This is a study in musical performativity, taking up the challenge from recent studies in musicology (and in related fields of theatre, art, and cultural studies). The author claims that the shift of perspective from score to performance is definitely fruitful and needs to be taken further. However, the distinction between musical writing and playing is not as sharp as some may think. The study of musical performativity needs to be rehistoricized, knowing that the 19th century idea of the 'work' already implies questions of interpretation, aesthetic experience, and change.

To enter this field, the author elaborates on two musical examples, taken from the practices of Pierre Boulez as a modernist composer and a conductor of classical music. The article’s first part discusses some important changes in the musical articulation of his work Pli selon pli from 1957 to 1989. The second part points to striking developments in Boulez’s Mahler conducting by comparing recordings of his 9th Symphony from different periods. The examples invite a reassessment of historical and interpretative questions. How does the Austro-German tradition influence Boulez’s conducting, as well as his own compositional practice? What was the role of his serialist writing in the first place? And what can be said about the interplay between composition and conducting in this case?

Keywords:Musical performativity,Boulez,Mahler,composition,conducting

What conceptions of musical articulation govern the serialist writing of Pierre Boulez in the 1950s? What happens when this diehard modernist composer takes up orchestral conducting and delves deeply into the late romantic repertoire of Wagner and Mahler through the late 1960s and ‘70s? What changes can be noted when Boulez then returns to Pli selon pli, his main work from the late 1950s, and profoundly revises its musical form in the 1980s? And what differences in musical articulation can be traced when Boulez the conductor again takes up Mahler’s orchestral works, and records them in new versions in the years around 2000? These are central questions that I wish to address in this article, which is mainly a study in musical performativity.

Interpretations of Performativity

During the last few years, the concept of ‘performativity’ (Performativität) has emerged as a weighty challenge to the traditional practices of musical analysis, historiography, and interpretation. The so-called performative turn is claimed to have changed the optics of several related fields, such as the studies of theatre, art, music, literature, media, social practices and rituals, and it has also involved cultural studies, anthropology, philosophical aesthetics and speech-act theory.1 Some call this a veritable shift of paradigm in the history of the humanities2 – from semiotics to linguistic performance (Austin, Searle), from structuralist to performative poetics (Derrida, Felman, Hillis Miller), from textual theory to performative aesthetics (Fischer-Lichte, Schechner), and from biological to performative theories of gender identity (Butler).

Despite its plurality of functions, the term “performativity” displays a deepening actuality in aesthetics and in arts. This may be seen as an indication of a general change from essentialist conceptions to a more dynamic understanding of the art work, which may eventually instigate a stronger interest in the temporal and transitory aspects of the work of art. Rather than focusing on “given” structural characteristics of the work, a performative perspective can enhance an inquiry into how the work takes on form through a play where the author, the performer, the interpreter, the reader, the spectator, or the listener are mutually conditioned by aesthetic practices, cultural processes, and regulations.

Also in the field of musicology, the discussion of terms like performance and performativity has almost reached the level of a new vogue.3 General performative theory, however, must be critically scrutinized if one tries to reinvest it in the field of musicology. Moreover, the views on music are not always as new as some may think.4 Instead of heralding in some completely new paradigm of performativity, there seems to be a need to reintegrate central concepts from a longer tradition of musical aesthetics. Arguably, in the field of musicology, several related questions of performance practice, musical competence, cultural context, aesthetic constitution, and tacit, bodily knowledge are methodologically based on categories that have penetrated the discourse on music throughout the last two centuries or more. Although these premises may not always have been explicitly stated, I will claim that categories of performativity are implicit in existing concepts of ‘work’, ‘musical form’, ‘interpretation’, ‘musical meaning’, ‘aesthetic experience’, and ‘tradition’ (all of which are formative concepts that emerged in musicological thinking mainly through the 19th century).

In this article I shall deal with the broad and promising methodological field of performativity – not through theoretical argument, but through the discussion of some carefully chosen analytical examples. Nevertheless, the theoretical impact of the performative turn, in all its different variants and versions, undeniably forms a set of methodological premises for my concrete treatment of the examples. A main starting point is the presumed “reversal” of the relation between the musical work “itself” (the musical composition) and its necessary performative enactment (the performance of the music). An influential musicologist like Nicholas Cook criticizes “the extraordinary illusion—for that is what it is—that there is such a thing as music, rather than simply acts of making and receiving it.”5 This illusion, he continues, “might well be considered the basic premise of the Western «art» tradition.”6 Here Cook challenges the common idea that in music, there is a “primary” level of written works, which are then in the second round reproduced in acts of performance. Conversely, he implies that the act of performance is constitutive to the idea of the “work” as such. Thus, the presumed “original” is dissipated into a field of performative practices.

This assumed reversal of a hierarchic opposition may seem to be drawing on the poststructural discussion of the reversal of the idea of “première” and “répétition”, or of “copy” and “original”, for instance in the early Derrida.7 However, the main point in Derrida’s argument is that a simple reversal of oppositions is of little use, since it hardly changes the very model of thought, but only reiterates it by negation. Any dichotomic division such as this suffers from the fact that both sides of the dichotomy have always already invaded the other side. The dichotomy emerges as a dynamic and temporal play (jeu) where one side always presupposes the other. There can be no idea of an original, or first, without the idea of a copy, or second, and in this sense, the second also constitutes the first. In the final analysis, the act of mimesis cannot be said to imitate some pre-existing original or essence, but only to reiterate the very scheme of imitation itself. On the other hand, this interpretation paradoxically confirms that there is a certain activity of imitation involved.8 The act must merely be given a different name, for instance the name of productive mimesis, of performative difference, or perhaps simply, of “becoming.” 9

The paradoxical play of re-presentation, or of philosophical constitution, has been known at least since Kierkegaard’s lucid discussion of “gentagelsen” (the “repetition”) in 1843, or in a different sense, since Hegel’s problematization of relational constitution in his Phänomenologie des Geistes in 1807, or even, and more likely, ever since the central dialogues of Plato (such as the famous discussion in Parmenides).10 This paradoxical play of productive difference is astutely taken up by thinkers like Derrida, Barthes, Deleuze, and Foucault. The mentioned “reversal” comes back later in a new disguise in the basic argument of Judith Butler, who coins it in her now famous distinction between “expression” and “performance,” before relegating it to the field of political interpretation of gender.11 As a theoretical model, Butler’s argument, in my view, offers little more than a translation into current American terms of what she herself denotes as “French theory,” obviously perfectly aware of the dangers of making simplifying unifications of disparate theories in her gesture of trans-Atlantic deportation.

The study of musical performativity offers an alternative to both historical performance practice and existing performance studies that are related to computer-based measurement.12 In current musicological thinking, it is also of key importance not to fall back into simplistic models where an assumed liberation from the “work” and a transition to the idea of “performance” (or of performativity) is taken for granted without any further development of the epistemological categories that are at large here. In short, inherited concepts of “musical work” and of “writing” are already part of dynamic circulations whose complexity by far transgresses current attacks that seem to misread them as closed and simple categories. In the following pages, I shall try to unfold some of the highly complex play (or jeu)13 that is always already taking place between compositional writing, musical work, performance, analytic interpretation, and aesthetic experience – and try to demonstrate the inherent performativity of this whole field of central concepts of romantic and modernist, musical-aesthetical thinking.

The Case of Boulez

The relation between the postwar modernism of Pierre Boulez (born 1925 in France) and the great tradition of 19th and early 20th century music has occupied my interest in several previous studies. I shall here invite to a reassessment of some fundamental aspects of Boulez’s musical practice from the late 1950s till today, focusing on the crossroads between his multiple roles as composer and conductor, and also drawing on his contributions as a theorist and an essayist on music. The main point of my article, then, is to discuss how the presumed performativity of Boulez the practical musician may be of fundamental importance in understanding his handling of musical enactment and shaping, both in the fields of composition and conducting.

