Cage = 100: Tudor and the Performance Practice of Concert for Piano and Orchestra

Cage = 100: Tudor and the Performance Practice of Concert for Piano and Orchestra

Many of John Cage’s scores seem to allow performers a degree of freedom that often leads to interpretations that, by the composer’s own admission, do not reflect the spirit of the work. This is a problem of both attitude and notation. If we are to continue or reconstruct the tradition, we must look to the one performer in particular who defined and was defined by the performance practice of Cage’s music – the pianist, composer, and electronic musician David Tudor.

Written By

Isaac Schankler

Cage in 1992

John Cage in August 1992, the last month of his life. Photo by John Maggiotto, courtesy S.E.M. Ensemble

In 1958, the premiere of John Cage’s Concert for Piano and Orchestra was marred by disruptive behavior from both audience members and musicians. By Cage’s own account, “some of [the musicians]—not all—introduced in the actual performance sounds of a nature not found in my notations, characterized for the most part by their intentions which had become foolish and unprofessional.”[1] These intrusions included, among other things, exaggerated corny blues riffs, prolonged and sarcastic applause, and a tuba ostinato from Stravinsky’s Le Sacre du Printemps.[2] This was certainly not the only incident of this kind in Cage’s life. In a 1975 performance of Song Books, soloist Julius Eastman proceeded to slowly undress his boyfriend onstage, and then to attempt to do the same to his own sister, who stopped him by protesting, “No Julius, no!” The next day, this prompted Cage to pronounce in a rare moment of anger, “I’m tired of people who think that they could do whatever they want with my music!”[3]

Scenarios like these bring to light a particular recurring problem in the interpretation of Cage’s music. Many of Cage’s scores seem to allow performers a degree of freedom that often leads to interpretations that, by the composer’s own admission, do not reflect the spirit of the work. This is a problem of both attitude and notation. In the first example, nothing in the notation of the Concert would allow for the Stravinsky tuba excerpt, so the event could be simply explained as the “unprofessional” actions of a disgruntled performer. However, in the second example, Eastman’s interpretation could in fact be, technically speaking, a valid reading of the score. Song Books consists of eighty-five solo pieces with various combinations of song, theatre and electronic accompaniment. Some of the theatrical solos ask the performers to make a list of verbs and nouns and perform actions based on those choices and indications in the score. If Eastman chose the verb “undress” and nouns “boyfriend” and “sister,” then his actions could have been perfectly within the bounds of Cage’s written notation. But both Cage and Petr Kotik, who directed the performance, denounced Eastman’s performance, Kotik declaring it a deliberate, malicious act of sabotage.

This incident had a profound effect on Kotik, who came to believe that Cage’s music demands a particular kind of performance practice that is not contained in the notation, and that in this respect, Cage’s music is “not much different from a Mozart score.”[4] The problem is that, as we have seen, even during Cage’s lifetime people had substantial difficulties with the performance practice of Cage’s music. Now, when there are few left who had direct contact with Cage, and fewer still who have lectured or written about it, much of this performance practice is in danger of being lost entirely. If we are to continue or reconstruct the tradition, we must look to the one performer in particular who defined and was defined by the performance practice of Cage’s music – the pianist, composer, and electronic musician David Tudor.

Tudor played a crucial role in the development of Cage’s music in the 1950s. Over and over, Cage acknowledges his debt to Tudor, as in this representative statement from 1970:

In all my works since 1952, I have tried to achieve what would seem interesting and vibrant to David Tudor. Whatever succeeds in the works I have done has been determined in relationship to him… Tudor was present in everything I was doing.[5]

There is a tendency to view the relationship between composer and performer as one way only, as a vector through which the composer’s intentions are transmitted to the performer, but Cage describes his relationship with Tudor as one through which the composer’s direction and determination was in fact defined by the performer’s interests and aspirations. If we are to understand Cage’s music from this period, Tudor is the key.

Because so much of his music in the 1950s was written specifically for Tudor, Cage’s notation is frequently opaque or hard to understand. Tudor’s rapport with Cage, not to mention Tudor’s serious devotion and meticulous attention to detail, may have made Cage less aware of the potential pitfalls involved in performance. Certainly, this was an issue with the premiere of Concert for Piano and Orchestra. As in most of Cage’s work from the 1950s, Concert employs indeterminate notation that gives the performers a certain degree of freedom, but Cage was not interested in just any kind of freedom: “I must find a way to let people be free without their becoming foolish. So that their freedom will make them noble. How will I do this?”[6] This is also one of the central questions that every interpreter, listener, or scholar of Cage’s music must eventually come up against: “How will my freedom make me noble?”

