutions designated is re-
stricted: In this case it is the small number of the referents that func-
tions as the first articulation. Since verbal language is a sememe con-
taining many more "things that have to be said," it therefore reguires
the first articulation in order to reduce the innnite multiplication of
sentences to the controlled amplitude of a lexicon. The cinema, like
language, has much to say, but, like sign boards, it actually escapes
the first articulation. It proceeds by "sentence," like sign boards, but,
like verbal language, its sentences are unlimited in number. The dif-
ference is that the sentences of verbal language eventually break
down into words, whereas, in the cinema, they do not: A nlm may
be segmented into large units ("shots"), but these shots are not re-
ducible (in Jakobson's sense) into small, basic, and specific units.
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 89
One can of course conclude that the cinema is not a language, or
that it is so only in a sense that is altogether too figurative, and, con-
sequently, it should not be dealt with through semiotics. But this is
a very negative point of view, particularly in the case of a social fact
as important as the cinema. The result of this attitude would be that
one would study traffic signals because they have a very obvious para-
digmatic structure, while paying no attention to a means of expres-
sion that after all carries a little more human weight than roadside
signs! The alternative approach is to look at the semiological endeavor
as open research, permitting the study of new forms; "language" (in
the broad sense) is no simple thing-whole flexible systems may be
studied as flexible systems, and with the appropriate methods.
Under these c"ircumstances-and despite the fact that the names
mentioned on these pages indicate a strong Saussurian legacy-prob-
lems of strict compliance can of course arise. But it suffices to point
them out. Naturally, anything that even approximately resembles a
linguistics of speech is a departure, it would seem, from the thought
of the Genevan scholar. The objection had to be pointed out. It is,
however, not insurmountable,'" and it would be to respect the great
linguist very narrowly indeed if one were to block all innovative re-
search under the pretext that one could not risk even grazing a study
of speech. And I say: grazing. For it often happens in the study of
nonverbal means of communication that the actual nature of the ma-
terial under consideration causes one to resort to a "lingUistics" that is
neither that of language nor that of speech, but rather is one of dis-
course in :Emile Benveniste's sense
95
(or even in the way Eric Buys-
sens used the word in a text in which he was attempting precisely
.. In the present state of research in linguistics and semiotics, it is becoming
less and less insurmountable. The fact that Chomsky'S work-and among other
contributions, his reformulation of the "language system/speech" duality in
terms of "competence/performance" with all that this implies-is now better
known in France is only one of the reasons. There is also the concept of dis-
course (neither pure language system nor pure speech) in mile Benveniste;
Zellig Harris's discourse analysis; Greimas's transphrastic analysis; Luis Prieto's
concept of the seme (extralinguistic unit of the same order as the statement),
etc. A certain, too brutal, or too literally Saussurian, concept of the "language
system/speech" dichotomy is becoming less and less tenable.
90 PROBLEMS OF FILM SEMIOTICS
to broaden the famous Saussurian bipartition in order to be able to
analyze more diverse "language systems"). Between words-pure
"sign events" as they are called in American semiotics, events that
never occur twice and cannot give rise to a scientific study-and lan-
guage (human language, or the more systematic and formalized lan-
guage of machines), which is an organized, coherent instance, there
is room for the study of "sign designs," sentence patterns,96 trans-
phrastic organizations, "writings" in the Barthesian sense, etc.-in
short, types of speech.
Conclusion
There have been, up to now, four ways of approaching the cinema. I
will leave aside the first two (film criticism and the history of the cin-
ema), which are foreign to my purpose even if some of their basic
notions, which fall under the category of "general ciIi\matographic
culture," are clearly indispensable to anyone who wants' to speak
about the cinema. One must go and see films, and one must of course
have at least an approximate idea of their dates. The third approach
is what has been called the "theory of cinema." Eisenstein, Bela Ba-
lazs, and Andre Bazin are its great names. The "theoreticians" were
either film-makers, enthusiastic amateurs, or critics (and it pas often
been pointed out that criticism itself is a part of the cinematographic
institution
97
): This is a fundamental point (about the cinema, or
about film, according to the case), whose originality, whose interest,
whose range, and whose very definition are after all derived from the
fact that the theory was made within the cinematographic universe.
The fourth approach is that of filmology-of the scientific study con-
ducted from outside by psychologists, psychiatrists, aestheticians, so-
ciologists, educators, and biologists. Their status, and their proce-
dures, place them outside the institution: It is the cinematographic
fact rather than the cinema, the filmic fact
98
rather than the film,
which they consider. Theirs is a fruitful point of view. Filmology and
the theory of film complement each other. There are even some bor-
der-line cases between the two, some of which are quite: important:
Who is to say whether Rudolf Arnheim, Jean Epstein, or Albert Laf-
THE CINEMA: LANGUAGE OR LANGUAGE SYSTEM? 9
1
fay were "filmologists" or "theoreticians?" Filmology proper also has
its great names: Gilbert Cohen-Seat and Edgar Morin. Both filmol-
ogy and the theory of the cinema are indispensable to the approach I
am proposing. Their division is justifiable only when it allows for a
reciprocity of perspectives. If it becomes a true separation, if it is
made into an antagonism, it can only be damaging. The major book
Jean Mitry has just published, Esthetique et psychologie du cinema,
a true sum of all the thought that the cinema has provoked up to
present, is an example of the deep reconciliation of these two comple-
mentary approaches, which one can only applaud.
Very much to one side of both filmology and the theory of the cin-
ema-unfortunately-is linguistics"" and its semiological extensions.
The discipline is an old one-it was known by Bopp and Rask-
and old age seems to suit it, since it is very much alive and well. It is
sure of itself-therefore it inspires confidence. That is why one has
sought its aid unhesitatingly-it can hardly be overburdened by a few
extra demands placed on it, and in any case, the study of the cinema
is far from being its only concern. It is a well-known fact that the
busiest people are always those who find the time to concern them-
selves with others-as Proust remarked about Monsieur de Norpois.
These few pages were written in the belief that the time has come
to start making certain conjunctions. An approach that would be de-
rived as much from the writings of the great theoreticians of the cin-
ema as from the studies of filmology and the methods of linguistics
might, gradually-it will take a long time-begin to accomplish, in the
domain of the cinema, and especially on the level of the large signify-
ing units, the great Saussurian dream of studying the mechanisms by
which human significations are transmitted in human society.
De Saussure did not live long enough to remark on the importance
the cinema has assumed in our world. No one disputes this impor-
tance. The time has come for a semiotics of the cinema.
".. In his Essais sur les princil'es ... (op. cit.), G. Cohen-Seat had very
clearly indicated what importance the linguistic approach would have for the
filmic fact. But there has been no development since then. One still speaks in-
nocently about "language" in the cinema as if no one had ever studied lan-
guage. Was Meillet then a service-station attendant, and Trubetskoy a butcher?