In the first part of the article, I will quite rashly present some of my findings concerning the use of serial techniques in Pierre Boulez’s work Pli selon pli – portrait de Mallarmé. These findings actually seem to shatter the very idea that his serialist composition in the 1950s and early ‘60s is primarily a matter of securing structural unity, coherence, and rational compositional control. On the contrary, central aspects of the generative procedures are tied to a dimension of radical unpredictability. This unpredictable dimension is a decisive part of the procedures, allowing the compositional subject to retreat and to delegate the choices to the “anonymous” mechanisms of serial generation. Then in the final round of composition, or of revision, salient features of the musical texture and surface seem to be decided on and articulated through free aesthetic choices, tacitly guided by accumulated experience, musical intuition, and aesthetic taste – dimensions that are all embedded in a necessary field of performance practice. In the light of these observations, there is a need to reassess the very idea of serial composition in Boulez, and to see it as a part of a different kind of aesthetic project than the usual textbook presentation of “modernist rationalism,” “structural coherence,” “technical control,” and “rupture with tradition.”14

The second step in the article’s argumentative strategy then follows: The exploration of possible parallels and interactions between his two performative practices of composing and conducting. There is, precisely in the decades between 1959 and 1982 (where he composes and later revises his grand “Improvisation III”) an intensive development in Boulez’s activity as an orchestral conductor. Starting mainly by performing contemporary music in the ‘50s, Boulez gradually expands his musical repertoire backwards, going through early modernist composers like Stravinsky, Webern, and Debussy, to the late 19th century music of Wagner. After his enhanced conducting practice in the 1960s and ‘70s, he then embarks on a new phase of intensified compositional work, resulting in new, large-scale pieces like Répons (1981–88) and …explosante-fixe…(1991–93), as well as in the revision of main works like Pli selon pli (in 1980–89).15 Through this phase, there emerges an emphasized feeling of formal continuity and organic phrasing in his musical articulation. My assumption is that this should be seen in the light of his experiences of conducting the repertoire of large-scale musical works, especially by Wagner and Mahler, and Debussy, Berg, and others. Not only the encounter with the musical works, but also the experience from collaborating with leading symphonic orchestras of the great Austro-German tradition may be seen as profound sources of influence here. In this light, a new discussion of Boulez’s relation to the musical tradition of the late 19th and early 20th century music becomes inevitable, not only as concerns his conducting practice, but concerning the very conception of musical form and articulation in his own compositional output.

During the 1990s and ‘00s Boulez then re-enters the symphonic world of Gustav Mahler in a remarkable series of recordings on Deutsche Grammophon. In the final part of this article I shall study one of these recordings more closely, comparing it both with a much earlier (and more awkward) recording by Boulez of the same Mahler piece, and with select performances by Bruno Walter, Herbert von Karajan, and Leonard Bernstein. This is another contribution to the question of the role of performativity in the conception of musical understanding.

Reassessing Serial Composition in Boulez

The five-movement work Pli selon pli – portrait de Mallarmé for soprano and orchestra was composed, re-built and revised in the course of several periods between 1957, 1959–60, 1961–62, 1981–82, and 1989. For Boulez scholars it probably no longer comes as a surprise that there are quite strange things going on in the compositional processes of this huge work, including a considerable recycling of material from earlier works by the composer (see references further down). Far from solely generating his musical material “top down” from some central, serial essence or some single, unitary, structural principle, the composer reuses musical textures from different earlier works, and combines these with freshly invented structures. The different textures are cut out, transformed, and pasted together in new combinations, according to an apparent logic of wide-ranging compositional pragmatism.

During the early 1990s, I studied the manuscripts and sketches to Pli selon pli – portrait de Mallarmé at the Paul Sacher Foundation, Basel. The vast sketch material (comprising thousands of pages) falls in different categories.16 For instance, the genealogy of “Tombeau” (the final movement) partly has its own, separate history, and the same goes more or less for the later versions of the opening movement “Don” (although this movement ostensibly also presents quotations from the three middle movements). I shall here focus mainly on the central corpus of the three “Improvisations sur Mallarmé”. Still, during this short article I can only mention briefly the many different technical procedures that are at play within these movements. The details of my analytical findings have been extensively presented elsewhere. My main study on Pli selon pli was published in Norwegian in 1997, and is currently being translated into English.17 Parts of my analyses of serialism have been presented in different international contexts.18 In my book, the technical analyses of compositional procedures are part of a wider interpretative context, including questions on musical form, of Boulez’s theory of composition, of his readings of Mallarmé’s poems and literary poetics, and of Boulez’s particularly complex relation to the European musical tradition and to certain non-Western musical cultures.

Throughout the last four or five decades, structural analyses of Boulez’s music seem to have taken for granted a certain image of serialism, that emphasizes the strive for structural unity and rational compositional control. The analytic literature on serialism is filled with terms like unity, coherence, consistency, integration, order, strictness, discipline, logic, necessity, consequence, and rational compositional control. This rather misleading picture I might call “the unity and control model” of serialism.19 A similar understanding also seems to dominate the general reading of Boulez’s own theoretical articles, from the Relevés dapprenti to Jalons and beyond.20

It is only recently that musicology has begun to see how wrong-headed this current image of serialism has been. The spell of this image has obviously rested on two main pillars, the one being Boulez’s and his proselytes’ own verbal rhetoric around the topic of “coherence and control”, the other more deeply embedded in the familiar formalist paradigm of structural analysis as an “explication” of serial music inside the frames of the ruling musicological practices in Europe and America from the 1950s on.

No doubt, the articles and rhetoric of Boulez as theorist and polemicist are partly responsible for the rationalistic image of Boulez the composer. Nevertheless, there is an obvious tension here, since many of his texts are distinctly ambiguous in this regard, also expressing the necessity of poetic interpretation, of different aesthetic influences, and of an irreducibly unpredictable dimension of compositional practice. For more detailed examples from the compositional processes of Improvisation I, II & III sur Mallarmé, I refer to my mentioned publications from 1997, and later. The principal findings may be summed up like this:

In generating the vocal material of “Improvisation I–III”, Boulez utilizes serial structures that are almost incredibly heterogeneous, when compared to the mentioned expectations of “unity and control”. The many structures comprise material as different as his Douze notations (pour piano), No. 5 and No. 9 (dating from 1945); the incidental music for the theatre piece L’Oréstiè (1954–55); and the flute piece Strophes (1957), plus the re-use of generative structures from Le Marteau sans maître (1952–55) and the Troisième sonate pour piano (1955–57), to mention only some of the more influential sources. Moreover, he handles these sources with an astonishing ease when it comes to the question of choice, transformation, and distortion of the “original” structures. It is a long journey indeed from these highly dissimilar materials and pieces (both in genre, style, and structure) to the instrumental and vocal textures of Pli selon pli. Any belief that all the different textures and techniques are applied to guarantee a structural unity from beginning to end, would seem strangely misplaced and irrelevant, indeed, an untenable position. Neither is there any rational compositional control over the final result right from the start. Rather, the final structural result is radically unforeseeable at the start of the generative processes. The possibilities of the opening material are to a great extent unpredictable, and appear to be produced through the very process of writing.

Here and there in his own theoretical texts, the unforseeable aspect in serial composition is actually underlined by Boulez, from very early on. In the important article “Éventuellement” (“Possibly…”) from 1952 he writes: “From the prescriptions we have been examining in detail arises the unforeseen.”21 In 1957 he writes in “Alea”: “In my experience it is impossible to foresee all the meanders and virtualities in the material with which one starts.”22 Such remarks by Boulez are particularly interesting here because they correspond to peculiar blind spots in large parts of the analytical reception. In the composition of Le Marteau, the row serves as a hidden structure generating a number of different textures, none of which can be completely controlled from the start.