In many ways, Tudor was the embodiment of this “noble” freedom (as distinct from other, “foolish” kinds of freedom). Not surprisingly, he discovered how to be “free” while working on one of Cage’s first scores for Tudor, Music of Changes, written in 1951. (While Cage used chance procedures in composing Music of Changes, it predates his experiments with indeterminate notation.) Here is Tudor’s own description of the discovery:

Music of Changes was a great discipline, because you can’t do it unless you’re ready for anything at each instant. You can’t carry over any emotional impediments, though at the same time you have to be ready to accept them each instant, as they arise. Being an instrumentalist carries with it the job of making physical preparations for the next instant, so I had to learn to put myself in the right frame of mind. I had to learn how to be able to cancel my consciousness of any previous moment, in order to be able to produce the next one. What this did for me was to bring about freedom, the freedom to do anything, and that’s how I learned to be free for a whole hour at a time.[7]

Tudor’s freedom actually arose from an unprecedented array of constraints, and the physical and mental discipline needed to obey those constraints. Tellingly, when asked about the proper interpretation of Cage’s music, Petr Kotik also refers to David Tudor and “discipline”:

The most important thing to understand about Cage’s music is the discipline required which is the exact opposite of the popular perception about chance music. It is the discipline, the exactness, the precision, the focus, the concentration, all of which Cage takes for granted when he writes his music. Everything that Cage wrote from the early 1950s until the early 1970s was written for or with David Tudor in mind. This is why these pieces are so difficult. Without a “Cageian discipline” the application of chance turns the music into nonsense.[8]

In many ways, the Concert for Piano and Orchestra provides an ideal framework to explore what “Cageian discipline” might mean in terms of Tudor’s performance practice. As we have already seen, it is problematic, uncovering some of the more unruly aesthetic issues in Cage’s work. It is imposing, as few performers have even attempted to take on the mammoth piano part. It is exemplary, a colossal compilation of Cage’s various notational gambits. It is transitional, paving the way for Tudor’s forays into electronic music. And it is perennial, as Tudor continued to revisit the piece even after he had largely left the piano behind, recording it a total of four times between 1958 and 1992, a time span that covers most of Tudor’s career.

The Concert has 14 instrumental parts and no overall score. While there is also a “part for conductor” that can in theory be used to alter the piece’s timing, Kotik asserts that Cage never used this part, at least not since 1964 when Cage and Kotik first met.[9] The instrumental parts can be used in any combination, with according changes in title, from the full complement (Concert for Piano and Orchestra) to smaller ensembles (e.g. Concert for Piano, 2 Violins and Bassoon) to single instruments (e.g. Solo for Cello, Solo for Sliding Trombone, etc.). Of these parts, the Solo for Piano is by far the largest and most difficult.[10]

The Solo for Piano was in many ways the culmination of Cage’s experiments with indeterminate notation. A kaleidoscopic compendium of graphic notational systems, it asks the pianist to compile a performance using selections from 84 different kinds of notation spread across 63 pages. Cage refers to these notations using letters A through CF in the key at the beginning of the score, which also gives instructions on how to interpret each kind of notation. This facilitates both identifying and executing the notation, since there are multiple instances of some kinds of notation. Tudor referred to these instances as graphs, presumably because each instance “constitutes a discrete graphic object,”[11] but also for the mathematical implications of the term, as we will see.

Rather than reading directly from the score, Tudor wrote out his own realizations to read from in performance. This is by far a more practical approach, since reading from the score would require internalizing all the different types of notation and being able to execute them instantaneously, a next to impossible task. But making realizations was also Tudor’s standard practice for indeterminate music. Christian Wolff, another composer fond of graphic notation, attempted to curb this tendency of Tudor’s by writing scores for Tudor that required spontaneous action in performance—but in the end Tudor simply wrote out all the possible choices.[12] For Tudor it was the only way: “Nothing else could work. When you’re looking at graphic notation, how are you going to do it? Either you make the realizations, the way I did, or you decide that whatever happens at the moment is the music. And that’s the way many people are looking at those graphic scores right now.”[13] Tudor is almost blasé about the idea of a spur-of-the-moment performance, but Kotik suggests that Cage might not have been happy with such a result:

[Cage] had absolute faith in what Tudor would make of it, and Tudor always made it what it ought to be… Their connection was perfect and Cage deliberately left some things open. But of course this presents us with a problem today… One of the basic foundations of Cage’s thought was the rejection of value judgments. He completely refused to judge things, and was utterly consistent about it. So when someone “messed up” his music in some ghastly way he wouldn’t stand up and start shouting “How dare you?” but would just sit there saying nothing, and then leave. The problem is that this attitude has often been regarded as agreement. It got to such a point that there are musicians Cage simply couldn’t stand who still think he was terribly fond of them.[14]

Initially, Tudor’s meticulous, ordered approach may seem like the antithesis of Cage’s carnival of possibility, but Tudor’s fastidious tendencies actually allowed him a great deal of flexibility. Each one of the four widely available recordings of Concert for Piano and Orchestra/Solo for Piano is quite distinct in both content and character. To fully understand how Tudor achieved this flexibility, one must look at two steps of the process: 1) the conversion from Cage’s score to Tudor’s written realization (interpretation), and 2) the conversion of the written realization into sound (performance).

Tudor eventually made two different written realizations of the Concert for Piano and Orchestra/Solo for Piano. The first realization was used for the 1958 Town Hall premiere, a few subsequent performances, and to accompany Merce Cunningham’s dance Antic Meet. To make this realization, first Tudor made a list of all occurrences of each kind of graph and their corresponding page numbers. Next he made a selection of which particular graphs to use. The justification for this is in Cage’s key: “The whole is to be taken as a body of material presentable at any point between minimum (nothing played) and maximum (everything played), both horizontally and vertically.” In other words, Tudor could use as many or as few graphs as he wished. John Holzaepfel suggests that Tudor made his selections so that he would have at least one of each “graph type,” since some kinds of notation were closely related to others.[15] (For instance, the instructions for AB specify “clusters as in Z,” and instructions for Z specify “dynamics as in T,” making T, Z and AB one “graph type” in Tudor’s estimation.) After a preliminary sketch, Tudor made a performance plan of which graphs to use in rehearsal and performance for the Town Hall concert (Figure 1).

