Scribd Logo
Fazer o upload

Bem-vindo(a) ao Scribd!

  • Temporal

    Enviado por

    Squaw
    2 visualizações38 páginas
    Temporal

    Direitos autorais

    © © All Rights Reserved

    Formatos disponíveis

    PDF ou leia online no Scribd

    Você considera este documento útil?

    Este conteúdo é inapropriado?

    Denunciar este documento
    Temporal

    Direitos autorais:

    © All Rights Reserved
    Baixe no formato PDF ou leia online no Scribd
    Sinalizar o conteúdo como inadequado
    2 visualizações38 páginas

    Temporal

    Enviado por

    Squaw
    Temporal

    Direitos autorais:

    © All Rights Reserved
    Baixe no formato PDF ou leia online no Scribd
    Sinalizar o conteúdo como inadequado
    de 38
    ‘Temporal Gestalt Perception in Musie James Tenney; Larry Polansky Journal of Music Theory, Vol. 24, No. 2 (Autumn, 1980), 205-241. Stable URL: httplinksjstor.orgsici?sici=01)22-2909% 281980239 2024%%3A2%9C 20523 ATOPIMISED0.CO%IBI-K Journal of Music Theory is curently published by Yele University Department of Music. Your use of the ISTOR archive indicates your acceptance of ISTOR's Terms and Conditions of Use, available at flip: feworwjtor org/aboutterms.htmal. ISTOR's Terms and Conditions of Use provides, in par, that unless you fave obtained pcior permission, you may not dowaload an cnt isus of @ journal or multiple copies of articles, and you may use content inthe ISTOR archive only for your personal, non-commercial uss. Please contact the publisher cegarding any further use of this work. Publisher contact information may be obtained at upsferwe.jstor.org/joumals‘yuden tel. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transtnission. ISTOR is an independent not-for-profit organization dedicated to creating and preserving a digital archive of scholarly journals. For more information regarding ISTOR, please contact support @jstor.org- up:thrwwjtor.orgy ‘Tue Fan 18.09:55:01 2008 TEMPORAL GESTALT PERCEPTION IN MUSIC James Tenney with Larry Polansky INTRODUCTION For the historian, time is not che undifferentiated “continuum” of the theoretical physicist, but 4 hierarchically ordered network of mo rents, incidents, episodes, periods, epochs, eras, etc.—te., time-spans whose conceptual “boundaries” are determined by the natute of the events or processes occuring within them (or of the historian’ inte pretation of these events or processes). Similarly for the musician, 2 Piece of music does not consist merely of an inarticulate stream of lementary sounds, but a hierarchically ordered network of sounds, ‘motives, phrases, passages, sections, movements, etc.~is, time-spans whose perceprua! boundaries are largely determined by the nature of the sounds and sound-configurations occurting within them. What 1s involved in both cases is a conception of distinct spans of time~ at several hierarchical levels~each of which is both internally cohesive and externally segregated fiom comparable timespans immediately preceding and following it. Such timespans (and the events or pro: esses which define them) will here be ealed temporal gestleunics (or "TGs" 205 In the years that have elapsed since the early papers on gestalt per- ception by Wertheimer, Kahler, and others", a considerable body of literature has accumulated which deals with the visual perception of spatial gestalt-units, although some of this Literature remains highly speculative. Much [eas has been written (even of a speculative nature) about the perception of fomporal gestalt-units. Some useful analogies hhave been drawn between visual and auditory perception, but such analogies provide litle insight into the besic mechanisms of temporal {gestalt perception, and many of the questions which might be the most relevant to musica! perception have not even been asked by perceptual psychologists, much less answered. How, for example, ate the per- ceptual boundaries of a TG determined? To what extent are the factors ‘bwolved in temporal gestalt perception objective, beating some measut- able relation to the acoustical properties of the sounds themsolves? Assuming that there are such objective factors, is their effect strong ‘eiuigh that one might be able to predict where the TG boundaries will be percelved, if one knows the nature of the sound-events that will occur? In an effort to provide some tentative enswers to such questions, a hypothesis of temporal gestalt perception will be proposed in See- tion 1 of this paper, and Section 2 will present some results of a com- puter analysis program based on this hypothesis. The program repre- sents a simplified model of this aspect of musical perception, and some of the implications, limitations, and possible extensions of this model will be considered ‘in Section 3. Although the hypothesis on which the model is based is very simple, it invalves some unfamiliar concepts and terms that will have to be explained before the hypothesis will bbe comprehensible. Some of these concepts were first stated albeit Jn eather embryonic form-in an eatlier papet* chough these have evolved consigerably in the intervening years." ® Others have emerged ‘more recently, in the effort to organize the more general music-theo- retical ideas into an explicit “algorithmic” form. Though T will not recount the history of the development of the model, I wil try to describe the conceptual transformations of these earlier ideas in a way ‘which paralielstheie actual historical development. 