This mode of writing is taken even a long way further in Pli selon pli. In the generative processes of the vocal line, one can find no continuous, unbroken necessity or unity at a structural level, but rather a brand-new generative technique, more or less capriciously invented at each new step. The same goes for the instrumental textures. In Pli selon pli a true multitude of transformation techniques are developed and put into play. This dimension of play is no less characteristic of Boulezian serialism than are the ideas of strict logical unity and rational control. Or to say it – somewhat bluntly – with the famous concluding remark in Michel Foucault’s essay: “[…] Boulez has demonstrated the force to let his rules be broken in the very act that puts them to play.”23

Contrary to what is often believed, then, the role of strict, theoretical deduction appears to be of curiously limited importance in Boulez’s work. The task of his methodic, serialist deductions is rather to produce largely unpredictable results, which all need to be further articulated by free-hand performative acts of compositional and musical intervention. In this sense, the question of musical performativity comes forward as a matter of central importance to the understanding of his compositional thinking.

Once the mere “raw material” of serial structures has been generated, Boulez proceeds to a further articulation of the musical surface. This next step in the compositional process is characterized by free-hand musical choices of a truly performative character, in most cases based on his practical experiences rehearsing the pieces with his musicians. This late phase of composition involves musical phrasing, articulation, the forming of instrumental textures, of musical gestures, of dynamic processes, and settings of the text. At its best, this way of composing is a practical process of performative trial and error.

Trying, Failing, Choosing: On the Writing of “Improvisation II”

In general, Boulez’s sketches and manuscripts tell us very little about this final musical forming of the work. There is at least one place, though, where the sketches are particularly revealing. In fact, they don’t appear to be so very different from the famous sketch books of Beethoven. The Beethoven scholar Maynard Solomon states that even in Beethoven it seems to be difficult to retain the idea of a teleological necessity through the compositional process from sketches to the final result.24 Similarly, I have shown this in detail in Boulez, taking as example the creation of the vocal line in Improvisation II sur Mallarmé: Une dentelle sabolit.25 As in many instances of Boulez’s work, this melodic line consists of two separate structures that have been combined. The “main notes” in the vocal part are written separately, and the “grace notes” are taken from a different source. Boulez sets out to fuse these two structures into one, and his sketches quite stunningly reveal a compositional process of simple trial and error. In that process we can find no less than eleven different attempts to finish just the first musical verse of the sonnet text.26

Here it is obvious that Boulez is working directly with musical phrasing and articulation. He seems especially to be struggling with the opening and the ending of the phrase, trying out different solutions and rejecting them, until he finds a musically satisfying result. One might ask: What criteria does the composer lean on here? Inevitably, the working process involves some kind of aesthetic choices. The criteria for these choices are apparently not formalized in any kind of system, and they seem to be rooted mainly in Boulez’s own aesthetic judgment, in his own musical taste, embedded in his conception of style, which is responding to – and at the same time transcending – tradition. There is no theoretical deduction, following some hidden rules of logical, serial generation that might “explain” the melodic choices. Rather, there is the performative practice of a skilled musician, investing his musical taste in a pragmatic process of trial and error, resting on musical experience and intuition.

Again, in the important article, Le système et l’idée (1986) Boulez writes that the system of generative procedures is nothing more than a crutch, a help for the imagination to get started.27 By this account, he requires serial writing to furnish himself with raw material of structural objects, and then in the next round he chooses from these objects. And what does he choose? “[I choose] what I judge to be good, beautiful, necessary”.28

These free, aesthetic choices in Boulez’s compositional practice seem to have been largely misrepresented in the main bulk of structural musical analyses.29 However, this dimension can contribute to a better understanding of the music, helping the analyst to see what kind of processes and choices are in fact necessary to make the music work. Moreover, the compositional dimension of phrasing and articulation, as well as the shaping into a more elastic formal continuity, seems to evoke the affinity to composers like Berg, Wagner, and Mahler, which is gradually increasing in his compositional practice (and in his theoretical texts) during the last fifty years.

Further on in Le système et l’idée Boulez writes that the work is nothing but a struggle between the system and the idea.30 I already mentioned Beethoven. In certain respects, Boulez’s conception of the work of art is here strikingly similar to the philosophic-romantic category of the work (das Werk) that emerged in early German romantic aesthetics before and around 1800. In that tradition, the musical work does not reveal itself at first glance. It can only constitute itself through repeated performances and listenings. As long as no single interpretation can pin it down, the work has a kind of virtual existence.31 The idea of a musical work stands out not as a fixed result, but as a performative procedure. Such aesthetic ideas are taken up in postwar modernism and even exaggerated there – not only in what a composer like Boulez sometimes says, but in the way he works. This radical openness and performative processuality of the concept of the musical work has, as I said, been largely misrepresented in the main bulk of analytic investigations of Boulezian composition.

Revision of “Improvisation III” from 1959 to 1982

After years of experience with conducting large-scale works and music dramas by Berg, Debussy, Schoenberg, Wagner, and others, Boulez turns back to several of his earlier works and rewrites them in a new mode – revealing a different feeling for long, unbroken lines, gradual transitions, and continuous formal processes. If we take a look at the revision of “Improvisation III” from the version of 1959 to the one of 1982, there are certain developments that spring to the eye and to the ear. The changes are significant both when it comes to detailed figuration of musical texture and to the large-scale conception of musical form.

With its approximately 21 minutes of playing time, “Improvisation III” is the longest and by far the most formally complex movement of Pli selon pli. The 1959 version has been recorded twice by Boulez (around 1969 and in 1981), and the 1982 version once (in 2001).32 What do we find in this movement when it comes to the formal outline of the music? In an ideal world, this question would scarcely be addressed without first looking at the composer’s use of the Mallarmé sonnet, A la nue accablante tu, which is itself unusually complex, even for Mallarmé, both with regards to its labyrinthic syntax, the multiple meanings of several words and their grammatical undecidability, and the poem’s musical play with the phonetic sounds of vowels and consonants. However, any further textual interpretation of the sonnet and its relation to the musical texture must here be omitted.33 Suffice it to say that in the 1959 version, Boulez sets this sonnet to music in a manner that is suitably complex, compared to the intricacy of the text. Focusing first on textual details, he only presents the first three of the sonnet’s 14 verses. Roughly described, the musical form falls into a mosaïque of countless short, separate episodes with sharp musical disruptions between them. In the revised 1982 version, however, these shorts bits and pieces are fused into longer, more continuous phrases and sections.

The overall musical form is profoundly altered in 1982, resulting in a movement that is more balanced, and at the same time emerges with an increased complexity in development, integration and differentiation. The impression of enhanced balance is partly due to the new introduction of verse 4, “Par une trompe sans vertu” of the Mallarmé sonnet towards the end of the piece. The instrumental texture of this verse is simply the “second alternative” of verse 1 taken from the open form of 1959. Many other alternative choices of textures from the former open-form structure are now further elaborated and fused into one. This goes for the textures of the harps, of the xylophones, of different percussion instruments, guitar and mandoline, and of the mixed ensemble of winds and strings, which earlier were mainly presented as distinct and separate structures. Very often, the short-breathed textures of a typical punctualist, serialist style are now rephrased and stretched into more continual, “diagonal” textures of a wider and larger kind of gestural language.

One trait that adds considerably to the increased formal complexity is the inclusion, apparently as late as in 1983–84, of all the sonnet’s 14 verses at the very beginning of the movement.34 Strangely, this presentation overlaps with completely different generical textures and formal sections of the piece. The relation between the chronological order of the sonnet verses and the many-layered presentations of these through the musical work gives “Improvisation III” a formal complexity that is extraordinary even in Boulez’s œuvre. As to the listener’s experience of the formal process, what mainly comes across is the striking abundance of flexible musical lines, the almost ungraspable rhythmical richness, and the “diagonal” textures and gestures, less reminiscent of the semi-pointillistic articulation of the 1950s or ‘60s (like in Structures I, Le Marteau, or Le Troisième sonate) and actually more similar to the dynamic musical form of post-romantic, Austro-German symphonic processes in Berg, Mahler, or the late Wagner. This dynamic impression comes across to the listener despite the fact that the musical material carries virtually no traits of the thematic-motivic language of tonality, regular pulse, or traditional polyphonic texture. Instead, the dominant texture is that of heteronomy, marked by the typical “attack-resonance” figurations.

Still, the overall gestural language retains some strong dynamic reminiscences of late romantic tradition. This free-hand disposition of a pre-existing, serially generated material, which has been further developed and rephrased in processes of compositional trial and error, partly through the nearly endless writing processes, and partly through rehearsing the material with his own ensembles, is another example of the acute role of performativity in Boulez’s composition.

