the tonal occurrence and not yet a real clarification of its meaning" (ibid., p.20). Schenker's impatience with Hoffmann is understandable, especially when we take into' account their divergent backgrounds. Schenker approached Beethoven's Symphony with the ear of an early 20th-century theorist; Hoff- mann, with the ear of an early 19th-century critic-composer. Our historical investigation of hearing on various hierarchical levels must include all avail- able evidence, whether or not it conforms to current standards. The task will be complicated indeed. We will have to take into account and assess opinions of composers, theorists, performers, critics, informed or uninformed amateur listeners, from their respective points of view. In applying their comments to the music, we will have to establish the hierarchical levels, which in turn vary for different periods. Whether or not our results will affirm "the in- creasing hierarchization of the musical space" remains to be seen. The attempted investigation must be fruitful, nevertheless, because we will be "educating our ears" in the process. Edward T. Cone, What is a composition? Not all arguments can be settled by agreement on the definitions of the terms involved, but the issues at stake can always be clarified thereby. Not all questions of definition can be settled by appeal to the dictionary, but, unless one is determined to play Humpty-Dumpty, that is the place to begin. When the concept under discussion is one that may well be affected by relevant historical considerations, as is certainly the case with "piece of music", one might well start by consulting the Oxford English dictionary. The entry under "piece" in the OED is, as one might imagine, lengthy. The references that concern us are to be found under two subheads: the first refers to the use of the word "in general sense; or followed by oj" and the second to "absolute uses ... without oj". In other words, we are asked to distinguish between "piece of music" and (musical) "piece". The former is listed under meanings relating to "a portion or quantity of any substance or kind of matter forming a single (usually small) body or mass". It is considered as "forming one body of finite dimensions" and as constituting "a separate part of the whole existing stock of the substance". In this sense, which is the earlier of the two to enter cornmon use, we find, for example, the locution "piece of song" (Twelfth Night, II, iv, 2), which, although obsolete today, shows clearly the way in which the term is conceived and how it differs from the more modern "absolute", usage. In the latter sense "piece" refers to "a production, specimen of handicraft, work or art", and is equated with "piece of work"; specifically, "a musical composition, usually short, either inde- pendent or for-ruing an individual part of a larger work". What interests me about the foregoing distinction is the fact while the 101 two senses of "piece" are closely related, the connection between them is not so simple and obvious as one would suppose. For although "piece of music" may be analogous to "piece of cake", "piece" tout court is more like "cookie". And the second usage seems not to derive directly from the first. It appears rather to arrive indirectly, by way of "piece of work" (cf. "opus"). The "piece" is thought of as the result of a limited of work, not as a de- limited portion of all music. Used in the second way, the word has been applied to plays, pictures, statues, and furniture. Its specific application to music comes on the scene comparatively late. The OED finds the word referring to a playas early as 1643, but the earliest citation for a musical composition is 1825. Grove, on the other hand, states, without evidence, that it has been in use "since the end of the 18th century". In any case, the "absolute" usage is more recent than the "general"; but this fact tells us less about music than about European thought, for the concept is recent with respect to all the arts, whether fine or applied. (According to the OED, the first instance of "piece of work", mean- ing "a product of work, a production", is 1540. This date coincides strikingly with the one cited by Miss Carpenter for the appearance of opus perfectum et absolutum for a musical work.) Let us now look a little more carefully at the definition: "usually short, either independent or forming an individual part of a larger work". (I shall return to the term "composition" a little later.) Grove says that the earliest use of the absolute term in English referred to the parts of a suite, but that the word is not generally applied to works in several movements, like a sonata, nor to the individual movements thereof-and further, that it is "not used of vocal music". Certainly all of these qualifications determine our ordinary use of the word in the absolute sense. Although we may occasionally refer to the "movements" of a suite, we usually feel that they have more independ- ence than those of a sonata-i.e., that they are pieces. We note Webern's care in distinguishing (Drei) Kleine StUcke from (Funf) Satze. We do not nor- mally refer to a symphony as a piece, nor as four pieces; but we do speak of the pieces that make up Le Tombeau de Couperin. We do not usually call songs, or cantatas, or operas, pieces. (Jocular use is always an exception: "The Ring is quite a piece !") But these distinctions are practical, not theoretical. If one should hear the Adagietto from Mahler's Fifth Symphony played alone on a program, one might well call it a piece. A Scriabin one-movement sonata might be considered a piece. Furthermore, the limitations of the word are not the same in all languages, not even all western-European ones. Beethoven referred to the first movement of Op. 27, No.2 as questa pezzo, and Verdi wrote four Pezzi Sacri for chorus. German dictionaries report that StUck (which dates, incidentally, from the 16th century) refers especially but not exclusively to instrumental music. And so.on. It now seems clear that both Miss Carpenter· and Professor Crocker have, to a certain extent, confused the two "pieces". For "piece of music" is as we see, by no means synonymous with (musical) "piece". The latter, once we 102 2£ WKtAJii-_ Z .