The salient, indisputable fact about philosophical studies of the dance is their conceptual poverty. I think it is accurate to say that the dance is the single principal art that is either very nearly unmentioned in comprehensive overviews of aesthetics or else treated (almost as a second thought) by way of adjusting arguments strongly and directly grounded in the other arts—principally, drama and music—or, by way of notions of representation and expression, linked even with the analysis of the literary arts.

     There are, also, some natural difficulties in theorizing about the dance. For one thing, there can be no question that notational efforts at scoring a dance are radically less interesting intrinsically than musical scores or the texts of dramas. The reason undoubtedly lies with the central importance of the tempered keyboard in Western music and the use of an articulated language in drama. For the sequences of actual notes in a musical score cannot fail to exhibit formal properties (often quite complex) that may be detected even in the absence of acoustic assistance, and the internal properties of the elements of language are, as the history of criticism and appreciation attests, very nearly inexhaustible. Dance scores are primarily heuristic devices for recovering a minimal sense of the principal positions and movements of a given dance; and, aside from the classical ballet, which has its own sense of a finite vocabulary of distinct bodily positions and movements and of the history of its own evolving style, the notational convenience of Labanotation, for instance, has rather little to do with the formal conditions of actually learning to dance or to perform a particular dance (see Hutchinson 1977; cf. Cunningham 1968). Dance notation, on the whole, tends to capture the movement of a dance by drawing the path or sequence of trajectories that statically approximates it—very nearly in the sense in which Henri Bergson despaired of any conceptual capture of human movement. Although there are efforts to correlate movements and distinct expression notationally, such efforts appear to be noticeably awkward and unconvincing (cf. Sirridge and Armelagos 1977: n.7). Furthermore, there is no sense in which the perception of related positions and movements, ordered perhaps in an interesting way and legible from the notation itself, could possibly be grounded in an understanding of the deeper structures of the dance, by means of a closer attention to the notation itself: no such structures are there presupposed, in any sense comparable to the structures of music and language. The reason seems to lie with the emphasis on visual recognition tout court in the notation and the requirements of actually generating dance movements in terms of the dynamics of motor activity controlled proprioceptively.

     This cannot be taken as a weakness in the dance, only as a weakness in attempts to treat the dance primarily as an art that can be satisfactorily construed as allographic—in Nelson Goodman's sense (see Goodman 1968: 121-22)—or as a performing art suitably subsumed under generalizations readily drawn from music and literary theatre. It is abundantly clear that even musical notation fails to achieve its ideally allographic function, insofar as it is both impossible to exclude notationally variations of tempo and tone that are musically pertinent, produced in apparent accord with notation, and quite irrelevant to insist on the allographic as far as the reidentification of particular compositions, or the serving of musical interests, are concerned (see Webster 1971, Margolis 1980). There are, of course, other complaints about putative allographic restrictions on identity—in particular, those bearing on the familiar practice of omitting portions of a drama or musical composition or dance, or of replacing dance steps or instruments or the like, in order to accommodate local contingencies—without the least damage to fixing the identity of a piece. Hamlet is often cut in performance; the Brandenberg Concerti are often played with modern instruments; the thirty-two fouettés that appear in Swan Lake are often reduced to sixteen, the remaining measures replaced by other steps. In fact, Georges Balanchine has quite regularly replaced steps in the choreography of well-established dances of his that, in rehearsal, seemed, because of the overriding movement and spirit of those very dances, to be required.

     These reflections point to a second feature of the dance that tends to set it somewhat apart from the principal performing arts. The divergence among performances of what purports to be one and the same dance tends, if anything, to be less easily discounted than the counterpart phenomena in the other arts. I should not deny that divergent performances or interpretations obtain in music and drama as well; but the abstract constraint of a common keyboard and a common language encourages the fiction that relevant divergencies in music and drama are either practical concessions to the imperfect conditions of performance (for instance, regarding acoustical differences between concert halls and the convenient slackness of requirements on tempo and pitch) or else degrees of interpretive freedom permitted once conformity to certain strict but minimal requirements is achieved. The illusion that this holds for dance as well may be seen in the following remarks by Mary Sirridge and Adina Armelagos, primarily intended to oppose what they term "the modified expression theory":