Figure 1

Figure 1. Tudor, first realization of Concert for Piano and Orchestra: performance plans for May 15, 1958 rehearsal and performance. Image reproduced courtesy of the Getty Research Institute.

These performance plans also included predetermined gaps where Tudor would play silence. Approximate durations of graph readings are given in increments of 5 seconds (the longest is 45 seconds, while many are as short as 10 seconds). The fact that Tudor made separate plans for rehearsal and performance is telling. Cage himself preferred not to have any rehearsals whatsoever, since they might cause performers to interact with one another in intentional ways. But when rehearsals were obligatory, Tudor’s differing performance plans were one method of disrupting this tendency.

For the realization itself, Tudor transcribed his readings of graphs onto separate small loose-leaf sheets of manuscript paper, which were then compiled in a ring binder notebook. This allowed Tudor to “vary both the internal order and the overall duration of his subsequent performances of the realization simply by adding, removing, and rearranging the pages in the notebook.”[16] Thus Tudor was able to be both very specific and flexible regarding timings and length of the piece. However, because each graph reading was bound to a specific length of time, in effect this put a cap on the maximum duration of the piece (without adding additional graph readings).

This became an issue when Cage asked Tudor to provide musical accompaniment derived from Concert/Solo for Cage’s Indeterminacy lectures, a series of 90 stories each one-minute long. The flexibility afforded by Concert/Solo was ideal for such a task, but Tudor did not have enough time to transcribe the many, many more graph readings this would have required (the longest previous performance of Concert was scarcely more than thirty minutes). This called for a new realization, and a new approach.

In creating this realization, Tudor was able to draw on his unique understanding of musical time, which he first developed while working on Pierre Boulez’s Second Sonata and Cage’s Music of Changes. Boulez’s Second Sonata employs conflicting rhythms to frustrate a sense of meter, and Tudor initially struggled with the performance of Boulez’s work:

I recall how my mind had to change… I realized that I could play everything, but I had to stop every two measures. I couldn’t put it together. And I wondered, What is wrong? Why not? … I saw that there was a different way of looking at musical continuity, having to deal with what [Antonin] Artaud called the affective athleticism. It has to do with the disciplines that an actor goes through. So all of a sudden I found I could play a movement through. It was a real breakthrough for me, because my musical consciousness in the meantime changed completely… I put my mind in a state of non-continuity, not remembering what had passed, so that each moment is alive.[17]

This sense of “non-continuity” is crucial not just to Boulez and Cage, but to all the composers who worked most closely with Tudor, including Karlheinz Stockhausen, Morton Feldman, Christian Wolff, Earle Brown, Henri Posseur, and others. While the Second Sonata predates Boulez’s work with electronics, Cage and others were at that time greatly influenced by the advent of magnetic tape music, which engendered a new and different understanding of time. Specifically, it caused composers to move away from the idea of rhythms that could be counted and move into what Boulez called “amorphous time” (as opposed to “pulsed time”) and what Cage called “time itself”:

Counting is no longer necessary for magnetic tape music (where so many inches or centimeters equal so many seconds): magnetic tape music makes it clear that we are in time itself, not in measures of two, three, or four or any other number.[18]

Composers tried various methods to make this new conception of time comprehensible to human performers. For example, Stockhausen developed a theory of “time-fields” to examine the psychology of time, while Feldman employed carefully notated shifting meters to create the feeling of “durational” (as opposed to “rhythmic”) time. But it was Tudor who first learned how to perform without counting beats, and showed that this time-conception was humanly executable, in his work with the Second Sonata and Music of Changes.

Cage wrote Music of Changes for Tudor after hearing his performance of Boulez’s Second Sonata, and the two works share a sense of temporal “non-continuity.” In some ways, Music of Changes takes this idea further, by using a prototypical version of proportional notation, here described by Eric Smigel: “A conspicuous notational feature of Music of Changes is the presence of evenly-spaced barlines in a non-metric context… the barlines simply articulate exact intervals of time-space, irrespective of the musical content of each measure.”[19] While there are tempo changes that disrupt this proportion, according to Cage, they apply to the “rhythmic structure, rather than with the sounds that happen in it.”[20]

While proportional notation is familiar, even ordinary, to any student or performer of contemporary music today, it is worth exploring why the idea was so radical at the time of its inception. Composers of the older generation, in particular, seemed to be bothered by it. Tudor describes the reaction of Stefan Wolpe, one of Tudor’s teachers, after studying a score of Music of Changes:

[Wolpe] met Cage at a party and he told him, “I love your music, but you’re a liar!” … What he meant to say was that he couldn’t feel it. But I could… I was watching time rather than experiencing it. That difference is basic. Even playing pieces which last an indefinite length of time your relationship to time is different, because you are now able to telescope some periods and to microscope others at will.[21]

This ability to “telescope” and “microscope” time, first developed while working on Music of Changes, became the key to the second realization of Solo for Piano:

I had already prepared a great deal of material from the Concerto [sic] for Piano and Orchestra but for John’s lecture he wanted quite a length of time, so taking the notion that the time of the performance had to be adjustable, I then looked over the material that I had and I even made more. The method was that I looked over all the graphs from the Concerto [sic] which would only produce single ictii [sic] (accents)… Then I looked at all the graphs containing single points or which would produce single ictii [sic] and I expanded each graph to the same proportion. I made a notation of this proportion like a book… With that in mind, I could play the whole thing in fifteen minutes if I were a genius or thirty minutes, or forty-five minutes, or an hour. Eventually we performed it for three hours and there was always plenty of sound material.[22]

That is, Tudor made a new selection of graph readings, with the criterion that they must be graphs that could produce single attack points, avoiding graphs that produce sequences of notes or other linear implications. This would allow Tudor to expand or contract the space between attack points, granting even greater flexibility of duration. Thus Tudor was able to play the entirety of Solo for Piano (i.e. every page of his realization) over various lengths of time. Though both realizations employ proportional notation, the complex, often dense graph readings employed in the first realization (what Tudor called “cursive figurations”) would be very difficult to expand or contract temporally. The single ictuses in the second realization present no such problem.

But how did Tudor choose where to place the attack points? Tudor hints at this when he mentions expanding “each graph to the same proportion.” Holzaepfel explains:

To determine the attack points of his readings of Cage’s graphs within the 90-minute time frame of his realization, Tudor measured the area or length of each graph, using whatever means of measurement he found appropriate to a graph’s individual form. Usually a decimal ruler, or sometimes a circular slide rule, would suffice… This gave him an area or length A for each graph. Next, Tudor measured the position of each ictus within the graph, usually in terms of its distance from the beginning of the graph. He then multiplied each position measurement by the total duration of his realization (5400 seconds) and divided the result by the A number. The quotient was the ap [attack point], in Tudor’s realization, of the ictus in Cage’s score. In other words, what was constant to each graph was not a multiplier but a divider which was, in fact, the area or length, depending on a particular morphology, of the graph itself… in this way, Tudor, devised the internal temporal structure of his new realization in terms of both specific attack points and order of occurrence of the source material from Cage’s score.[23]

The linear, sequential format of graph readings in the first realization gave way to a format where graph readings were superimposed. Essentially, all graphs were performed simultaneously. To use Tudor’s analogy of expansion, it is as if each graph was magnified (or shrunk) until all the graphs were all the same size, and then laid on top of each other. (Interestingly, this is at around the same time that Cage started to experiment with using transparencies in his scores, as in Fontana Mix, Cartridge Music, and others.) Tudor saw Cage’s notations literally as graphs that could be measured and plotted in space. This equivalency of the spatial and temporal dimensions is consistent with Tudor’s approach to performance as “watching” (not “experiencing”) time. In fact, in his notes for a lecture given at Darmstadt, Tudor describes the basic formula for interpreting graphic notation as “starting out from space = time.”[24]

Tudor’s next step was to create a Master Table listing the locations of all attack points in the new realization (the first page of which can be seen in Figure 2). The first column of the table shows the location of the attack point (in seconds), the second column identifies the kind of graph that the attack point came from and what part of the graph (e.g. T-1 refers to the first attack point found in a graph labeled T), and the third column gives the page number where the graph is found in Cage’s score. Tudor was then ready to transcribe his readings from content sketches into the realization.

Figure 2

Figure 2. First page of Tudor’s Master Table for second realization of Concert for Piano and Orchestra/Solo for Piano. Image reproduced courtesy of the Getty Research Institute.

From the 787 total attack points, Tudor actually made two versions of this realization, one containing 472 attack points and the other using the remaining 315 attack points. Version 1 is on loose sheets of manuscript paper, while Version 2, like Tudor’s first realization, uses a ring binder. (Version 1 was used for the initial performance of the Indeterminacy collaboration at Columbia Teachers College, but Version 2 was used for the recording and all other subsequent recordings. Therefore, whenever I refer to “the second realization” I am generally referring to Version 2 of the second realization, unless otherwise specified.) But unlike the first realization, this time Tudor used blank paper instead of manuscript paper. Holzaepfel offers an explanation why:

The contents of both versions of the second realization, consisting as they do entirely of discrete events, no longer needed a continuous staff of lines and spaces but only a means of denoting the time scale… If the notation of a reading was in graphic or verbal form, Tudor could also dispense with the lines and spaces.[25]

Using blank paper had the added advantage of making the realization easier to read, since the notation “pops” more against the white space surrounding it.

Tudor’s second realization takes up 90 pages, making it ideal for accompanying Cage’s lecture at a rate of 1 page per minute. There are also several sets of numbers written on the second realization that suggest he worked out many possible timings for other performances. Each set of numbers is consistently placed on its own area of the page (e.g. upper-right corner, end of every third system, etc.), and each set generates a different total time for the piece. The quickest of these adds up to 22’30”, and requires the performer to play two pages every thirty seconds. Since the preparation of some attack points may take several seconds (allowing time for picking up or putting down beaters, preparing harmonics in advance, etc.), it is easy to see why Tudor would balk at performing the whole thing in fifteen minutes.

As a result, Tudor had to look for other ways to condense the piece without dramatically increasing the density or difficulty of the material. Fortunately, the ring-bound format of Tudor’s realizations made it extremely easy to simply omit pages: “You simply turn the pages and … select what material you want.”[26] Another set of numbers in the realization indicate one possible 30-minute version, which includes a selection of 30 pages played at a rate of one minute each. Tudor gives the timings at the right edge of each of those pages, sometimes with arrows leading to the next page number when the next page is not clear (presumably, where there are not arrows Tudor simply removed the intervening pages from the booklet). In this way, with these various sets of timings Tudor gave himself several possible courses of action for executing the realization (see Figure 3 for a table of these timings). None of the available recordings, however, use any of these timings.