{ J .,} have milked it history, retains little conceptual power. To argue over what is or is a piece is, today, an argument about usage: does one nor- mally refer to or that-a song, a symphony, an opera-as a piece? It is equally pointless{to argue over what constitutes a "piece of music"-for any ,piece of music,: be considered as constituting a "piece of music", or, as one migh,t better write it, "piece-of-music". (It is difficult to define the term ". without 'circularity, for what else is one to call what one is trying to define?) A piece; then, is what. one can usefully call a piece; a piece of music is what one chooses to regard as a piece of music . .. What we ought to be discussing-in fact, what Miss Carpenter and Pro- fessor Crocker are often, but not always, discussing-is, What is a composi- tion? Pieces, songs, symphonies, operas-all are compositions. Must a piece be written by one, and only one, composer, as Professor Crocker implies? Perhaps-but many compositions have been written by more than one. (We don't have to wait for the collaboration of Luening and Ussachevsky; we can point to the Handel-Brahms and Haydn-Brahms variations.) Does a Bach fugue, as Miss Carpenter seems to believe, represent the ideal piece? Never mind-a Willaert ricercar is certainly a composition. (Let that state- ment stand for now; we may wish to modify it later.) Must a piece be con- tinuous? No doubt, but the Ordinary of the Mass can constitute a composition in several parts, distributed over a wide span of time, interrupted by other music. Is a trope a piece? The query may be translated thus: "Is it valid to apply to a special medieval form of vocal interpolation a modern word nor- mally used to distinguish short, discrete, instrumental works?" When the is stated in this way, the answer is obvious; but what we should really be asking is, "Is introit-cum-trope a composition?" What is a composition? Returning to the OED, we find that the word "composition" can be applied to music in two ways. Referring to the pro- cess, to the action of writing music, the term has been current at least since 1597, when it was so used by Morley. Referring to the product of such an activity, its first citation is from 1666-67-in Pepys's diary, of all places. By way of definition this dictionary offers little help: "a musical production, a piece of music". Grove offers nothing. Is "composition" then synonymous with "piece of music", as the OED would have it? When the dictionary becomes vague, or at odds with ordinary usage, one must refer to one's own experience with words in order to answer such a question. Are the terms synonymous as we normally use them? A folk ballad is certainly "a musical production, a piece of music"; do we call it a composition? Hardly. Why not? Because we feel that a composition must indeed be composed-by a composer. The composer may be known or un- known, single or multiple, human or (Heaven help us) mechanical-but the composition must have a composer. If a piece of music is the result of accident or, like the ballad, of evolution, tradition, and constant variation, we do not normally apply to it the term "composition". The only way in which composition, in this sense, can be practiced is by 103 ) \ } \ some sort of recording. In the past this recording has been in the form of writing; now other methods are available to us. J'he real poi.rli, I think, is that a composition must be repeatable-and id(entifiably repeat- able. Without recording, no repetition can be identified/cas a form "0-_-,- original, since the original would exist only in the slipper\- and unreliable form of memory. The composition must endure in some kiiid of tr;lodel (a ,-, score, a master tape) against which every performance can be checked:->. ' Perhaps, then, one can simplify matters and equate the composition''Yith its model. If there is a score, the composition is the score, and every per- formance can be measured as a more or less accurate realization of its directions. (In the case of a mechanical recording as a model, any other per- formance would be measured as an imitation of the original.) But ifwe accept such an equation, then what of Stockhausen's Klavierstilck XI? There is something intuitively wrong in saying, of several pieces of music that sound completely different from one another, that they are all valid realizations of one and the same composition because they were all obtained by faithfully following the directions of the same score. Or, to make the issue clearer, here is the score of a new composition of my own: PLAY ANY PHONOGRAPH RECORD IN YOUR COLLECTION. Are all possible performances of this score realizations of one and the same composition? I think that we must conclude, first, that a score is not a com- position, but only a series of