1. THE FUNDAMENTAL HYPOTHESIS, As in my earlier weitings,Ishall use the terms “element,” and “clang,” and “sequence” to designate TGs at the first three hierarchical levels ‘of perceptual organization. An element may be defined more precisely as a TG which is not temporally divisible, in perception, into smaller TGs. A clang is 2 TG at the next higher level consisting of a succession 206 of two or more elements, and 4 succession of two or more clangs— heard as a TG at the next higher level—constivutes 2 sequence. In the ‘earlier writings names were not given to TGs at levels higher than that of the sequence, but recently we have been using the terms “segment” and “section” for units at the next two higher levels. The TG at the highest level normally considered is, of course, coextensive with the piece itself, although situations are’ certainly conceivable where stil larger gestaltunits might be of inteest—e.g,, the series of pieces on & concert or the set ofall pieces by a particular composer. In Meta / Hodas (1961), 1 designated praximity (in time) and sim- dariey (with respect to any oral other parameters) as the two “primary factors of cohesion and segregation” involved in musical perception (or, more specifically, in clang-formation) a8 follows: In 4 collection of soundelements, those which are simultaneous fof contiguous will tend to form clangt, while relatively greater weparations in time will produce sepregation, other factors being equal, Those which are similar (with respect to values in some ppacameter) will tend to form clangs, while selative dissimiayity ‘vill produce segregation, other factors being equal Aside ftom certain other differences between these early formulations and my mote recent ideas (eg. that two or more simultaneous elements do not necessarily constitute a clang, but more likely what T would now call a “compound element,” several problems had to be solved before the current algorithm could be designed.” Fist, the principles, as stated, were not “operational,” but merely descriptive, That is, although they were able to tell us Something zbout TGs whose bound: ‘aries. were already determined, they could say nothing about the process by which that determination was made. They described the results of that process, but not its mechanism. Second, “similarity” was not defined in any precise way, except by reference to “values in some parameter.” The assumption here, of course, was that the simi- larity of two elements is an inverse function of the magnitude of the interval by which they differ in some parameter. This remains 4 plau- sible assumption, though it was never made explicit—hut even such a ‘corcelation of similatty/dissimilarity with interval-magnitude does not, by itself, allow for the simultaneous consideration of more than cone parameter at a time. This rather profound difficulty was implicit in the “other factors being equal" clause appended to the two state- ments. At the time, this qualification scemed necessary. in onder to rule out cases where two or more parameters vaty in coniicting ways, ‘or where two or more “factors” function independently. Although ‘this was 2 useful device for isolating and studying some important aspects of temporal gestalt perception, it imposed a very severe 207 limitation on the renge of musical examples whose gestalt structure rolght be predicted. In most real rousica situations, ather factors are ‘manifestly nor equal, and our perceptual organization of the music is 1 complex result of the combination and interaction of several moce- orless independent variables, Third (and finally), these early formulations referred to one hier- auchical level only—the grouping of elements into clangs—although it was obvious to me even then that the similaity-fuctor, at east, was of treat importance in the pesceptual organization of TGs at ail higher levels. Tn a later peper,* an attempt was made to generalize these principles, restating them ia a way that would be applicable to all hierarchical levels. Thus (fram Proposition TI, p.4), The perceptual formation of TGs at any tievarchical level is de- termined by @ number of factors of ecitesion and segregation, the ‘most important of which are proximity and similarity; their effects niay be described as follaws: relative temporal proximity [ard] relative similarities of TGs at a given hierarchical level will tend to sroup them, perceptually, into a TG at the next higher level, Con- versely, relative temporal separation andjor differences between TGs will tend to segregate them into separate TGs at the next higher level Although these Jater “propositions” served to extend the exrlier formu: lations ta higher Jevels, they suffered all of the other deficiencies of the later formulations: they were non-operational in character, impze- cise with cespect to the concept of “similarity.” and restricted to one Parameter (ar factor) at a time. The first of these problems has been solved by 2 shift of emphasis from the unifying effects of proximity and similarity to the segrepative effects of temporal separation and parametcle diseimilarity, and by a ‘more careful consideration of these effects as they must accur in real time. In the ongoing process of perception in time, TG-boundaries are determined by successive TG-inttitions. This obviously applies to the beginning of a TG, but also to the end of it, since the perception that i has ended is determined Gin the monophonic case, at least) by the ereption that a new TG at that same hierarchical level has begun, In this new light, the effect of the proximity-faetor (at the element/ clang level) might be restated as Follows: In 4 monophonic succession of elements, 2 clang will tend to be initiated in perception by any element which begins after a time- interval (irom she beginning of the previous element, ie. after 4 delay-tome) which is greater than those immediately preceding and following ic, “other factors being equal.” 