Currently, there are four widely available recordings of Tudor performing the Concert for Piano and Orchestra or Solo for Piano. The earliest is a recording of the infamous 1958 Town Hall premiere, and can be found on The 25-Year Retrospective Concert of the Music of John Cage, a three-CD set released by Wergo. The next is a recording of the 1959 Tudor/Cage collaboration, Indeterminacy, on the Smithsonian Folkways label. The third is a 1982 recording of Solo for Piano made in Amsterdam and released in 1993 on David Tudor Plays Cage and Tudor by the Atonal label (out of print but still obtainable). The last is a 1992 recording of Concert for Piano and Orchestra with conductor Ingo Metzmacher and Ensemble Modern which appears on The Piano Concertos from the Mode label. Each recording is remarkably distinct from the others, and when taken together, they trace an evolutionary trajectory in Tudor’s performance practice, so it is worth describing the general features of each.

Figure 3

Tabulation of page numbers (left column) and various timings written in Tudor’s second realization of Concert for Piano and Orchestra/Solo for Piano.

In the recording of the Town Hall premiere, the “foolish behavior” of the musicians and scattered laughter and applause from audience members is clearly audible. The overall impression is quite busy and raucous. For Tudor’s part, there are periods of frantic, virtuosic activity interspersed with gulfs of silence. The result is less than ideal. Tudor’s playing, while infallible from a technical standpoint, presents a dialectic between silence and sound that seems jerky and forced. It is easy to see why listeners and performers could have mistaken it for comedy. This is the only recording of the four that uses the first realization; after he made the second realization, he seemed to vastly prefer it.

The version heard on Indeterminacy uses the 90-minute version of the second realization to accompany Cage’s lectures. Possibly because Tudor was worried his new accompaniment would be too sparse, he supplemented his performance with tracks from Cage’s tape piece Fontana Mix. These are triggered on and off at specific points corresponding to Tudor’s readings of graph BY, for which Cage’s instructions read: “Any noises, their relative pitch given graphically (up = high, down = low).” Obviously, Tudor took “any noises” to include electronic noises. However, Tudor seemed to pay no attention to the specified relative pitch, and he departs from the idea of each reading as a single ictus. In the realization, Tudor penciled in “on” and “off” beneath or above various instances of BY. In other words, each reading of BY would either activate or deactivate an electronic sound source. As a result, unlike most attack points in Tudor’s realization, these sounds could continue for quite some time. The end result on the recording is almost as busy as the Town Hall premiere, but in the context of Cage’s engaging and often witty lectures, the accompaniment seems more appropriate, with many extraordinary coincidences of word and sound (such as the fortissimo chord cluster that follows Cage’s utterance of “My problems have become social, rather than musical”).[27]

Holzaepfel also points out that on the Indeterminacy recording, when readings of multiple graphs coincided, Tudor modified his performance practice out of necessity:

Simultaneous occurrences of graph readings were interpreted with considerable flexibility, even freedom… Tudor sometimes spreads the contents of coincident readings over one or two full seconds. In fact, at times the effect is not that of a discrete sonority but something very like a phrase.[28]

Other than this, however, on this recording Tudor follows closely the timings laid out by the realization.

In some ways the 1982 recording of Solo for Piano hews most closely to Tudor’s second realization as written, with no electronic or verbal accompaniment. It is not difficult to follow along with the recording as if one were reading a conventional score. However, Tudor occasionally omits graph readings in his performance. There seems to be no systematization to the readings that he omits, though he tends to ignore graphs like BY that have vaguely specified pitch. This may have been due to Tudor’s reluctance to do anything unprepared, anything improvisatory that might lead to an undesired sense of intention. He certainly showed no reluctance to perform unpitched events if they were clearly specified in the realization (e.g. percussive effects on the body of the piano).

Furthermore, on this recording it seems as though Tudor allowed himself to be more liberal with time. The realization proceeds at a much faster rate than in Indeterminacy, with Tudor playing all 90 pages in less than 40 minutes. He also does not appear to be consistent about time corresponding to spatial proportion. The time scale is initially obscured on the recording by the first sound event taking place immediately (in the realization it comes after a system and a half of silence), but regardless, it is clear that here Tudor is no longer strictly following the “space = time” principle. Sound events that are spaced far apart on the page are sometimes performed in quick succession, and some that are spaced close together are often separated by large gulfs of silence (or other forms of sonic space, like long uninterrupted reverberations). If he is expanding and contracting time in a systematic way, it is not clearly audible, and I have found nothing in his notes to suggest such a systemization.

In the 1992 recording of Concert for Piano and Orchestra, Tudor seems to allow himself even more artistic license. He omits more graph readings, and often omits parts of graph readings (e.g. he may only play the top three notes of a five-note chord, or the outer tones of a cluster chord). This is well within the bounds of Cage’s notation (recall his instruction that any selection of material can be made “horizontally or vertically”). Tudor also rolls many chords rather than playing them as single attack points. Additionally, this is the only recording of the second realization that does not use the whole 90 pages. Tudor begins at page 1, but skips to page 18 a minute and a half later. From that point on, he reads pages sequentially until the end of page 61, when the piece ends. The duration of performance is almost exactly 30 minutes, suggesting that the time was predetermined but not the number of pages, and Tudor simply stopped when the conductor signaled the end of the piece.