directions; and second, that a score can be said to embody a composition only if its directions, correctly followed, lead to results that can be aurally recognized as performances of the same piece of music. A further condition for a piece of music to achieve compositionhood has to do with its beginning and ending-extremes, as I call them. As Professor Crocker says, "All pieces have beginnings and endings, simply because every- thing does." But so does any fragment of music. Hence his demand "that the beginning sound like a beginning, and especially the end like an ending". By this, I take it, he means that the extremes must be musically determined and musically satisfYing as extremes. (To be sure, in vocal music the extremes are also partly determined by the words; but if they are totally determined by the words, the result can hardly be called a musical composition. The task of the song composer is to make the musical form sound natural in spite of the exigencies of the text. See, below, the discussion of what I call "essentially vocal music".) The last requirement is a preliminary way of saying that a composition must have form. This condition is necessary and might even, when properly interpreted, prove sufficient. True, at this point I may have overstepped the line that separates the definition of a class from the specification of a good example of that class. Yet I should find it as hard to define "musical com- positron" without some'reference to form as to define "poem" without some reference to rhythm. To call a piece of music a composition is the same as 104 calling it a work of art, as I use the term. When a work of art is fully formed, there is no doubt of its status: it is a work of art. But it is always hard to de- cide what to call less fully formed works. We may wish to call some of them imperfect; others, unsuccessful; still others, bad. And at some point we are bound to say, "This is not a work of art at all". So, perforce, the term "work of art" implies some normative judgment; so, too, I hold, must the term "composition". And indeed, if we take the trouble to follow the OED back from "composition" (product) through "composition" (activity) to its source in 'fcompose", we find the term, as applied to music, defined as "to invent and put into proper form"; from which we could certainly infer that to call a "musical product" a composition is to attribute to it at least a modicum of structure. It is of further comfort to refer to Webster (Third) and to find that, prop- erly glossed, its definition offers some support to my own. According to this authority, a composition is "a written piece of music; esp. an odginal work of some magnitude and to whose formal structure appropriate attention has been given". Let us ignore "magnitude"; most of us would reject the implica- tion that our word should be reserved for fairly long works and be denied to much of Webern. But the rest is apposite. "Written": this refers to my re- quirement that a composition be recorded in some way. "Original": pre- sumably this is intended, not as a stipulation of original quality, but as an indication that a work should not be secondhand-i.e., that it should be neither deliberately copied nor traditionally handed down. In other words, it must be "composed by a composer". "Formal structure": this restriction can be interpreted as applying to my own demands for recognizable repeatability as well as for musically determined form. There is, then, some justification for the position that to call a piece of music a composition is to suggest that it can be considered a work of art. But now a new question arises. Why have we, in the case of music, only the term "composition", which has, etymologically speaking, no specific relation to music, and has, practically speaking, associations with most of the other fine arts? There is no common noun-a music-comparable to a picture, or a building, or.a statue. (Even if one protests that the last does not apply to re- liefs, one must admit the usage "a sculpture", which does.) One answer, if not the only one, is suggested by a comparison with another art suffering from a similar disability. What term have we-save "composi- tion" again-to cover all products of imaginative literature? Poem, novel, short story, drama-there is no single word that covers all these and only these. But in the days when literature was primarily poetry, before the de- velopment of prose fiction and prose drama, no such word was needed; "poem" was always at hand (in whatever language one was using) and covered almost all cases. Every imaginative writer could accordingly be called a poet. But when a new word was needed to cover the wider range of possibilities, none was readily available-except the generalized "composi- tion". 