Excluding the case of explicit mime, which is remarkable precisely because it is not continuous with the traditional vocabulary of ballet, the ballet dancer may be said to be concentrating on two things: doing the right thing and staying in line. (Sirridge and Armelagos 1977: 16)

With an important qualification, Sirridge and Armelagos follow Goodman's (1975) theory of style, holding that "[a] general dance style, such as ballet, solves the problem [of expression] by providing the dancer with an inventory of movements or sequences of movements, which we will call a 'spatial vocabulary'" (Sirridge and Armelagos 1977: 18). Against the "classical expression theory," they hold that "the dancer [need not] actually be feeling what he projects;" and against the "modified expression theory," the dancer need not be concentrating on expression at all. In fact, they say, "[I]t is quite important that the dancer not concentrate on expressing something or 'playing to the audience'" (Sirridge and Armelagos 1977: 15, 17). They go so far as to hold that "the expression or projection of personal feeling or emotion has nothing to do with the dancer's expectations of himself or with the focus of the dancer's artistic concentration." He restricts himself somehow "almost entirely" to "the kinesthetic sensations associated with executing steps properly" and to the comparable "sense of being en rapport with the other dancers or with the music" (Sirridge and Armelagos 1977: 16).

     There is no reason to hold to the classical expression theory. In fact, it is rather difficult to find advocates nowadays. The modified expression theory is similarly occupied with a contingency that is either marginal or associated with the work of a special sort of dance, possibly not unlike the mime Sirridge and Armelagos have in mind. But their emphasis is designed—deliberately or not—to conform with the allographic model of the dance; for, by the ordered regularities of a putative style, they suppose that the aesthetically relevant expressive qualities of a performance may be grasped by a sensitive audience without affecting the sense in which performing dancers conform only with the notational requirements of a given piece. The movements and positions of the body, then, notably in the classical ballet but by no means confined to that, serve as the extensional counterpart of sequences of musical notes and of sentences.

     But this is a decidedly defective view—in fact, a view which Sirridge and Armelagos in effect undermine in an apparently contemporaneous paper. (We shall turn to it shortly.) There simply is no reliable correspondence between a dancer's performing a set of movements in accord with a mere notation and an audience's seizing the expressive qualities somehow conveyed by those movements; and there are no kinesthetic sensations that could guide a dancer with regard to much more than congruence with a rhythmic beat and with the beat of other dancers. Also, a dance coach or director normally does not correct a dancer with respect only to the notational minima of a dance; and to say that he corrects the dancer with respect to the requirements of style is a) to say, at least, that the dancer is expected to attend to the expressive features of a dance; and b) not to say that the style invoked is systematically or algorithmically coordinated with the performance of movements demonstrably correct on notational grounds.

     The body is not an instrument for a dancer in the sense in which a violin is for a musician or a brush for a painter. We cannot help but see the peculiarly intimate and sui generis sense in which the dancer and the mime use themselves in performance. The same sense is at work in the singer, the actor, and the instrumentalist—and even, if the point needs to be pressed, among visual and literary artists; for all artists compose and perform with the organized materials of their experienced life. To watch Rampal perform on the flute, for instance, is to appreciate the sense in which he is obviously in command of the idiosyncratic energies and habits of his own body that he must tap in order to bring his notes into accord with the notational requirements of the score and his own conception of the appropriate style of a given composition. The authority of a performing artist characteristically extends to the invention of an expressive style, not merely the confirmation of a canon—which, in any case, is too subtly linked to the niceties of particular performances to be captured generically. In the same regard, the dancer does not work simply with positions and movements, but with the naturally expressive use of positions and movements generated (not altogether consciously) by that person's use of his own body through the accumulated grooming of a continuous life; and, imposed on, and altering, this natural and acquired expressiveness, individuated through an exclusively privileged experience with a single body, the dancer learns the expressive possibilities of different styles and the way in which his own performing body, stylistically informed, is viewed by knowledgeable observers.

