2. Revision of the concept and audiovision theory

The problems which emerge in relation to the concept of diegesis and the alleged distinction derive directly from the philosophical premises that we have set out. Since we are dealing with a concept and a distinction which are crucial to film analysis, it is important to tread carefully. It is usually the practitioners of cinema who propose doing away with the concept altogether; some argue that the terminological opposition of diegetic/nondiegetic serves no practical purpose (Thom 2007), although they are clearly motivated by a fundamental prejudice against the very idea of analysis and theory. Clearly such an attitude is of no help in any reflection on the concept. The fact that the concept can easily be eluded in practice does not by itself imply that it is totally invalid. Rather, such criticism should prompt reflection on the limits of its use, which in turn will help to reach a better understanding of its spheres of applicability. Better still, this reflection can help in achieving an alternative theorization of the discipline’s premises.

The concept of audiovision proposed by Michel Chion ([1990] 1994) is significant in this respect, and we can try to clarify its philosophical basis. In a hypothetical formulation, an entirely coherent theory of audiovision should be endowed with a constructivist conceptual framework whose foundations can be deduced, for example, from the philosophical elaboration of Nelson Goodman (1968; 1978). A theory of this type should first of all call into question the thetic attitude, focusing on the audiovisual narration as the primary construction. From this perspective diegesis could maintain some of its prerogatives while taking on a less ambitious ontological status, in keeping with its condition of secondary construction. Diegesis is based on an act of inference which cannot lay claim to any kind of objectivity: it is a subjective act, and hence merely hypothetical. As a matter of fact this theoretical perspective is not alien to the scholars who use the concept in the customary sense. Indecision concerning the position of certain elements of audiovision – particularly the audio – often conceals an arbitrary shift from a thetic perspective to one which is prevalently constructivist. A full commitment to the constructivist model would probably do away with most of the difficulties associated with the theoretical concept of diegesis. The result would be a reappraisal of the concept of diegesis and the diegetic/non­diegetic distinction: from being fundamental concepts, they would be deliberately turned into tools or theoretical constructs to be empirically negotiated from one situation to the next. This is in fact the orientation adopted recently by two scholars who have focused on the collocation of the diegetic/nondiegetic distinction in their own theoretical-analytical models (Smith 2009; Neumeyer 2009).

In terms of the theory of audiovision, the diegetic/non­diegetic distinction should renounce its topological significance and be reformulated in a functional sense. In the final analysis, distinguishing between what is diegetic and what is nondiegetic means abstracting from the audiovisual configuration as the primary construction, elaborating a theoretical model whose applicability is limited (detecting some relations while ignoring others) and consciously focusing on a secondary construction based on an act of inference. In other words, all the affirmations concerning the significance of the distinction have to be contextualised in a situation which is able to make the conditions of the concept’s applicability explicit. Moreover, from the standpoint of the theory of audiovision, the diegetic and nondiegetic aspects cannot be distinguished at the ontological level; rather, they cooperate in the audiovisual narration, within which they are constantly interacting. In any case, the constructive moment must be primary with respect to the representative moment, which is a derivative. The fact that in many cases (but not “always”) the traversing of the boundary “does […] mean” (Stilwell 2007: 186) depends strictly on the theoretical construction applied to the particular audiovisual situation, and not on the claimed perceptive objectivity of the diegetic/nondiegetic threshold. Deriving a universally valid concept of diegesis from this means passing off a merely hypothetical inference for objective knowledge of an actual reality.

It is only in the context of audiovisual narration that the audio can express all its aesthetic and constructive potential. In the most famous scene of Alfred Hitchcock’s Psycho (1960), “The Murder”, an analytical appraisal should consider not the represented contents but primarily the construction of the audio­visual sequence in the sense of an on screen narration in which Bernard Herr­mann’s music plays its full part. This music is clearly nondiegetic with respect to the diegetic sounds. If the sequence is deprived of the audio, we immediately see all the diegetic insufficiency of the images with respect to the audiovisual impact of the complete sequence. The piercing E flat in the violins – which in the space of eight bars organizes itself into a cluster involving all the strings on the subsequent three degrees of the chromatic scale spread over four octaves – is perfectly synchronised with the killer’s hand pulling aside the curtain. It is precisely the metallic grating sound, heard just before as the victim pulled the curtain to, which suggests the musical timbre. The clusters in the following eight bars, made all the more dissonant by rapid rising glissandi in all the instruments, prove to be not only narratively but also diegetically more convincing than the representation of the stabbing, an abstract sequence stripped of the potential violence of the blows. In a sense this has been absorbed by the nondiegetic music, even though there is no strict synchronization. The position of the music with respect to diegesis is perfectly clear, but its function does not stop here. It actually determines the audiovision in its representative capacity, and thus exerts an influence on the very possibility of inferring a diegesis, even though of course there are other elements which collaborate in equal measure: the woman’s screams, the noise of running water, the metallic rasp of the curtain and the slump of the moribund body. The very sense of the diegetic sounds is reconfigured in the interaction with the nondiegetic music. Thus it is perfectly evident that in terms of narrative the diegetic/nondiegetic distinction has a function which is quite marginal with respect to the sequence in question. We experience the threshold only in the limited degree to which we perceive its secondary status in terms of the aesthetic experience of the film, which is in fact our experience of the audiovision.