208 ‘Thus, in mm, 24-28 of Varése’s DENSITY 21.5 (Example 1), where clangiinitiations are determined almost entirely by the proximity- factor, it can be seen that the clements which initiate successive clangs ate i fact, icvariably those whose delay-times are “greater than those immediately preceding and following” their own (the delay-times associated with each element are indicated in the example by the rtumbets below the staff, in triplet sixteenthnote units; those which sue ciscled are for the clangénititing elements). Note that the first ‘occurrence of D (at the end of m. 25) does not initiate a new clang, in spite of its fairly long delay-tie (12 units), because the delay: time which follows itis still longer (19 units). As stated above, the proximity-factor begins to take on a form that i “operational.” In 4 musical situation where no other parameters are varying (say, 2 drum solo, at constant dynamic level), this principle can provide an lunazubiguous procedure for predicting clang-boundaries. Tn an analogous way, the effect of the similarity-factor (at the clement/clang level) may_ be reformulated 2s follows (and note that this statement can actually inchide the previous one as-a special case, if the parameter considered is time, and the “interval” isa delay-time): Ina monophonic suecession of elements, 2 clang will tend to be initiated in petception by any clement which differs from the previous element by an interval (in some parameter) which is greater than those (interelement intervals) immediately preceding and following it, “other factors being equal.” ‘This, too, is “operational,” in that it suggests an unambiguous pro- cedure for predicting clang-boundaries, though it is limited to special cases where only one parameter is varying at a time. Consider, for example, the first 12 measures of Beethoven's Fifth Symphony. Bx- ample 2 shows the melodic fine, abstracted from all eontrapuntal/ textural complications—as it would be heard, say, in a piano transcrip tion, Because of the considerable difference in tempo here, compared to the Varése example—and thus in the actual duration of notated time-valuos-relative weights are used that give the value of 1 to the ‘ighth-note (as well as to the semitane, as before), The clanginitiations during the first six bars are obviously determined by the proximity- factor alone, but beginning in m. 6, the proximity-factor can have no effect on the clang-organization (except in m, 9), because the delay- times ate all equal, This passage is not heard simply as two clangs, hhowaver, but as 2 succession of clangs (indicated by the brackets above the staff), each consisting of four elements, And note that, for every clanginitiating element, the pitch-interval associated with itis greater than those immediately preceding and following i. “The parallelisa: of the proximity- and similarity factors, as restated 209 (21-9 “unui spraronur-qond Aq vay S-1 uy) soun Keep 44 14 paula suareRTEEUeTD “z adureng 0 0@r 50 500@ o0@t 00@» 00e@ roto fot | cirrerperre peer arte @ @ 1! Ld | | : fy SE te eee sowp-dojp Aq pouuarap suopeprUETuMy “1 awry © @ : ‘- _ in Ti oe. + ——; a 7 we te Pian fig +5 SSF — above-and the fact that the second staiement can be considered to include the first one as a special ease—is extremely important. In both, itis the occurrence of a local maximum: in interval magnitudes which determines clanginitiation. An interval is simply a difference, and ‘whether this is a differance in stating-times, or pitch, or incensity— cor any other attribute of sound—is not what is important. Rather, it is relative differences {in any parameter) that seem to be crucial. We live in a “universe of change," but whether a particular change marks the beginning of a new temporal gestalt-unit oF simply another “turn” in the shape of the current one depends not only on its sbsolute magni- tude, but on the magnitude of the changes which precede and follow “The restriction to one parameter (or factor) at a time, stil implicit in the last formulation, remains to be overcome before our principle ccan be of much use in predicting clanginitiations in any but a very Tintited set of musical situations. What is needed is some way to com- bine or integrate the Interval magnitudes of af parameters into a single measure of change or difference. The solution to this problem involves 4 concept that has been employed by experimental psychologists for several decades now~that of a multidimensional psychological oF perceptual “space. The “dimensions” of this space are the several parameters involved in the perception and description of any sound, 1. time, pitch, and intensity. Other parameters (cg. timbre) could be added to this list, if they satisfy certain conditions, but T shall limit my discussion here to these three basic ones. The set of parametric values characterizing an element serve to locate that element at some “point” in this multi

    Você também pode gostar