     Such considerations obviously affected the appraisal of Tanaquil Leclerc's efforts to play the principal role in Swan Lake. Leclerc's tall and extremely slim body naturally trained in a sort of humorous and informal manner had much more to overcome than the more classically proportioned body and more formal manner of Maria Tallchief. In this sense, quite contrary to Sirridge and Armelagos, since the body is naturally expressive (essentially in the sense that Wittgenstein emphasized), dancers are obliged to be concerned with the expressiveness of their movements—even if both the classical and modified expression theories are false and in addition to whatever convention about denoting emotions and the like may obtain. Leclerc managed to produce a rather deviant version of the Swan Queen, whose expressive features may very possibly have required an adjustment in taste and perception precluded by her sudden malady. The adjustment, however, was no different in principle from the general diachronic shifting of balletic style through the historical influence of strongly authoritative dancers. Still, it is certainly clear that she produced the requisite movements of a relevant notation, exhibited the generic features of the Romantic style favored in the transatlantic version of the Russian ballet, and remained kinesthetically en rapport with other dancers. What Sirridge and Armelagos fail to account for is, precisely, that the Leclerc performance had to be orchestrated around the novel bodily expressiveness of Leclerc's manner of dancing; and this both required attention on Leclerc's part to expressive considerations and now exposes the sense in which style and expression in dance cannot be satisfactorily captured as the correlates of correct movements. Here, it is extremely important to notice—it is often misunderstood—that the expressiveness of performance is intentional in a sense that cannot be satisfactorily captured in terms only of the psychologically deliberate or consciously intended, while at the same time it cannot be satisfactorily represented by any extensional notation. It has to do rather with the individuality and idiosyncrasy of human expressiveness as the intersection of biologically and culturally contingent processes, raised at least to an essential place in the medium of the dance. Conceptually, this signifies that the history of production, including particularly the personal resources and expressive motivation of individual artists, tends to dominate the characterization of whatever may be called the style of performance.

     There is, of course, a logical distinction between denotation and expression: neither entails the other. But in a highly conventionalized art like the classical ballet, what is reflexively representational at the level of mental states tends to be expressive because it is denotative. This is, in one sense, a matter of style; but it is neither necessary nor even characteristic of more recent ballet, which has accommodated the freer forms of so-called modern dance. It is rather more a matter of certain genres of representation. The expressiveness that concerns the natural movement of the body belongs, first, to the human animal and, second, to the groomed member of a particular culture. It is conceivable that, in very highly stylized societies, conventions obtain that are at once both denotative and expressive of emotion—perhaps in the public behavior of the Geisha or of the temple dancers of Southeast Asia, for example—in which the natural expressiveness of one's own body is suppressed or ignored. Logically, however, that expressiveness is implicitly recognized wherever, with regard to actual mental states, the expressive and the denotative are not viewed as coextensive. But when the division obtains in the dance—conceivably, even in the classical ballet, if the point about Leclerc and Tallchief is well-taken—"personal style" cannot be treated as "a dancer's characteristic articulation of a more general spatial vocabulary" or of some generic style that fixes the particular denotative vocabulary of a distinctive form of dancing (Sirridge and Armelagos 1977: 19). For that thesis involves a confusion. There may be a conventional representational system of beliefs and emotions, in which what is expressed is no more than what is denoted; style in this sense will be correlated with correctly executed movements, as Sirridge and Armelagos claim. But where the expressiveness of bodily movement intrudes, either in the sense of natural expression or in the sense of culturally favored modifications of such expression (not incorporated in the denotative conventions of a dance), what we designate as "style" cannot, for conceptual reasons, be merely correlated with the proper execution of notationally specified movements. There is, therefore, a deep equivocation about the notion of style: it may refer to a conventionally developed system of denotative (and therefore, expressive) correlates of notationally specified sequences—which of course relieves the dancer of any need to attend to expression; or it may refer to a mode of performance that combines the natural and culturally groomed expressiveness of the body with some approximate denotative convention—which of course imposes on the dancer a need to attend to the proprieties of expression. In the latter sense but not the former, expressive qualities need not be construed in terms of denotative symbols at all, though they will clearly depend on the conventions of a particular culture. They are, when present, literally present—as extensions of the naturally expressive—in whatever sense cultural phenomena are actual phenomena irreducible to purely physical systems, with respect to which they might well have been treated as metaphorical.