Boulez, Performativity, and Mallarmé’s Poetics of Writing

The dimension of “productive writing” in Boulez’s serialism throws important light on his explicit affinity with the Mallarméan poetics of literary écriture. His explicit reading of Mallarmé’s literary poetics, as well as his use of Mallarmé sonnets in Pli selon pli is of great interest here. The deep kinship between the German romantic idea of “absolute Musik” and Mallarmé’s idea of a poésie pure has been pointed out by, among others, Carl Dahlhaus,35 but it is rarely discussed in any closer technical detail in the literature on Boulez. Like the many earlier studies on Boulez and Mallarmé from the 1960s and ‘70s, also the otherwise lucid literary study of Mary Breatnach from 1996 lacks a closer analytical reading of compositional structure.36

In a certain sense, Mallarmé considers the poem not as a fixed result, but as a strategy for reading.37 To read a poem in this sense amounts to rewriting it, which suggests a parallel to the act of interpreting a score by playing it.38 Applied to Boulezian serial composition, this idea involves a fundamental shift in perspective, from regarding the work as a fixed result, to regarding it as a performative procedure – for playing, for interpretation, and for further compositional writing. This also gives an important background to Boulez’s frequent use of James Joyce’s “work in progress” category.

My point is not a simple dichotomy between compositional freedom and serial control. There is no “opposition” between serialism and chance in Boulez. The unpredictable dimension is already inherent in the play of the serial procedures, even when these procedures are functioning at their, supposedly, most strict. Boulez’s way of serial writing amounts to an idea of “anonymity” on behalf of the author’s voice, as Mallarmé famously put it, giving away the initiative to the words, or, in Boulez’s case, to the productive play of the structural material itself.39 But then, in the second round, the composer intervenes and articulates the final textures according to his musical taste and judgment. This dimension of play (jeu) that penetrates both the early and the final phases of the compositional process amounts to a performative practice, a dimension of performativity that is bound to the compositional “here and now” and to the singular musical choices, forming a constitutive dimension that is indispensable even to the “strictest” modernist practice.

Through this specific, performative practice, the very categories of “work” and of “art” are actually thrown open, drawing some radical – and largely unacknowledged – consequences from the tradition of early German romantic aesthetics. To grasp the particular ambivalence and strangeness of Boulez’s serialist practice, there is a need to integrate (in our analytical understanding) certain aesthetical insights from modern literary textual theory, and especially from Mallarmé’s poetics of writing, and to combine them with the most thorough technical analyses of the musical works.

In its own, peculiar way, the postwar modernism of Boulez turns out to confirm certain central categories of Western musical tradition, at the same moment as it transcends them, opening up for new kinds of musical productivity and experience. The apparent performativity of Boulez’s musical practice is of fundamental importance in understanding his handling of musical form both in the fields of composition and conducting.

The performative perspective can shed light on his surprisingly free-hand application of serial rules and techniques from the 1950s on, and on his surprisingly organic way of conceiving musical form and articulation in his later compositions and revisions of his works. As indicated, this perspective may cast new light on his increasingly close relation to the Austro-German tradition of Western art music from the 19th and early 20th century, and also on his connections to central traits of literary poetics (of Joyce, Proust, and Mallarmé) and of pictorial aesthetics (of Kandinsky, Klee, and the Bauhaus movement) in the aesthetic modernity of the early 20th century. Moreover, the performative perspective has also encouraged me to endeavor studying Boulez’s recordings of the already mentioned late romantic repertoire, searching for possible parallels there.

Musical Performativity in Boulez: The Mahler Example

It would be surprising if Boulez’s experiences as a conductor didn’t have traceable influence on his later conception of musical articulation and form. The changes and developments in his preferred repertoire as a conductor is one aspect that is of interest. With certain deviations, Boulez the conductor has largely explored the Western musical tradition backwards, starting with his own contemporaries like Stockhausen, Berio, Nono, and Messiaen in the 1950s, going back through composers like Webern, Stravinsky, and Varèse (in his own concert series Le domaine musical in Paris 1954–67), and then continuing in the direction of Berg, Schönberg, Debussy and Ravel (in productions with increasingly important orchestras, opera houses, and record labels). He embarked on Wagner with Parsifal in Bayreuth in 1966–70, he gave the famous Ring in Bayreuth 1976–80, conducted the world premiere of the completed Berg Lulu in 1979, and then proceeded on to Bartók, Mahler, and a single Bruckner’s 8th in 1996.40 Boulez’s deepening involvement with the great music tradition from modernism and back to the 19th century is also evident in his abundant writings on musical matters, on composition, conducting, performance, on opera, and on other arts (to give a complete overview of his textual production is, however, impossible here). Of course, he has never cut off the contact with the newest music, regularly premiering new works by himself and by younger composers, not least in his work with his favored Ensemble InterContemporain.

Boulez’s involvement with Gustav Mahler’s symphonies and orchestral songs through the 1990s and 2000s is of considerable interest here, not least because his Mahler recordings on Deutsche Grammophon (for instance of Symphony No 6, of No 9, of Das Lied von der Erde, and the brand new one of No 2) are widely regarded as setting new and impressive standards for Mahler performance sui generis. Till now (May 2006), he has in fact recorded all the symphonies and orchestral song cycles of Mahler except Symphony No. 8.

Spontaneously, one would hardly think that Mahler’s music, with its immense expressive intent, its versatile programmatic pretexts, its theatrical play with stylistic figures, its self-conscious inclusion of ostensibly burlesque and trivial genre elements, its dwelling on pain, longing and farewells to life, and its intense personal flair and frenzied temperature of feeling, sometimes verging on the hysterical – one would hardly think that all this would particularly appeal to a presumably crisp, modernist temperament like Boulez’s. Not only is this a matter of Boulez’s stated positions as a modernist composer and theorist right since the late 1940s. It is also indicated through his actual passage to the art of conducting, influenced by certain early mentors belonging to the pointedly anti-romantic generation between the wars, like Hans Rosbaud, Hermann Scherchen and Paul Sacher, as well as George Szell at the Cleveland Orchestra.

For a long time, interpretation of Mahler was rather considered the almost personal domain of characters like Willem Mengelberg, Bruno Walter, Rafael Kubelik, Herbert von Karajan, and Leonard Bernstein, conductors that obviously belong to different interpretative and expressive traditions than Boulez. It is also true that Boulez came quite late to the music of Mahler. He explicitly states that he arrived at Mahler “backwards”, via musical modernism and the Second Viennese School, instead of discovering him “forwards”, from the developmental viewpoint of Beethoven, Schubert, Wagner’s Tristan, and Bruckner.41

On this occasion I am not going to go into the interesting discussion of romantic historical performance practices and their development since the late 19th century. My chosen line of investigation here is mainly to compare earlier recordings by Boulez with later ones by himself. This approach is more directly relevant to the discussion of his compositional development in the same period. In fact, there are highly significant musical differences between Boulez’s recordings of Mahler’s 5th Symphony from 1968 and 1996, of his 6th from 1973 and 1994, and of his 9th from 1972 and 1995, to mention three particularly striking examples.42 Similar and parallel developments can, by the way, be observed in Boulez’s “two generations” of complete Debussy recordings (from the 1960s and ‘90s, respectively) and of the complete Webern (from 1969–70, and 1995–99). For the sake of concentration, in this article I shall focus on his two mentioned recordings of the first movement of Mahler’s 9th Symphony.

