* I wish to thank Carolyn Abbate,
Masako Hayashi-Ebbesen, and Simon Morrison for their help in writing
this article. I am also grateful to two anonymous readers for their
thorough and insightful comments on an early draft of this paper. Finally,
thanks to Susan McClary and Jacqueline Warwick for their encouragement.
In Monteverdis Orfeo, the opening instrumental Toccata
is marked to be played thrice, with all the instruments before
the curtain is raised. The marking presumably reflects the functions
it was to play when the opera was first performed in Mantua: first,
to call the attention of the audience; second, to announce the arrival
of the Duke; and third, to sanction the beginning of the drama.
In discussing the role of music in bacchic rites, Nietszche went
of course well beyond this. He says: The music of Apollo
was Doric architectonics in sound, but only in the kind of hinted-at
tones characteristic of the cithara. It keeps at a distance,
as something un-Apolline, the very element which defines the character
of Dionysiac music (and thus of music generally): the power of
its sound to shake us to our very foundations, the unified stream
of melody and the quite incomparable world of harmony. In the
Dionysiac dithyramb man is stimulated to the highest intensification
of his symbolic powers, something that he has never felt before
urgently demands to be expressed: the destruction of the veil
of maya, one-ness as the genius of humankind, indeed of nature
itself (21). Nietzsches reference to the intensity
of the effects produced by music reflects the nature of the rites
he is discussing. However, it also betrays his view that music
functions not so much as a sign that a certain kind of
behaviour is allowed and indeed encouraged, but rather as a direct
cause of a certain state of consciousness. I will take up
the difference between causal function and mediating function
below (see also footnote 7).
This may well reflect the irreversible shift from cult value
to exhibition value postulated by Benjamin as a condition
of modernity. Artistic production, wrote Benjamin, begins
with ceremonial objects destined to serve in a cult. One may assume
that what mattered was their existence, not their being on view
Debussy orchestrated the first and the third Gymnopédies
in 1896, Roland-Manuel the second. Satie composed the original
piano pieces in 1888. It is worth recalling in this context that
the very word Gymnopedie refers to the dances of a
festival in ancient Sparta, thus linking the music programmatically
to a nostalgic celebration of lost, ancient communal rituals.
Allen exemplifies here what the Russian formalists called estrangement,
changing the place of things to heighten their effect. Allen was by
no means the first to adopt this strategy with respect to the beginning
credits. Both Ingmar Bergmans Persona (1966) and Robert
Altmans Thieves Like Us (1972), feature complex pre-title
sequences which call for a new way of understanding credits and their
music (or lack thereof).
In his essay on the credits sequence of Renoirs Une partie
de campagne, Roger Odin suggestively characterizes the title theme
as an enigma, arousing in the spectator the desire for
Here I am echoing Gilbert Rougets observations on the relationship
between music and trance: Demistifying the conception, too often
adopted, of the role played by music in inducing trance states will
be one of this books aims. The importance of music will not be
diminished for as much; quite the contrary. Music will ultimately appear
as the principal means of manipulating the trance state, but by socializing
much more than by triggering it (xviii). See also my observations
on Nietzsches interpretation of the Dionysiac dithyramb in footnote
French philosopher Etienne Souriau and his followers first elaborated
the notion of diegesis in specifically cinematic terms. It is
worth adding at this point that there is an unfortunate graphic
coincidence in the English language between the term diegesis
as used by Souriau, i.e. meaning story world, and
Platos notion of diegesis as pure narrative,
i.e. narrative without dialogue. In French, the former
is spelled diégèse, the latter diégésis.
See Genette 18.
The perception of cuts in the image track also depends on the viewers
allocation of his/her attention. Often the cut is not attended to as
such because the visual information contained in the new shot takes
up the whole of the viewers attention, especially if this information
is new or carries the narrative forward in significant ways. An extreme
instance of diegeticization of the cut occurs when the viewer
understands it as representing a characters shift of attention.
For Gombrichs classic discussion of the psychology of visual representation
see Art and Illusion. I hasten to add that Gombrich never intended
the term illusion to suggest that the beholder is deceived
or ensnared by the verisimilitude of the representation. He makes one
exception only in his discussion of the Greek story, if the painter
who had imitated grapes with such a degree of verisimilitude that the
birds came to pick at them (Gombrich 206). Psychologist James J. Gibson
discusses this example and Gombrichs interpretation of it in The
Ecological Approach to Visual Perception 281.
For a theory of representation as make-believe see Walton. I am heavily
indebted to Waltons discussion of imagining and I share his conviction
that the differences among varying forms of representation must
be seen against the background of their commonality, the fact that all
prescribe imaginings, generate fictional truths (Walton 51).
Here I am taking up Malcolm Turveys Wittgensteinian claim
that verbs such as to believe, to understand,
and so on, should be interpreted as referring to forms of behavior
toward concrete object or situations, not as occult mental processes.
That said, I believe that the verb to imagineæin
the sense defined above also points to a form of behavior, and
thus I feel no need to dispose of the term imagination
altogether, as Turvey seems at times to imply (See Turvey 456).
Scruton discusses imagination as a parallel to metaphor in language.