Other changes made by Tudor on this recording are more baffling. At a few points, Tudor plays attack points in a different order than they appear on the page. Even stranger, occasionally Tudor plays a sonority that does not seem to appear anywhere in the written realization. The realization contains many additional graph readings in pencil, presumably tacked on throughout the years (most carried over from Version 1 of the realization), and the thinness of the paper also sometimes makes it possible to see notations on the opposite side of the page, or even the next page. It is possible that Tudor was deliberately reading through the page to create new sonorities. It is also possible that Tudor mistook some sonorities on the other side of the page for penciled additions, since both are faintly visible. (Austin Clarkson reports that in 1993, Tudor’s eyesight was beginning to fail.)[29] It is even possible that Tudor is operating on some plane of thought that transcends my understanding. At any rate, whatever the cause, it adds a new level of indeterminacy to the proceedings!

Tracing a path through Tudor’s recordings of Concert for Piano and Orchestra/Solo for Piano, Tudor appears to allow himself more freedom as the years progress, particularly with respect to the dimension of time. In the later recordings, Tudor no longer seems bound to the visual or spatial dimension as a guiding temporal principle. It is my belief that this is a result of his experiences with making electronic music. To illustrate this, I will closely examine a few graphs and how their interpretation evolved over time.[30]

T is one of the graphs that appears in both of Tudor’s realizations, and one where it is relatively easy to see the relationship between Cage’s notation and Tudor’s interpretation. Cage’s instructions for T read: “Influence in pitch and time notated as shapes with center points, to be audible as clusters, a single one changing in its course. Numbers refer to loudness (1-64) (soft to loud or loud to soft).” Figure 4 shows an instance of T on page 12 of Cage’s score, and Figure 5 shows its interpretation in Tudor’s first realization. Tudor’s interpretation is fairly literal here, except that he uses Cage’s “center points” to put the clusters in order from left to right, instead of using the leftmost points of the shapes. Tudor also chooses to move from one end of a shape to another and re-orient that into a left-to-right trajectory. For example, the rightmost shape curves around to the left at the top, but Tudor chooses to transcribe it as if it were continuing to the right, i.e. moving forward in time. Tudor also translates Cage’s numbers into his own dynamic scale, which runs from 0.0 to 10.5.

Figure 4

Figure 4. Graph T from page 12 of Cage’s Solo for Piano. John Cage, Concert for Piano and Orchestra © 1960 by Henmar Press / Edition Peters. All rights reserved. Reprinted with permission from C.F. Peters Corp.

Figure 5

Figure 5. Interpretation of graph 12 T from Tudor’s first realization of Concert for Piano and Orchestra/Solo for Piano. Tudor converts the 10 shapes in Cage’s score into specific clusters, here spread across two systems. Image reproduced courtesy of the Getty Research Institute.

This graph (12 T in Tudor’s notation) does not, strictly speaking, consist of single attack points, but Cage’s “center points” make it possible for Tudor to measure and plot them as if they were. Initially, 12 T appeared in Version 1 of Tudor’s second realization, but not Version 2, and so it does not appear on the Indeterminacy recording. However, Tudor at some point must have decided he liked the results of this graph, and made it one of the penciled additions to Version 2. Figure 6 shows the eighth shape from 12 T as it appears in Version 2 of the second realization. It can be heard on both the 1982 Solo and 1992 Concert recordings, but Tudor interprets it differently each time. On the 1982 recording, he plays the full figure at its written dynamic level (quite loud in Tudor’s scale). On the 1992 recording, it is the last sound heard, and Tudor only plays the very end of the gesture, a short rolled cluster ending on A-sharp.[31] Tudor also ignores the dynamic marking, and the gesture as it is performed is quite hushed, somewhere between piano and mezzo-piano.

Figure 6

Figure 6. Page 61 from Tudor’s second realization of Concert for Piano and Orchestra/Solo for Piano. A later addition at bottom right is penciled in, corresponding to the eighth shape from 12 T. Read-through from the next two pages is also faintly visible. Image reproduced courtesy of the Getty Research Institute.

BT is another graph used in both realizations, and it is one of the most unusual notations in Cage’s score. His instructions state that “notes give place of performance with respect to the piano,” but the drawing shows the outline of two grand pianos and a collection of points that, for the most part, do not intersect either piano (Figure 7).

Figure 7

Figure 7. Graph BT from page 54 of Cage’s Solo for Piano. John Cage, Concert for Piano and Orchestra © 1960 by Henmar Press / Edition Peters. All rights reserved. Reprinted with permission from C.F. Peters Corp.

Tudor chose to interpret those points which intersect the curve of the first piano as effects on the strings or body of the piano, points which come close to the keyboard of the second piano as effects on the keys, and points away from both as auxiliary sounds, non-pianistic in origin. Figure 8 shows the interpretation of graph BT 54 as it appears in Tudor’s first realization.

Figure 8

Figure 8. Interpretation of graph 54 BT from Tudor’s first realization of Concert for Piano and Orchestra/Solo for Piano. Top-to-bottom in Cage’s score becomes left-to-right. Image reproduced courtesy of the Getty Research Institute.