105 In the same way, when music was thought of as primarily, if not exclusively, vocal, no term was needed other than "song". I believe it is no accident that the adoption of the generalized word occurs simultaneously with the appear- ance of independent instrumental music. But there is a much more profound lesson to be read here. So long as music was essentially vocal (and, when instrumental, tied to dance or ritual), there was no occasion for what Miss Carpenter calls "the self-conscious notion of the musical work and the culti- vation of autonomous musical form". By essentially vocal music-which I take almost all pre-Renaissance music to be-I mean essentially textual music: music whose form is determined primarily by textual (or ritual) exigency. (The key word in this formulation is, of course, "primarily". A vocal style founded completely on textual requirements, as in certain forms of liturgical chant, would have no musical interest whatsoever. At the other extreme, much l5th- and 16th-century polyphony, while primarily textual, achieves forms that are either satisfactory from a purely musical standpoint or at least independent enough to be listened to as musical compositions.) Vocal music written since the rise of a specifically instrumental style is composed on quite different principles. (Indeed, perhaps one need say only that it is composed.) Except in the case of the pure recitative, its form is primarily musically determined. Even when the demands of the text seem paramount, those of functional harmony and tonal structure are more insistent. (The rare excep- tions, of which Satie's Socrate may be one, only point up, by contrast, the force of the general rule.) In other words, the concept of a musical composition is, as Miss Carpenter says, a late arrival on the Western musical scene. But this arrival need not necessarily represent a change from process to object, or from becoming to being, or from participation to observation. The comment attributed to Johann Gottfried Walther, that music "has gone through a change in its usage from adjective to substantive", can be interpreted in a simpler and more objective way. A substantive can stand alone; an adjective must modi- fy a substantive. Modern music may thus indeed be a kind of "object"; medieval music embellished, varied, expanded, or otherwise modified an- other object: the text. Or perhaps one should say that music, text, and ritual together constituted the object. The loose forms of the ricercar and canzona are primarily due, not to an ideal of syncretic form, but_precisely to the fact that there was, in the 16th century, as yet no ideal of instrumental form. As has often been pointed out, the only models the composers had were vocal. Imagine these compositions with suitable words and you will realize that they embody what I call essen- tially vocal forms, adapted for instruments. Lacking the structure of the word, they seem to us imperfectly realized as compositions-though com- positionhood is certainly the condition toward which they aspire. The late Professor A. M. Friend (of the Princeton University Department of Art and Archaeology) used to insist that what I have been arguing for music was true of all the arts. The Middle Ages produced not works of art but 106 a synthesis of the arts, most completely embodied in the cathedral. The cathedral was architecture, it is true, but it also embraced sculpture (in relief and the round) and pictorial art (as mural and as stained glass). Its liturgy comprised drama, literature, and music. Already with the 14th century, each of these arts was beginning to realize itself individually, and the process was completed by the Renaissance. The communally built cathedral yielded to the individual art, the individual artist, and the indi- vidual work of art. We may now be entering a definitive post-Renaissance stage of Western culture, but I find it misleading to look on what is happening as in any sense a return to older and perhaps more natural modes of perception. Rather, we are confronted by an attack on the whole concept of art. If the attackers win, not only the work of art as we know it but art itself may disappear. Some composers-I use the term only because I do not know what else to call them, except perhaps noncomposers-are loudly proclaiming the Death of Music in a manner that recalls certain stylish theological positions, and they are en- couraging their followers to complete its doing-in. Others, more reticent, are nevertheless apparently trying to hasten the process by insisting that whatever one wants to call music is music, that what one calls a composition is a com- position. John Cage's position is more honest. A few years ago, in conversa- tion, he said, "I don't claim that what I am doing is music, or art-or that it has any value. I maintain only that it is an activity, and that it is the one in which I happen to be engaged at present." Such a position is, from a purely personal point of view, unassailable; but if generally accepted by those who call themselves musicians, it means the end of music. Let us not deceive ourselves. The extreme avant-garde is not trying to offer new definitions of what constitutes a work of art, or to create new forms, or to encourage new modes of perception. The extreme avant-garde has only one attitude towards the arts: it wants to kill them. Bernard Stambler, Debts and transfigurations The materials of Miss Carpenter's article suggest certain quodlibets on the terms and structures of our responses to the arts. Miss Carpenter is quite adept with the two large structures, or frames of reference, she is manipulating and combining for the purposes of her paper: the sense (or concretizing) of musical form that appeared in the 18th century, and certain analytic methods of experimental psychology, essentially of the 20th century. Somewhat less clear and explicit in her article (as, generally, in "histories" of music) is how we are to see the relation between 18th-century and pre-18th-century forms. The 18th century developed the "great formal types of musical organiza- tion" that by and large we still live by; it might be more accurate to say, the 18th century provided the ears with which we still hear music. The forms 107