     Interestingly enough, Sirridge and Armelagos qualify their acceptance of Goodman's well-known thesis about expressive qualities, namely, that a work expresses those properties that it metaphorically exemplifies (Goodman 1968: 85). For, conceding the natural expressiveness of the body, they are prepared to admit that "human movement may be literally sad." They go on to say that a good performance of Swan Lake may have properties that "are surely metaphorically exemplified, hence expressed. With others [that is, other properties], the answer is not at all clear-cut. Since both kinds of properties are aesthetically important ones to which the work is symbolically related, it makes little difference" (Sirridge and Armelagos 1977: 18). Here, however, we see how the complexity of the dance refuses to yield to their formula and how, at the same time, it betrays a characteristic weakness in Goodman's theory of expression. For, if a certain movement may be literally sad, then it must "possess" the property without needing to refer to it; in that sense it possesses it without exemplifying it, without being in any way symbolically linked to it. But that, precisely is what is denied by Goodman (whom Sirridge and Armelagos profess to follow) (see Goodman 1968: 85-86; cf. Margolis 1980: Ch 5). The point is not a mere quibble, however. For, if the performance of a dance exhibits such properties, dancers cannot restrict themselves to what may be notationally represented. Given that dance notation has a rather restricted use, refining our sense of the nature and role of expression points directly to the sui generis nature of the art of dance.

     The mystery of this aspect of the dance, the intrusive expressiveness of the body, of which the dancer may not be initially entirely cognizant, is probably most nearly captured in Merleau-Ponty's well-known essay "Eye and Mind" (1964)—except that Merleau-Ponty never mentions the dance and apparently believes (for rather curious reasons) that the art of painting best exemplifies the primordial sources of artistic expression. Nevertheless, his insight concerns the interior, the lived, effort of the body to orient itself at once perceptually and in movement, with regard to external objects and other incarnate systems, originally and also in a way that cannot completely overcome the essentially tacit dimension of our bodily orientation and expression.¹ Put this way, the theory of the dance and of bodily performance cannot be completely formulated without attention to the deeply intentional, ultimately autographic, features of the expression and significance of bodily movement itself. In effect, notably in our own time, the dance cannot be appreciated without some sense of how movements are actually generated; and this perception, however disadvantageously managed in third-person situations, relies very heavily on our sense of the manner of movement of distinct and particular bodies. This is why, perhaps, one feels that Martha Graham could hardly have been replaced in certain of her dances. In a second paper, Armelagos and Sirridge admit as much about Graham's "Frontier," though they persist in thinking that a dancer's personal style (style₂) is, somehow, "both an articulation and a further development of style₁ [that is, an inventory of acceptable positions and position sequences, including, one supposes, denotative import]" (Armelagos and Sirridge 1978: 130-31). On their own admission and on the preceding argument, moreover, this is quite impossible: either dances of Graham's kind cannot be notationally fixed, for if they could, another could dance them; or else what may be noted must be confined to questions of style₁, which then loses for us the specification of what is often crucial both to the appreciation of a particular dance and to a group of its distinctive properties. It is of no importance that Armelagos and Sirridge hold that "some surrogate explicitly designated by [Graham]" (Armelagos and Sirridge 1978: 130) might replace her, because, precisely, the intended sense is that the lucky replacement could not be judged suitable on the basis of notationally specified details. It is also not particularly telling that they hold, reasonably enough, that José Limón's "The Moor's Pavane" could be performed by dancers other than Limón and his company; although, allegedly because it failed to preserve the intended style, the classical balletic performance of "Pavane" by Nureyev and Fonteyn either radically transformed the work or in a sense destroyed it (Armelagos and Sirridge 1978: 131). The reason is simply that it is not essential that token instances of the same (type) composition agree in all aesthetically relevant features, or even (again, contra Goodman) that different performances exhibit some common set of necessary and sufficient features in order to be construed as "compliants" of the same work.² In fact, there is every reason to believe that the performing style of nineteenth-century ballets, to the extent that they may be thought to have been preserved, has evolved under the pressure of the personal style of leading dancers. And, although on an argument already given, the dance may be less accommodating notationally than music and drama (because it is more autographic in spite of being a performing art), there is no antecedent reason to preclude a classical rendition of "The Moor's Pavane." If it failed, Nureyev's did so because it was a failed performance—even an inappropriate one—not because it was in principle impossible to perform it in such a different style. Or, alternatively, if the classical balletic style was indeed inappropriate, other styles might have proved unexpectedly congenial; and other dances might show a greater hospitality toward divergent styles of performance. It is hard to see that a Shakespearean performance of Julius Caesar and a contemporary one—even Orson Welles's modern-dress version—could not fairly be construed as alternative, stylistically quite different performances of one and the same drama. In any case, resistance to the propriety of the Nureyev version can hardly be restricted to notational constraints and points in the direction of the autographic.