Writing and Performance

No doubt, the question of symphonic form in Mahler’s Spätstil is inextricably tied to the question of orchestration. There are even strong indications that the traditional hierarchy that sees compositional structure (musikalischer Satz) as a primary level and orchestration as a secondary one, must here be reversed. Analytical arguments support the view that orchestration in Mahler may be of an ontologically primary importance, whereas the thematic-motivic web of the musical Satz may be of a secondary order.43 Then there is the question of performance in relation to the written score. The “work itself” can hardly be meaningfully grasped, or even imagined, without regarding how it is actually thought to be played. There can exist no valid access to the musical form outside a performative representation of it, through an actual orchestral articulation. In this sense, performance is constitutive to what the work is, or rather, to what it becomes.44

Still, in my methodical procedure I shall not here pursue any rigid purism, solely referring to the audible dimension of the music. Along with my evaluation of the sheer audible output of orchestral sound, I do also work with an eye on the written structures of the score. Contrary to some statements by Nicholas Cook, there is no methodological abyss between performance and written score, no “essential incommensurability” here. In the introduction to their book, the editors sum up Cook’s own argument like this:45 “Where established approaches look for a more or less smooth translation from analysis to performance, he [Cook] sees an essential incommensurability between music as writing and music as performance.” However, the idea of “performance” is larger than this or that single concert or particular recording, since it also already involves the whole field of performance tradition, as well as the theoretical condition of thinking of the performative dimension as such.

Such wider dimensions of tradition are akin to a kind of epistème, a horizon of possible expectations that broadly constitutes what the particular recordings can possibly mean to the socialized members of the culture. The faculty of reading and interpreting written scores (and even of writing them) is one constitutive factor among others inside that historical and epistemological horizon. A recorded performance, then, could rather be regarded as another kind of writing – of écriture – as I see it. This audible inscription or écriture is supplementing – and not supplanting – the score, and both kinds of “writing” are inscribed in the cultural traditions of performance practice.

Boulez’s Recordings of Mahler’s 9th in 1972 and 1995

In the large-scale Andante (1st movement) of Mahler’s 9th Symphony, the form is well known to be utterly complex, with an abundance of multilayered structures and developmental ambiguities that have been discussed through nearly a century of heated analytical controversy.46 The formal complexity probably even surpasses that of Mahler’s other “ungraspable” pieces, like the opening movement of the 7th or the finale of the 6th symphonies. Till now, it has proved difficult even to decide whether the piece at hand (Symphony 9, I) is some advanced and distorted sonata form, some double variation with extra development sections, or some third and singular solution.47 The controversy goes right down to the mere identification of the central thematic figures, of their developmental versions and analytically slippery variants (the term “Variant” is here taken in Adorno’s sense, where the intervallic, harmonic and rhythmic details may be altered, whereas the global identity of the theme remains recognizable as the “same”).48

Between the great orchestral buildups of the movement there are thin and fragile textures of a nearly unheard-of disparity and transparency. For the conductor, these are no easier to handle than the grand climaxes. At every point of the temporal process, the conductor has to make decisions concerning the priority of leading voices and subordinate figures, the handling of rhythmical breathing and agogics, of buildups and peaks, of direct emotionality versus irony or distance, and of micro-level phrasing versus long-range dramaturgical disposition.

One might want to counter, here, that these kinds of difficulties are scarcely foreign to any symphonic music. The particularity of Mahler’s Spätstil, though, is tied to a more radical primacy of the dimension of orchestral doing over the chimerical stability of the written text. The performative choices and solutions radically define the profile of the orchestral constituents that determine what the form of the piece is seen to become. In the late Mahler there is also the paradoxical relation between the (almost legendary) exactitude of the score’s verbal indications on the one hand, and the actual undecidability of formal articulation on the other. In any case, every major performance of the work has to present its own solution to these formal and musical riddles.

What We Hear

Boulez’s version from 1972 (in the following called B72) is a live recording with the BBC Symphony Orchestra in London.49 When it comes to sound, the production is actually quite bad, with a rather muffled timbral body mainly centered around the middle register, supplied with a somewhat shrill treble profile, a restricted range of amplitude, and a conspicuous dryness and lack of acoustical reverb. Still, it is perfectly possible to single out all the orchestral incidences and phrasings that are of major significance to the formal development and interpretation.

In a short article like this, it is necessary to concentrate on certain sections of the work. First, though, I shall offer a general characterization of my global impression. In one word, B72 represents what I will call a literal reading of the score. What it offers is a succession of different orchestral textures, following, with a virtually minute precision, the markings and the processual changes and ruptures that are indicated in the score. The performance is solid, reliable, “objective”, no-nonsense, and matter-of-fact in its attitude. Obviously, such a kind of interpretation has its clear advantages. Any subjective, expressive, personal, or emotional additions to the written text may seem to be abolished.

However, this is not quite true, and that for two reasons. First, the musicians of the orchestra quite naturally add to the feeling of warmth and expression. This is particularly obvious in some of the solos, especially in violin (for instance in the Fourth return of the A-theme group (mm. 268–70) and in the Coda (mm. 434–440 and 445–458)), where the soloists play with their outmost expressive vibrato. As a general rule, one gets the impression that the living musicians themselves, embedded in a long orchestral tradition, add to the spontaneous expressivity of the textures with their own performative gestures. Anyway, Boulez wouldn’t stop them, partly since there are indications in the score saying for instance, “sehr zart, aber ausdrucksvoll hervortretend” (mm. 269–70). (And in any case, he couldn’t stop them either, since this is a live recording.)

Second, there is the nearly structural expressivity of the composition itself, which can hardly be held back, even if one tried to (and there is no particular reason to believe that this is attempted here). Mahler has a way of writing out orchestral crescendos and intensifying processes (Steigerungen) by the sheer building of texture and shortening of rhythmical pace, phase and phrase – an orchestral écriture that must result in the familiar, expressive peaks and collapses nearly by itself. Mahler’s own performative experience, as a lifelong conductor of the highest standard, furnished him with the ability to dispose over the orchestral forces in exact, written prescriptions. Better than many other composers, he knew exactly what his score demanded from the musicians and what the probable outcome would be. And still, the score is not everything, far from it. There would be no expressive climaxes without the performers making them that way. And at least, B72 does present the symphonic peaks with considerable force.

Still, the interpretation in many senses differs sharply from what a conductor like Bruno Walter contributes to the piece – although not always as much as one might have expected. Walter’s legendary recording with Columbia Symphony Orchestra from 1961 will serve as a reference here. Bruno Walter (1876–1962) was Mahler’s own pupil, friend, and conductor assistant from 1897 on, and it fell upon him to premiere the Ninth symphony with the Vienna Philharmonic early in 1912. He also premiered Das Lied von der Erde in November 1911, just after Mahler had died. Walter represents the closest one may come to the notion of an unbroken Mahler tradition – whatever that may be. (Famously, Mahler was fighting against the tradition of euphonic orchestral unity of his times, and to this day it is – admittedly – difficult to speak of a “Mahler tradition” in this sense.) On the other hand, Walter’s closeness to his teacher does not mean that his understanding of Mahler’s Ninth may not have developed or changed during the years, nor that it represents some original truth of the piece. (Naturally, the idea of such a truth runs contrary to the basic idea of interpretation as a productive, performative process, both in historical and artistic senses of the word.)

The key words of Walter’s late recording, as I hear it, are musical integration, warmth, humanity, and ethical commitment, terms that will scarcely surprise anyone that are familiar with his musical attitude and practice. Technically spoken, Walter’s version is satisfying in nearly every sense (I state this with the danger of exposing my own prejudices, since I have adored this LP version since I was 20). Perhaps surprisingly, the Walter version is not particularly slow. Neither does it straighten out the sharp edges and disparate ruptures of the score, creating some false harmonious unity of the piece. It is not “romantic” in such a sense (and here I want to add that the concept of “romantic” music has too long been dragged in a direction of sentimentality, which is a grave misunderstanding of an aesthetical world – from Schubert to Mahler, or from Schlegel to Nietzsche – centered around death, lack, longing, and romantic irony).