The experience of metaphor, in turn, is for Scruton a paradigm
of the double intentionality that permeates the experience of
music as well. [O]ne and the same experience takes sound
as its object, and also something that is not and cannot be sound:
the life and movement that is music (Scruton 96).
In insisting on the double intentionality of the perception of
diegetic music, I am adapting Scrutons claim to a new context
(see footnote 13 above). It needs adding that
throughout his book, Scruton shows no interest whatsoever in functional
or incidental music, nor in music that is artfully exploited as
part of a combinatoire of elements (to use Claudia
Gorbmans felicitous expression). This is a troubling aspect
of the book, especially when Scruton ventures into general claims
about musical representation, expression, and the relationship
between music and language. For upon embarking on such claims,
one should either define clearly what is meant by music or be
inclusive enough to cover a sufficient number of repertories,
uses, and ways of listening. Scruton seems instead to assume that
the reader will agree that music refers to so-called
Western classical repertory and that music listening
is best exemplified by absorbed contemplation of tones in a virtual
For a classic analysis of belief as a dispositional, as opposed to occurrent,
mental state, see H. H. Price, Belief, esp. chaps. 1 and 2, series
II. In a chapter called Half-belief, Price also offers an
illuminating analysis of the experience of make-believe
Noël Carroll and Murray Smith have both argued for the need
to sever the link between belief in the actuality of what is shown
on the screen and the capacity to engage cognitively and emotionally
This is also why such avowedly external elements as the written
credits or the voice-over reinforce what Christian Metz calls
the fiction-effect (and I will later call absorption).
They reassure the spectator of the distance between her and the
story world. See Metz,101.
In the terminology of philosopher Kendall Walton, the nondiegetic
music in Jaws makes it fictional that the shark
is attacking the unsuspecting swimmer. Diegetic music, on the
other hand, makes it fictional that there is music, that
a band is playing, for instance (172).
The effectiveness of the music cue is also dependent on ones memory
of the credits, where the music is used for the first time. Recalling
the music already heard is thus part of the listening experience and
reinforces our confidence in the intuition that the music is referring
to the shark attacking.
I am borrowing the idea of film music as vector from Cesare
Brandi, Teoria generale della critica, 270. Readers fluent in
Italian will find many illuminating observations on film music in this
important work of aesthetics. On this phenomenon, see also Claudia Gorbman,
32. Note that my use of the term vector differs from Michel
Chions term vectorization in his book Audio-Vision.
Chion uses vectorization to refer to how sound in general
endows the image with a unilinear temporality as well as a horizon of
expectation. See Chion, 1314 and 1820.
In his treatment of the relation between depth staging and attention,
David Bordwell unfortunately ignores the role of music altogether. See
See for instance Alfred Newmans observations as reported in Hagen,
Scott Paulin raises indirectly a similar question when he says that
anxieties over unity and effacement have usually made producers
of film more anxious than consumers [...] (64).
See, for instance, Stephen Heath, Questions of Cinema.
For sharp criticism see Noël Carroll, Mystifying Movies:
Fads and Fallacies in Contemporary Film Theory and Bordwell,
The famed orchestral pit of the Festpielhaus at Bayreuth, the
darkness of the theatre during performance, and the elimination of lateral
views to achieve frontal engagement with the stage may be interpreted
not as necessary steps to ensure heightened attention on the part of
the spectator but, rather, as signs or warnings that such
continuous attention was being solicited by the Puppet Master (i.e.
The expression mutual implication is Gorbmans and
it effectively highlights the fact that music too is transformed in
the process of being matched to a cinematic scene. See Gorbman, 15.
What I am theorizing here is implicit in two important historical studies
of live musical accompaniment by Martin Marks and Tim Anderson. See
Marks, especially chapter 2; and Anderson, 3-22.
The question is, of course, immense and not new to the 20thcentury.
Pondering why animated objects create terror, Carolyn Abbate has recently
made her own Stanley Cavells argument that automata suggest how
we could look down to find our own chests covered by brass plates, ripped
open to expose an elegant clockwork within. Thus the perfected
mechanical man robs us of a prize, our soul, and in so doing injures
human individuality and consciousness (476). Note that is terrible
here is the display of human features on the part of a mechanism, not
the mechanical per se.
The impact of the crime is considerably reduced when the scene is viewed
In Chapter Two of my dissertation, I examine at length the function
of film music as a prosthesis of the spectators mind.
This is not always the case, of course. In the films of Jean-Luc Godard,
the beginning and stopping points of the music are sometimes so unconventionally
placed that a narrational effect is replaced by ones awareness
of the hand twiddling a dial. For two illuminating discussions of Godards
use of sound, see Williams, 193208; and Bordwell, 330332.
It remains unclear whether the music is internal or external to the
diegesis and whether the primitive creatures first react to the sight
or to the sound of the monolith.
Ligetis Requiem was completed in 1966 and was still relatively
unknown when the film was released in 1968. It was unknown and unfamiliar
also in a literal sense to all but the most sophisticated and musically
In 1941, Bertolt Brecht ironically observed: The only thing that
one can say in defence of the dominance of so much music in film is
basically that no one hears it anymore (11).
See Chion, Audio-Vision, 40. Rick Altman has challenged Chions
complementary claim that the sounds of a film, taken separately
from the image, do not form an internally coherent entity on equal footing
with the image track (ibid.). See Altman, with Jones and