     The upshot is that, once we give up the notational constraint on the reidentification of a dance from performance to performance, there is no logical necessity or advantage in holding that personal style (in effect, expressiveness) is nothing more than, or is primarily, "an articulation and a further development of style₁," or that considerations of style₁ set fixed notational constraints on the identity of a dance. Treat the dance in a frankly autographic way and one accommodates at a stroke the history of the dance (and of every other performing art) and the peculiar importance of personal style in the dance (even in the classical ballet, where perhaps the strongest resistance might have been possible).

     This, then, is a third distinctive feature of the dance. Adjusting Goodman's category considerably, it is a largely autographic art (with some notational facilities) rather than an allographic one. I say that Goodman's category must be adjusted, but then it must be adjusted on independent grounds. For Goodman introduces two quite different notions of the autographic/allographic contrast: in one, art is autographic "if and only if . . . the most exact duplication of [a work] does not thereby count as genuine"; and in the other, a work ceases to be autographic, becomes allographic, "only when . . . [it] is fully determined independently of history of production, in terms of a notational system" (Goodman 1968: 113, 198). Insofar as there are no works of art that are "fully determined independently of history of production," there are no allographic arts, though there are arts with partially allographic features. All artworks are such in virtue of some history of production, even the intention that they be susceptible to notational constraint. But apart from this, it is certainly clear that contrast with forgery is not essential to considerations bearing on the history of production. If one stressed the notational features of the dance, then it would still be possible to insist that an unauthorized dance performance of Graham's "Frontier" might well be a forgery—much in the sense in which a printing of a Dürer etching, pulled from the true plate but unauthorized, might be a forgery. And if one stressed the productive history of "Frontier," then one could still insist that there was no useful sense in which an unauthorized performance must be a forgery. Useful as the notion of forgery is in obliging us to admit the nonperceptual features of all the arts, it is not really central to the effort to characterize particular arts as autographic or allographic. The truth is that ascriptions of forgery—even more benign characterizations such as "in the style of," "of the school of," "influenced by"—are contingently affected by prevailing practices, productive control, market value, and the like, in a way that cuts across the distinction. Forgery, for instance, is a perfectly reasonable charge against the reproduction of unauthorized copies of couturier designs—which (contra Goodman) are particularly apt to be regarded allographically.