What Walter takes magnificently good care of, like the Karajan recording from 1982 (see below), is the large symphonic waves of the piece (in a Kurthian sense of the word).50 Contrary to B72, the Walter version takes the necessary time at the crucial moments of change and transition that are so abundant and so important in this movement. This is not a question of slowness, nor of a measurable number of seconds. In Walter, there is just this diminutive agogical waiting, or inhalation, just at the point where something important is going to happen. Unfortunately, B72 does not have this. Or, to be precise, B72 also uses some rubato, but the rubatos are somehow not the right ones, they are not always convincing, and not organic – to use another term that might deserve to be taken into the warmth of musicological actuality again.51

Walter is by no means the expressive opposite of B72. The two extreme opposites here would rather be B72 and the version of Bernstein from 1986. Bernstein’s is intensively packed, not only with expression, but with personal emotion. One only fears that it is rather the personal emotions of Bernstein that we are facing, more than the inherent emotionality of the symphony (to the extent that this emotionality is constituted through the work’s écriture and its wide-ranging performance tradition). In its own style, Bernstein’s version demonstrates a masterly effort in handling the stormy masses of the orchestra. He also addresses the grotesque and the tragic dimensions of the music with great authority. And Bernstein’s is, not surprising, the slowest of all (with a duration of 30’03”). The problem is rather the emotional attitude. Sticking to agogics as topic, Bernstein is constantly waiting or dwelling on the upbeats before any new section of expressive and formal significance. This nearly becomes a kind of mannerism, as if ostensibly pointing to its own emotionality, in a kind of expressive, gestural showmanship.

Compared to this, B72 is miles away. But even Walter, as well as Karajan 1982, presents a rather non-subjective, over-personal commitment to the music itself, as if going through the horrors and desperation of the symphonic waves and transitions with heroic steadiness and courage. This gives a strong foreboding of what Boulez presents in his new version from 1995. At this point, let me briefly sum up the five recordings I am referring to:52

Boulez 1972 (B72), BBC S.O. (26’57”) Boulez 1995 (B95), Chicago S.O. (29’17”) Walter (1961), Columbia S.O. (29’17”) Karajan (1982), Berliner P.O. (28’10”) Bernstein (1986), Concertgebouw O. (30’03”)

One here notes immediately that Boulez 1995 and Walter have exactly the same duration. This sheer quantitative fact should not be over-estimated in itself. However, I shall return to some other, incredibly strong similarities between precisely these two performances on the level of musical understanding.

Orchestration, Phrasing, and Signification

Till now, my remarks have mainly offered some very general characterizations of these different interpretations. It is of little use to remain on such a general level in the following. I shall give some few examples as to the role of orchestration in the formal process, and to the handling of these details in the different recordings.

Under the note “Etwas frischer” (measure 80),53 Mahler marks the transition to a new large section, a section that is variably depicted as the Schlußgruppe of the Exposition (granted that this is a Sonata form) or as the B2 section (i.e. the second variant of the B theme group, granted that this is a double variation movement).54 In any case, this transition is enhanced as a constitutive shift of texture and character through several compositional means. To mention only some of them: An added, extra measure of 2/4 (m.79) acts as a written ritardando with a solo muted horn; woodwinds take over from vlni and vle with contrasting arpeggio figures; the harp and deep strings introduce a broken figure of accompaniment; and then (m.81) the highly expressive b-theme enters in vlni 1, opening up for the strong buildup towards the first main climax of the work (expanding ever stronger from m.92 to the composed collapse at mm. 105–7). At the point of “Etwas frischer” both Walter, Karajan, and Bernstein give every indication of an important transition, charged with expression and dark forebodings – mainly through the means of timing and of expressive profiling of the new motivic figures. However, B72 seems to lack any feeling of an expressive turn here: The transition is passed over without any particular profile whatsoever.

Then, when we enter the famous point of rupture (at m. 108) after the written collapse, both Walter, Karajan, and Bernstein underline the radically new impulse of the silent horns (mm. 108–9) that present the e1-theme from the strange introductory bars of the movement. The deep submediant drone of Gb from m.110 they give out like a heavily loaded blow – although it is written in p. Only the next time (mm. 119f) it strikes through in real written-out ff, after the uncanny fanfares and muted signals of the horns, but this blow is already prepared in their first presentation of the figure, which is, by the way, generally regarded as the beginning of the Development section.

B72, on the other hand, lacks any palpable feeling of dark meaning at this main transition. The e1-theme at m.108 is presented much quicker, keeping up the hectic pace of the shrill trumpets three measures before. Then at m.110, again somewhat abruptly, B72 changes to Tempo I. Subito, just as it says literally, but the musical handling of all this is strangely unconvincing. B72 doesn’t wait, doesn’t listen, doesn’t load the passage with the lurking terror of looming apocalypse that would generally be expected out of the interpretative tradition.

This is our paradox, then: B72 demonstrates his “correct” and “literal” reading of the score, and thereby it manages to place itself outside of the halfway unspoken, unwritten, or half tacit tradition of Mahler performance practice. How should this fact be evaluated and understood? Maybe B72 is literally right – if not, why did Mahler write it like that in the first place? Perhaps Walter, Karajan, and Bernstein (notwithstanding all their other, mutual differences) are doing some sort of “symbolic” or “hermeneutic” reading that is, strictly spoken, beside the point? Perhaps they misunderstand the radicalism and the modernity of Mahler’s score in projecting some idea of “deeper” meaning onto it, presupposing some romantic, metaphysical paradigm of essence versus surface, of inside versus outside, of inner truth versus outward expression, which this particular composition from 1909 is actually struggling to get rid of? Maybe B72 comes closer to the radicality of Mahler’s modern, ontological displacement of the metaphysical figure of interiority and expression?55

Maybe; but I don’t think so. What is mainly at stake here is rather the problematic idea of a division between written text and historical performance practice. This is no alternative of “either-or”. There can be no such thing as a direct performance of the written text. B72 also presents an interpretation, only a different one. The conditions and knowledge of the inherited performance tradition is an important, indeed an indispensable part of the interpretation of this music, and the criteria for a good performance – or an adequate or relevant or important one – are a natural part of this whole performative epistème. If we take a closer look at the musical differences again, we may note that in fact, neither Walter nor Karajan play the p more loudly than B72 does. This is hardly a question of measurable amplitude. They somehow give it an unmistakable intensity, despite the p. And they don’t play the passage very much slower than B72 either. It is just that it sounds weightier in some way. At no point are they going against the written text.56 The score is rather spoken out, it is narrated, in a certain tone, that supplies it with the mentioned kind of meaning.57 It is not what they say, but the way they say it, that differ them from B72. As regards Bernstein, he has his own problem in actually exaggerating these things, like an actor using too big gestures to make his point, thereby threatening to destroy the effect.

Mahler’s Modernity Revisited

It is not possible here to expound these questions in greater detail. What remains in this article is to briefly illustrate how Boulez some 20–25 years later, in his B95 version, has actually internalized the fundamental constituents of what might still be called a Mahler performance tradition. Had this been a more extensive study, and not just part of an article, countless examples could have been taken from this symphonic movement’s rich succession of musical textures and developmental processes, for the benefit of investigating and discussing the performative differences.

Throughout B95 there are some traits that distinguishes it clearly from B72, on the one hand, and Bernstein and Karajan on the other. B95 has a tendency of not automatically positioning the violins, or the strings, in the foreground. Already in the second part of the A theme (mm. 18f.) this is obvious. Woodwinds often come to the fore, but at any rate the voicing of the vlni is generally much more cantabile than in B72 (second presentation of A, mm. 47f.). Contrary to B72, the B95 version often offers dramatic, expressive peaks at the turning points in the formal process, for instance at the coming of the B theme (mm. 29f.) and the first D motif (mm. 39f.). At the brutal ending of the Exposition and the sudden twist into the Development (m.108f.), the weighty authority in B95 even outshines both Bernstein and Karajan, and almost ridicules B72’s palpable lack of sensibility and depth. Still, in B95 the important e2 motif in horns (mm. 112–13) is played with a weak p, like it says literally, whereas Karajan and Bernstein tend to increase the dynamic output to underline that kind of significant moments. A typical trait of Boulez, both in ‘72 and in ’95, is his peculiar interest in the discomforting bass clarinet figure (mm. 122–4), letting presumably marginal instruments and registres come to the fore.