In Tudor’s notation, a rectangle containing a left-facing arrow indicates an auxiliary sound source placed to his left, while a rectangle with a right-facing arrow indicates a sound source to his right. Unfortunately, on the 1958 premiere recording, most of Tudor’s reading of BT 54 is not audible over the raucous sounds of orchestra and audience, except for the use of an amplified Slinky toy (referred to as “coil” in Tudor’s notes). Holzaepfel describes Tudor’s use of this device:

Not until I recently saw… Tudor performing [Concert for Piano and Orchestra] did I realize that one of the most “abstract” electronic sounds… is produced simply by hanging a “Slinky” toy from a microphone stand, attaching a contact microphone to it, manipulating it by hand, and amplifying the resulting sounds.[32]

This distinctive sound is one of the few common elements between almost all the recorded versions of Concert/Solo. In Indeterminacy, it is again associated with graph BT 54. Figure 9 shows page 84 of the second realization, which contains one such reading from BT 54 (corresponding to the second-to-last reading in the first realization). On the 1982 recording of Solo, Tudor skips over this particular reading of BT 54. The reason why involves an unusual interpretation of graph P.

Figure 9

Figure 9. Page 84 from Tudor’s second realization of Concert for Piano and Orchestra/Solo for Piano. The rectangle containing a right-facing arrow is a reading from 54 BT. The Y-shape with the number above it is a reading from 9 P. Image reproduced courtesy of the Getty Research Institute.

Graph P is similar to the previously mentioned BY, and like BY, it is one of Cage’s least specific graphs (see Figures 10-11). Instead of specifying pitch area, P only specifies dynamics. Readings of P can also include “any noises (including auxiliary).” Instances of graph P are ignored more than any other graph in Tudor’s second realization. It is not linked to the activation of electronic sources, so it is not used in Indeterminacy. For the most part, it is completely ignored in the 1982 recording of Solo as well, with one exception. After Tudor skips over the reading of graph BT 54 on page 84, he activates a sound of unknown origin at the reading of P 9 on the same page. The sound can be described as a cross between a loud motor and a ratchet, with noise focused around the low end of the frequency spectrum. The sound is preceded by about 40 seconds of silence, and Tudor lets the sound continue uninterrupted for over a full minute, during which it winds down, becomes quieter and more textured. There is no other sound on the recording like it, before or after. After the sound has mostly faded away, Tudor compresses the next four sound events into six seconds, even though they are spread out over two pages. In the 1992 Concert recording, the “coil” sound serves a similar function; he ignores the readings of BT 54 and uses readings of P 9 to trigger the amplified coil at a similarly climactic moment near the end of the piece.

Figure 10

Figure 11

Figures 10-11. Graph P from page 9 and its continuation on page 10 from Cage’s Solo for Piano. John Cage, Concert for Piano and Orchestra © 1960 by Henmar Press / Edition Peters. All rights reserved. Reprinted with permission from C.F. Peters Corp.

Here, Tudor seems to break with Cageian tradition by exerting intention over sounds, and to break with his own tradition by letting time pass without “watching” it. But this may have been an outgrowth of Tudor’s experience working with electronics in his own compositions. Tudor was drawn to unpredictable sounds that took on a life of their own, which accurately describes the character of the climactic sounds in his last two recordings of Concert/Solo. Tudor was not averse to including climaxes in his own work, according to Matt Rogalsky:

Tudor had much more of a romantic soul than Cage and was quite shameless (his word) about deploying very traditional musical gestures—for example, his instruction to John D.S. Adams regarding Neural Network Plus (1992), that there should be from four to six climaxes within the performance. Perhaps this is not surprising, given Tudor’s love of nineteenth-century piano repertoire, which friends recall him playing during the 1950s for his own enjoyment late into the night.[33]

This is at odds with the perception of Tudor as the performer who chose not to maintain a repertoire of classical music because the time-conception of contemporary music was so radically different.[34] How is it possible to reconcile this contradiction?

Cage also had a knack for the contradictory, for crafting koan-like aphorisms that initially seem nonsensical before revealing meaning. One that seems especially relevant here is “Permission granted, but not to do what you want.” In the Concert for Piano and Orchestra, Cage’s goal was to open up a universe of possibility not driven by desire or intention. But even for Tudor, it was impossible to completely eliminate intention from his performance. In practice the goal became the practice itself, the process rather than the end result. (Why else, after all, would Tudor continue to perform and Cage continue to compose?) In the midst of this non-linear process, it makes perfect sense for the radically rational Tudor and the retro-Romantic Tudor to peacefully coexist. Or, if permission is granted for me to bastardize an aphorism for Tudor: “Let sounds be themselves, but some more than others.”


1. John Cage, “Indeterminacy,” Little Cambridge Design Factory, p. 17.

2. John Holzaepfel, David Tudor and the Performance of American Experimental Music, 1950-1959 (Ph.D. dissertation, City University of New York, 1994), p. 208.

3. Joe Panzner, “Crises of Authenticity,” Stylus Magazine.

4. Petr Kotik, April 1992 interview with Eric Salzman, contained in CD booklet notes to Concert for Piano and Orchestra/Atlas Eclipticalis, Wergo WER6216-2.

