The Harpist in the Closet: Film Music as Epistemological Joke
giorgio biancorosso2009, Music and the Moving Image
When music first heard as background scoring ("nondiegetic" music) turns out to be coming from a source inside the fictional world of a film ("diegetic"), the switch may be hilarious, disconcerting, mildly surprising, or all of the above. Examples abound. Think of the "closeted" harpist in Woody Allen's Bananas or the fiddler in Ingmar Bergman's Fanny och Alexander. In this article, I address these and other examples in the context of the century-old debate on perceptual ambiguity. As is the case with ambiguous images, my examples elicit two conflicting perceptions of the same sensorial input. But unlike the textbook cases examined in psychological and philosophical discussions, they are also part of a rich and highly specific spatio-temporal framework. Embedded in a cinematic narrative, they both receive and project meaning. Far from being a curious, marginal instance of a well understood psychological phenomenon, the ambiguity between diegetic and nondiegetic music forces us to reconsider ambiguity in the light of a theory of knowledge. In Edward Yang's Yi Yi, over images of Ting-Ting arguing with her sweetheart on the street, we hear a slow, improvisatory introduction to a piano rendition of Gershwin's Summertime. 1 The music colors the scene with a sense of pastness and enhances the impression of distance conveyed by the extremely high angle of the shot. It is, one would think, a classic use of nondiegetic scoring, but a cut to a shot of Ting-Ting's auntie rummaging through her CD collection at home poses a question: was the music an instance of displaced diegetic sound instead, functioning here as a sound advance? Yes, but yet another cut takes us to NJ Jian's home, where Ting-Ting is now seen playing the very same music at the piano. The music, then, would seem to have been working ambiguously, not in the sense of being vague or indeterminate but rather of being equally fit to perform different tasks at different stages of a transition from one scene to the other: nondiegetic underscoring first, then recorded music, and finally a piano performance. Following a suggestion first put forward by Jeff Smith in these pages, one need not think of the music as ambiguous, but only initially misapprehended as the result of the absence of crucial information about its status and provenance. 2 Smith's suggestion relies on the classic formalist distinction between the narrative, or fabula, as an implied spatiotemporal world constructed by the spectator; and narration, or syuzhet, as the total sum of sounds and images that literally make up a film as it unfolds in time. Given this distinction, upon examining a number of cases of reversals or multivalence, Smith effectively explains away ambiguity as an attribute of the music. Ambiguity, put another way, impinges not on what sort of narrative space the music inhabits in the first place, but rather on how communicative the narration is; it lies not in the stimulus but arises out of the way in which the narration shapes our construction of the story world, and this is why it is eventually resolved. The status of the music in Yi Yi as a diegetic performance is never really in doubt; rather, due to the low level of communicativeness of the narration in this instance, it is revealed in successive stages. An ambiguity that is finally solved is no less ambiguous for that, however. In fact, it is common for ambiguous configurations to elicit contradictory or mutually exclusive perceptions, only one of which turns out to be right or at least plausible. As I will argue in this essay, multistability, the constant "flipping" of perspectives, obtains only in laboratories or under extremely controlled, and thus unrepresentative, circumstances. Moreover, something would be amiss in a reading that would privilege solely the site of provenance of the music as eventually disclosed at the expense of the process of disclosing itself, all the more so if such disclosure were to be foregrounded and thus were to call attention to itself as a nexus of meaning. In these cases, as Robynn Stilwell has argued recently, and as Edward Branigan pointed out a few years ago with reference to sound, "the misreading is relevant to-functions in-the reading of the text and in that sense is not a misreading at all but a necessary part of an ongoing process of interpretation." 3 When a reversal occurs, we do respond to the sudden shift, adjust to a new perspective, and take stock of its implications, thus recognizing, if only on the level of the fabula, our initial perception as a misreading. The experience of the shift retains its meaning, however, for it draws us in the picture, as it were, and is a reminder of the precariousness of our reliance on conventions. In the almost whimsical example from Yi Yi, for instance, the irreversibility of the successive changes undergone by our perception of the music, and the experience of not one but two successive false perceptions, intimate the possibility of having to reconfigure the world of the fabula surprise after surprise, as if endlessly at the mercy of the director-as-deus-ex-machina. The shot of Ting-Ting playing the piano would seem to settle the matter of where the music is coming from; but how can we be sure that the director, having done so twice already, will not pull the rug from under our feet once again? To be sure, if the film allows it, we will eventually forget about the possibility and comfortably settle into an attitude of trust in the stability of the world constructed on screen. But this is no more than a pragmatic solution to what is ultimately a metaphysical question, one that no empirical data or statistically ordered count will help us