On the level of greater formal waves and buildups, B95 gives the slow and gradual reconquering of a steady musical allure towards A3 (mm. 136–148) in a long-range crescendo of masterly dynamic control, and the steady pace continues with devastating effect through the following phases, right up to the Schlußgruppe (“Mit Wut”, mm. 174ff). In the “Leidenschaftlich” section (mm. 211ff.) there is less passion than in Bernstein, but more expression all the same. B95 brilliantly respects Mahler’s preeminent polyphonic writing in the following section, where the structural question of which voices are to be leading and which are subordinate is rendered literally undecidable. There is, on the other hand, the returning weakness in B95 of letting the trumpets stick out too sharply from the orchestral masses at the peak of certain symphonic waves (mm. 295–98; 357–364). In addition, Boulez struggles with his old inability (known from his many Stravinskys, not least Le Sacre) to be sufficiently burlesque, macabre, and grotesque when such theatrical figures are demanded (see the “Wie ein schwerer Kondukt” section, mm. 321–336).

In a more extensive article than this is, some further examples that would have been particularly telling would be Boulez’s and the other conductors’ different renderings of the important transition to the first B theme group (mm. 27f.), as well as their different handling of the shifting balance between the groups of strings, woodwind, brass, and percussion in several significant passages of the piece. Of interest is also the different renderings of the “Schattenhaft” section (mm. 254f.), of the third grand climax of the movement (mm. 308f.), and not least, of the strange and famous episode (marked “misterioso”, mm. 376f.) which is inserted at the point in the symphony where a fourth and final emotional climax might have been expected.

One particularity of this movement’s compositional texture is the combinatorial play of small motivic cells, the size of one or two measures, mainly in syncopated and complementary rhythmical patterns. This mosaic-like structuring of larger textural panels has certain similarities to the modernist compositional thinking of Boulez (not least in Pli selon pli), and must be like gefundenes Fressen to his task of performing the symphony. At the same time, in B95 we encounter a nearly unprecedented feeling of continuity, mastering the major, large-scale developments and waves of the symphonic web much better than in 1972. The performance is verging on a kind of paradoxical perfection through its unique combination of small-scale disparity and large-scale balance. After years of working with late romantic and early modernist repertoires in important orchestras and opera houses, Boulez at this point is much closer to Karajan than B72 ever was. In any case, Karajan’s version is a strong recording that will remain a reference.

However, the most striking of my findings this far, is B95’s closeness to Bruno Walter. In my view, Walter and B95 are not only the most similar ones in their understanding of central traits of the movement: they are also the victors. At least, I have never heard such a strange and convincing exposition of the eerie, episodic interlude (the misterioso, mm. 376–390) as in B95. Furthermore, like Walter, B95 at one and the same time excels in the exactness of detail as well as in the overall mastering of the long-range temporal development. Also, with their non-seductive attitude, they are both effectively working against the wrong-headed and otherwise dominating convention of reiterating a euphonic symphonic ideal that is actually deeply foreign to the specificity of Mahler’s orchestral thinking.

The convergence of B95 and Walter is a multiple paradox, involving some strange cross-playing of modernity, Werktreue, historical radicality and performative recreation. Not only does B95, with the taste for multiple textural webs and expressive “anonymity”, budge Mahler’s Ninth a long step closer to his own compositional world of distinct postwar modernism. Also, with this recording, Boulez’s musical œuvre itself is set in a new light that makes his own music (like in “Improvisation III”) take on a certain similarity to the actual strangeness of Mahler. And thirdly all this turns out to be already present in the old Bruno Walter interpretation of the Mahler piece from 1961 (!), conjuring up a musical kinship that must have been unintelligible to everyone at that historical point – between Boulez the young French rebel and Walter the old defender of Austrian romantic repertoire and culture.

The year 1961 is one year before Walter died, at the age of almost 86, and one year before Boulez completed the first full version of his main orchestral work, Pli selon pli. Somehow, these two characters were working on the “same” project, but this virtual “sameness” is constituted as such only in retrospect, après-coup, through the multiple differences in time (history), place (culture), structure (work), and the particular performances: Walter’s conducting and Boulez’s composition. Only time could have made this peculiar kinship visible and audible, and the process has taken at least four decades even to become possible to think. It is all a result of, or a question of, constituting parallel musical experiences in a way that demonstrates – beyond comparison – the productive power of difference (or should I say, différance)58 in the faculty of performativity. The very question of difference turns out to produce an ephemeral stability of sameness, but only as a changing effect of the power of performative enactment.

Conclusions and Further Questions

Boulez’s version from 1995 of Mahler’s 9th (the first movement) is here taken as an example of a large bulk of performances and recordings by the late Boulez. The B95 recording, as I hear it, pinpoints an important turn (Wende) in his understanding of the large-scale, continual, dynamic formal processes of the Austro-German tradition. The revision of his own movement “Improvisation III” after 1980, as well as the creation of large-scale works like Répons, of …explosante-fixe…, and of Sur incises in the decades after 1980 point in the same direction, although there it is at the level of composition. Without letting go of a modernist and non-European flavor in his handling of texture and sound, Boulez still takes a long and decisive step towards the idea of dynamic musical form (Ernst Kurth: Formdynamik) inherited from the tradition of romantic symphonic performativity. At the same time, he changes that tradition, or rather: changes our perception of that tradition, through the way he reconceives it in his own interpretations. However, an assertion like this probably creates more new questions than it solves.

The further investigation of those questions entails not least the task of developing more precise terminologies to describe relevant performative differences. Principal questions arise concerning the distinctions and the dynamical play between composition, work, score, and writing on the one hand, and of performance, playing, and listening experience on the other, and both kinds of questions are again related to the historical, aesthetical, social, and cultural contexts and reconstitutions of these factors. Instead of remaining on the level of subjective criticism (or even, of personal idiosyncrasies), the question of aesthetic experience – following Fischer-Lichte – is to be lifted up to the domain of principal discussion by studying these experiences in their generative context of performativity.59

The possible connections between the changing conceptions of musical articulation in Boulez’s conducting and in his compositional practice probably demand their own methodological discussions. At least, there are no automatic lines of influence to be taken for granted here, although some kind of turn seems to be apparent – towards a greater care for continual formal developments and transitions, akin to the sensibility of Austro-German music taken from the First to the Second Viennese School.60 A possible methodological link would be to apply a critically sharpened concept of performativity in both fields (of composition and conducting, that is), suggesting that the scope of systematic performative study be elaborated in a direction that encompasses the better part of a wider field of musical practices. Such a task would ideally involve all the factors of compositional creation and writing, performance practice, listening experience, cultural tradition, and even of analytical concepts, technologies, aesthetical and musicological interpretations. It should also, if possible, set these factors into play in new and unpredictable ways, hopefully helping us to hear, and to see, what was hitherto inconceivable.

Music scores and manuscripts

Boulez, Pierre: Pli selon pli – portrait de Mallarmé:

1958a: “Improvisation I sur Mallarmé. Le vierge, le vivace et le bel aujourd’hui” (small version), Universal Edition 12855, London

1958b: “Improvisation II sur Mallarmé. Une dentelle s’abolit”, UE 12857, London

1967: “Don”, UE 13614, London

1971: “Tombeau”, UE 13616, London

1977: “Improvisation I sur Mallarmé” (large version), UE 16641, London

1982: “Improvisation III sur Mallarmé. A la nue accablante tu”, UE 19521, London

Boulez, Pierre: Sketches and manuscripts:

Sammlung Pierre Boulez, Mappe G, Dossier 3, Films No. 136 (pp. 717–1027); No. 137 (pp. 1–803); and No. 138 (pp. 643–863); in: Paul Sacher Stiftung, Basel

Mahler, Gustav [1912]: Neunte Symphonie. Universal Edition, 1969


Boulez, Pierre: Pli selon pli. Halina Lukomska (sopr.), BBC S.O., Pierre Boulez (cond.)