5. John Cage, For the Birds (Salem, NH: Marion Boyars, 1981), p. 178.

6. John Cage, “Indeterminacy,” p. 17.

7. David Tudor, “From Piano to Electronics,” Music and Musicians 20, no.12 (1972), p. 24.

8. Kotik, CD booklet notes for Concert, pp. 25-26.

9. Petr Kotik, personal communication with author, Feb. 5, 2009. The conductor’s part contains a column labeled “clock time” and one labeled “effective time,” and the conductor, acting as a “human clock,” converts the former into the latter. For example, converting a clock time of 1’30” to effective time of 15” would require the conductor to move his left hand from directly above his head (the position for 0 seconds) to pointing directly left (the position for 15 seconds) over the span of a minute and a half. The part also contains a third column of “omission numbers,” but even Kotik admits that he has “no idea” what Cage meant by this, and Kotik’s advice is to simply ignore it.

10. Since Tudor recorded the piece both with orchestra and without, for purposes of this paper the titles Concert for Piano and Orchestra and Solo for Piano are basically interchangeable.

11. Holzaepfel, “David Tudor and Performance,” p. 205.

12. Holzaepfel, “Reminiscences of a Twentieth-Century Pianist: an Interview with Tudor,“ Musical Quarterly 78, no. 3 (1994), p. 636.

13. Ibid., p. 634.

14. Tereza Havelkova, “Petr Kotik’s Umbilical Cord,” Czech Music (Jan-Feb. 2003).

15. Holzaepfel, “David Tudor and Performance,” p. 212.

16. Ibid., p. 216.

17. Austin Clarkson, “Composing the Performer: David Tudor and Stefan Wolpe’s Battle Piece,” Musicworks 73 (Winter 1999), p. 31.

18. John Cage, “History of Experimental Music in the United States,” Silence (Middletown, CT: Wesleyan University Press, 1961), p. 70.

19. Eric Smigel, Alchemy of the Avant-Garde: David Tudor and the new Music of the 1950s (Ph.D. diss., University of Southern California, 2003), p. 113.

20. John Cage, Music of Changes, New York, London: Henmar Press / C. F. Peters Corporation, 1961.

21. David Tudor, “From Piano to Electronics,” p. 24

22. David Tudor, interview with Teddy Hultberg, Electronic Music Foundation.

23. Holzaepfel, “David Tudor and Performance,” pp. 239-40.

24. David Tudor Archive, Getty Research Institute, Box 107, Folder 10.

25. Holzaepfel, “David Tudor and Performance,” p. 243.

26. David Tudor, “From Piano to Electronics” p. 24.

27. John Cage, “Indeterminacy” p. 18.

28. Holzaepfel, “David Tudor and Performance” p. 309.

29. Austin Clarkson, “Composing the Performer,” p. 27. In fact, many of the modifications and omissions Tudor makes to his realization may be justified by his increasing physical frailty.

30. An exhaustive catalog of all graphs interpreted by Tudor is beyond the scope of this paper, but chapter 4 of Holzaepfel’s “David Tudor and Performance” covers one reading of each graph type from Version 2 of the second realization.

31. The note written is a G-sharp, but Tudor either misreads it, or consciously chooses to extend it, since the edge of the outline does extend slightly higher.

32. Holzaepfel, “David Tudor and Performance,” p. 312.

33. Matt Rogalsky, “David Tudor’s Virtual Focus,” Musicworks 73 (Winter 1999), p. 21.

34. Eric Smigel, “Alchemist of the Avant-Garde,” p. 148.

Bibliography and Discography

The David Tudor Papers, 1994-1998 (bulk 1940-1996), Getty Research Institute, Research Library, Accession no. 980039.

Books and Articles

Cage, John. For the Birds. Salem, NH: Marion Boyars, 1981.

______. “Indeterminacy.” Little Cambridge Design Factory,

______. Silence. Middletown, CT: Wesleyan University Press, 1961.

Clarkson, Austin. “Composing the Performer: David Tudor and Stefan Wolpe’s Battle Piece.” Musicworks 73 (Winter 1999): 26-31.

Havelkova, Tereza. “Petr Kotik’s Umbilical Cord.” Czech Music (Jan-Feb. 2003),

Holzaepfel, John. “David Tudor and the Performance of American Experimental Music, 1950-1959.” PhD diss., City University of New York, 1994.

______. “Reminiscences of a Twentieth-Century Pianist: An Interview with David Tudor.” The Musical Quarterly 78, no. 3 (Autumn 1994): 626-36.

Panzner, Joe. “Crises of Authenticity.” Stylus Magazine,

Rogalsky, Matt. “David Tudor’s Virtual Focus.” Musicworks 73 (Winter 1999): 21-23.

Smigel, Eric. “Alchemy of the Avant-Garde: David Tudor and the New Music of the 1950s.” PhD diss., University of Southern California, 2003.

Tudor, David. “From Piano to Electronics.” Music and Musicians 20, no.12 (1972): 24-26.

______. Interview with Teddy Hultberg. Electronic Music Foundation,


Petr Kotik, e-mail correspondence, 5 Feb 2009.


Cage, John. The 25-Year Retrospective Concert of the Music of John Cage (Wergo WER 62472).

______. Concert for Piano and Orchestra/Atlas Eclipticalis (Wergo WER6216-2).

______. The Piano Concertos (Mode 57).

Cage, John and David Tudor. Indeterminacy: New Aspect of Form in Instrumental and Electronic Music (Smithsonian Folkways SFW40804).

Tudor, David. David Tudor Plays Cage and Tudor (Atonal ACD3027).