     Armelagos and Sirridge appear to concede the point of the difficulty of construing the dance as allographic, in the sense that Goodman favors, that is, that a dance score is more or less the equivalent of a music score. They stress the unlikelihood of incorporating, notationally, reference to such "integral 'incidentals'" as the bearing of the accompanying music, lighting and costumes, and—in the same breath—the personal style of individual performers. But their intention, ultimately, is to correct Goodman's thesis, not to abandon it, for they go on to argue that "choreography, correctly understood, is one of the main criteria for identifying a performance as a member of the class of allographic compliants [italics added] which counts as a performance of the work" (Armelagos and Sirridge 1978: 133). If one thinks of the work of Fokine and Nijinsky, however, it is difficult to see how the choreography of a particular artist can fail to anchor the very features of a work in the history of its production, in features of personal style that cannot satisfactorily be incorporated into some generic notation—even if the choreography yields a partially instructive notation. Armelagos and Sirridge admit as much in at least some cases, but they are committed to the gradual development of the dance as a fully allographic art. The principal counterconsiderations are reasonably clear, however. Even the notion of a style, style₁, loses all sense of an allographic nature if, where relevant, it is produced by a dominant choreographer, is associated with a particular dance ensemble, tends not to become a canon for the profession as a whole, is closely linked with the personal expressiveness of this or that dancer, is principally taught by example, is noticeably easily lost in performances guided solely by the notation of correct positions and movements, is essentially linked with a relatively transient spirit or style of some distinct society or group, maximizes improvisational forms and the recognition of stylistic congruence within actual performance. Armelagos and Sirridge speak of style as a choice of "spatial vocabulary and characteristic kinesthetic motivation;" but, in effect, they admit that, at the present time, it appears quite impossible to free reference to "kinesthetic motivation" from biographical reference to a particular choreographer, that is, from the personal expressiveness of the style (style₂) of a particular choreographer. "Style as we construe it," they concede, "is difficult, if not impossible to record in a notational scheme"; but if so, they go on, "then identity criteria based solely on notational scores are inadequate" (Armelagos and Sirridge 1978: 138). Nevertheless, they look forward to "the increasing homogenization of dancer training already underway," which, they feel, "will gradually free dance works from the idiosyncratic control of their creators and increase the number of persons who can adequately interpret inadequate scores" (Armelagos and Sirridge 1978: 138). But this may well be a betrayal of the inherent difficulties they themselves take notice of, as well as of the essential sense in which dancers use their personally and culturally idiosyncratic selves as the very medium of their art—not steps, movements, positions, or styles primarily focused on denotative and symbolic import. What reason is there for thinking that the evolving styles of the ballet or the plural styles of modern dance can be fixed canonically and freed from the expressive styles of an enveloping society? The two-storey picture of style₁ and style₂ is intended to accommodate the allographic conception, but it fails utterly to come to terms with the obvious fact that the dance is not insulated against the world and that dancers are recruited from the living culture that sustains it. One need think only of the eclectic, mixed, and pop-culture features of the work of Jerome Robbins, Anna Sokolow, and Twyla Tharp to see the implausibility of the expectation (see Siegel 1979). New dance forms are bound to exploit the expressive styles of life that themselves emerge contingently and change without end. Georges Balanchine, who wrote the preface for the original edition of Hutchinson's Labanotation, had already tactfully observed there the inability of the extraordinarily detailed extensional analysis of dance movement to capture "the style of the finished product and . . . the general overall visual picture and staging." He admired the precision of the notation, particularly its fidelity with respect to "time values," but he also saw the advantage of a film record—which is to say, he saw the impossibility of a notation for the intentional, autographic features of the dance (Hutchinson 1977: xi–xii). The point is easily misunderstood. Of course, it is quite possible that a kind of dance notation may well facilitate a grasp of the intentional and expressive complexities of a particular dance. What is at stake is not notation but allographic notation, notation confined to purely extensional features of the dance (as of position and path of motion) that may be supposed to be theoretically adequate to identifying particular pieces, or identifying them in an aesthetically relevant way. It is instructive, for example, that the enlargement of Rudolph Laban's extensional dance notation into so-called effort/shape analysis and choreutics presupposes relatively fixed correspondences between psychological attitudes and dispositions and movement complexes.³ But the adequacy of such a system is certainly doubtful beyond the most elementary forms of natural expressiveness. For aesthetic purposes, it is not particularly important to confine notation allographically; in fact, in musical notation, we normally supplement putatively extensional notation with such notational instructions as andante, allegro, and the like, which are intended to capture a quality that belongs to a certain historical practice or spirit (or dance style) that normally cannot be fixed allographically.

     In any case, it is impossible to advance the allographic ideal without attention to the theory of how dance is produced and related to the expressiveness of human life itself. To consider these matters—in all the arts as well as the dance—is to grasp, I should insist, the sense in which dance is essentially autographic, with some notational conveniences. The allographic model is not mistaken because it is too inflexible or because the dance has not yet emerged at a stable enough level to be properly thus characterized. It is mistaken because it tends to separate questions of identity from questions of aesthetic interest and because it ignores the profoundly intentional nature of all art—not primarily in the sense of the psychologically deliberate but in a sense grounded at once in the biologically and culturally shaped forms of historical human expression. The very fact that some of the arts are recognizably autographic in nature demonstrates conclusively that the reidentification of particular works—to say nothing of their appreciation—does not require compliance with any allographic system. Reject the extraneous complication of the theory of forgery, and no barrier exists against admitting the autographic nature of the performing arts as well. To restore the proper sense of that characterization is to restore the sense in which the production, recognition, and appreciation of art depend upon the emergent and personally shifting powers of the individual human members of historically contingent cultures.