Boulez, Pierre: Pli selon pli. Phyllis Bryn-Julson (sopr.), BBC S.O., Pierre Boulez (cond.), recorded 1981

Boulez, Pierre: Pli selon pli. Christine Schäfer (sopr.), Ensemble InterContemporain, Pierre Boulez (cond.), recorded 2001

Mahler, Gustav: Neunte Symphonie. Columbia S.O., Bruno Walter (cond.), recorded 1961

Mahler, Gustav: Neunte Symphonie. BBC S.O., Pierre Boulez (cond.), live recording, 1972

Mahler, Gustav: Neunte Symphonie. Berliner P.O., Herbert von Karajan (cond.), live rec. 1982

Mahler, Gustav: Neunte Symphonie. Concertgebouw O., Leonard Bernstein (cond.), rec. 1986

Mahler, Gustav: Neunte Symphonie. Chicago S.O., Pierre Boulez (cond.), rec. 1995


Adorno, Theodor W.: 1960: Mahler. Eine musikalische Physiognomie (Frankfurt/Main)

Adorno, Theodor W.: 1977: Alban Berg. Der Meister des kleinsten Übergangs (Frankfurt/Main)

Andraschke, Peter: 1976: Gustav Mahlers IX. Symphonie. Kompositionsprozess und Analyse (Wiesbaden)

Austin, John L.: 1980: How To Do Things with Words (ed. Urmson and Sbisà) (Oxford)

Barthes, Roland: 1982: L’obvie et l’obtus (Paris)

Bekker, Paul: 1921: Gustav Mahlers Sinfonien (Berlin)

Boulez, Pierre: 1966: Relevés dapprenti (textes 1948–62), (Paris); 1964: Penser la musique aujourd’hui (Paris); 1975: Par volonté et par hasard (Paris); 1985: Points de repère (Paris); 1989: Jalons (pour une décennie) (Paris); 2005: L’Ecriture du geste (Paris); 1995: “Mahler, Debussy, Webern”. CD mit Vortragen, Deutsche Grammophon; 2005: Régards sur autrui (Paris); 2005: Leçons de musique (Paris)

– English translations: 1991: Stocktakings of an apprenticeship (transl. Stephen Walsh) (Oxford); 1971: Boulez On Music Today (transl. S. Bradshaw and R. Bennett) (London); 1986: Orientations (transl. Martin Cooper) (London)

Bowen, José A.: 1993: The Conductor and the Score (Stanford)

Bowen, José A.: 1999: “Finding the Music in Musicology: Performance History and Musical Works”, in: Cook and Everist (eds.): 1999: Rethinking Music Oxford)

Bradshaw, Susan: 1986: “The Instrumental and Vocal Music”, in: W. Glock (ed.): Pierre Boulez – a Symposium (London)

Breatnach, Mary: 1996: Boulez and Mallarmé. A Study in Poetic Influence (Aldershot)

Brown, Clive: 1999: Classical and Romantic Performance Practice 1750–1900 (Oxford)

Butler, Judith: 1990/1999: Gender Trouble. Feminism and the Subversion of Identity (New York and London)

Cook, Nicholas: 1999: “Analysing Performance and Performing Analysis”, in: Cook and Everist (eds.): Rethinking Music (London)

Cook, Nicholas: 2003: “Music as Performance”, M. Clayton (ed.): The Cultural Study of Music, London (pp. 204–214)

Dahlhaus, Carl: 1974: “Form und Motiv in Mahlers Neunter Symphonie”, in: Hermann Danuser (ed.): 1992: Gustav Mahler (Berlin)

Dahlhaus, Carl: 1978: Die Idee der absoluten Musik (Kassel)

Dahlhaus, Carl: 1988: Klassische und romantische Musikästhetik (Laaber)

Danuser, Hermann (ed.): 1984: Neues Handbuch der Musikwissenschaft, Band 7 (Regensburg)

Deliège, Célestin: 1986: “The Convergence of Two Poetic Systems”, in: W. Glock: ed.: Pierre Boulez – a Symposium (London)

de Man, Paul: 1979: Allegories of Reading (New Haven and London)

de Man, Paul: 1983: Blindness and Insight, 2.edn. (London)

Derrida, Jacques: 1967a: De la grammatologie (Paris)

Derrida, Jacques: 1967b: L’écriture et la différence (Paris)

Derrida, Jacques: 1972: La dissémination (Paris)

Fischer-Lichte, Erika: 2001: Ästhetische Erfahrung. Das Semiotische und das Performative (Tübingen/Basel)

Fischer-Lichte, Erika: 2004: Performative Ästhetik (Frankfurt/Main)

Foucault, Michel: 1982: “Pierre Boulez, ou l’écran traversé”, Le Nouvel Observateur, 2.10.1982 (Paris)

Goehr, Lydia: 1992: The Imaginary Museum of Musical Works (Oxford)

Guldbrandsen, Erling: 1991: “Spor av en helhet? Om noen sentrale lesninger av Mahlers 9. symfoni, 1. sats”, in: Studia Musicologica Norvegica 17 (Oslo)

Guldbrandsen, Erling: 1997a: Tradisjon og tradisjonsbrudd. En studie i Pierre Boulez’ Pli selon pli – portrait de Mallarmé (Oslo)

Guldbrandsen, Erling: 1997b: “New Light on Pierre Boulez and Postwar Modernism. On the Composition of Improvisation IIII sur Mallarmé”, in: Søren M. Sørensen (ed.): In the Plural (Copenhagen)

Guldbrandsen, Erling: 2002: “Verkets åpning. Om verkbegrep, estetisk erfaring og transcendens”, in: Studia Musicologica Norvegica 28 (Oslo)

Hatten, Robert S.: 2004: Interpreting Musical Gestures, Topics, and Tropes (Bloomington)

Hegel, G.W.F.: 1807/1971: Phänomenologie des Geistes (Frankfurt)

Hillis Miller, J.: 2001: Speech Acts in Literature (Stanford)

Hovland, Erlend: 2002: Le déplacement ontologique de l’œuvre musicale. Une étude de l’écriture orchestrale de Gustav Mahler (Oslo)

Johnson, Barbara: 1980: The Critical Difference (London)

Kierkegaard, Søren (1843/1964): Gentagelsen. In: Samlede Værker, Vol. 5 (København)

Koblyakov, Lev: 1990: Pierre Boulez – a World of Harmony (Harwood)

Krummacher, Friedhelm: 1981: “Struktur und Auflösung. Über den Kopfsatz aus Mahlers IX. Symphonie”, in: Hermann Danuser (ed.): 1992: Gustav Mahler (Berlin)

Kurth, Ernst: 1925/1971: Bruckner I–II (Hildesheim)

Mallarmé, Stéphane: 1945: œuvres Complètes (Paris)

Morgan, Robert P.: 1991: Twentieth-Century Music (New York)

Plato: 1939: (ed. F.M. Cornford) Parmenides (Indianapolis)

Ratz, Erwin: 1955: “Zum Formproblem bei Gustav Mahler. Eine Analyse der ersten Satzes der IX. Symphonie”, in: Hermann Danuser (ed.): 1992: Gustav Mahler (Berlin)

Rink, John: 1999: “Translating Musical meaning: The Nineteenth-Century Performer as Narrator”, in: Cook and Everist (eds.): Rethinking Music (Oxford)

Schechner, Richard: 2003: Performance Theory (2nd ed.) (London and New York)

Scherer, Jacques: 1957: Le “Livre” de Mallarmé (Paris)

Searle, John R.: 1969: Speech Acts. An Essay in the Philosophy of Language (Cambridge)

Solomon, Maynard: 1991: “Beethoven’s Ninth Symphony: The Sense of an Ending”, in: Critical Inquiry 17, No. 2, Winter

Stacey, Peter F.: 1987: Boulez and the Modern Concept (Aldershot)

Stoïanowa, Iwanka: 1973: “Pli selon pli”, in: Musique en jeu 11 (Paris)

Vermeil, Jean: 1989: Conversations de Pierre Boulez sur la direction de l’orchestre (Paris)

Wirth, Uwe: (ed.) 2002: Performanz (Frankfurt/Main)

Wittgenstein, Ludwig: 1958: Philosophical Investigations (ed. G.E.M. Anscombe) (Oxford)

Zeller, Hans R.: 1960: “Mallarmé und das serielle Denken”, in: Die Reihe 6