The pianist Peter Stadlen worked intensively with Webern on the Variations for Piano op. 27 during the autumn of 1937 prior to giving the first performance of the work, and his testimony is the most detailed of any of the first generation of interpreters of the music of the Second Viennese School:

His comments echo those of the pianist Edward Steuermann and the violinist Rudolph Kolisch, who were able to work directly with Schoenberg and emphasised his reluctance to engage in any discussion of twelve-note technique. One might speculate that this uncharacteristic coyness could well derive from a realisation that emphasis on the organisation of a single component – pitch – could inhibit the integration of all the musical parameters into a convincing interpretation. Be that as it may, the unsubstantiated assertion which opens Stadlen’s account is perhaps more revealing about his own attitude to serial processes than it is about their relative function in the works of Webern and Schoenberg, yet its implications are certainly worth further exploration. Notwithstanding the fact that the order of the basic series itself has not been without musicological controversy and misinterpretation, the Variations op. 27 demonstrate a rigour in the application of dodecaphonic technique which render a purely pitch-based analysis unproblematic (cf. Reference SmithSmith 1986: 210–11). By comparison, even Schoenberg’s first twelve-note works, with their comparatively greater textural variety and occasional use of serial permutation, can present intractable difficulties in terms of analysis of pitch order. The problem is certainly not confined to the Second Viennese composers: the first piece of Boulez’s Structures could almost have been conceived with the lecture hall in mind, yet commentators have been wary of attempting a similar cataloguing of serial procedures in the remainder of the cycle, and with good reason, since it is stating the obvious to point out that any perceived overemphasis on technical matters presupposes that the serial processes themselves yield their secrets to the analyst.

If, in the case of Webern, the transparency of his serial technique can be credited with a ‘vital contribution’ to the perceived ‘beauty’ of a work, how are we to approach the acknowledged link between aesthetic effect and technical means? Fortunately, Webern’s own comments can assist in considering this question:

To pursue this analogy with traditional form a little further, the recapitulation in classical music can indeed be sensed by the listener, but a greater level of awareness of the technical processes is desirable in order to appreciate, for example, the sophisticated humour of a false recapitulation in Haydn or the structural logic behind the subtle changes which might occur in the recapitulation itself. Similarly, whilst the ‘recapitulation’ which occurs in the closing section of the Webern Variations can be sensed intuitively on one level, richer layers of meaning can be revealed by technical investigation. Thus, an analytical strategy which takes account of the return of the series and its retrograde at the original pitch but goes on to a holistic consideration of its deployment in the context of the parameters of rhythm, texture, register, and dynamics is likely lead to an enhanced appreciation of this coda and its function not only as a recapitulation within the third movement, but also as a series of allusions to the opening of the work. Such an approach has none of the certainties of an analysis focused exclusively on dodecaphonic technique, since it is inevitably subject to a personal interpretation of the balance between the component elements, yet it reflects the fact that the score can yield its richness on many levels. Similarly, the characteristic canons and crab canons which are such a feature of the form of each of the three movements of the Variations can be shaped as component elements in the overall phrase structure without the need for understanding the means by which this balance is achieved, or conscious awareness of the convergence between the ‘form’ of the series itself and its elaboration during the course of the work. However, given this perceived convergence between form and technique, Webern’s distinction between ‘how the piece ought to be played, not how it is made’ is a somewhat artificial and potentially restricting one. A performer seeking ways into the style of the music might well decide that precisely such an investigation can lead to a more musically aware interpretation of phrasing and musical structure.

Stadlen’s experience is particularly significant, as his comments form part of the introduction to an annotated score, consisting of markings in the composer’s hand and a transcription of additional comments made during the course of these coaching sessions. As such, although the Stadlen score has been the subject of earlier studies, it is well worth revisiting for the number of issues it raises (Reference WasonWason 1987; cf. Reference Boynton, Grassl and KappBoynton 2002). First, it must be acknowledged unequivocally that the annotated copy is of the greatest practical help to any performer of the work, with its clarification of phrase structure, indications of tempo modifications, expressions of musical character, and additional dynamic and pedal markings. Judging by the extent and nature of the markings, Webern must have worked in fanatical detail, and yet it is no disrespect to the value of the Stadlen score to affirm, as Neil Boynton has observed (Reference Boynton, Grassl and KappBoynton 2002), that the annotations themselves are the product of a composer’s reaction to the playing of an individual during a series of what were in effect coaching sessions on an instrument with its own unique properties of touch and sound quality, in acoustic conditions which cannot be replicated. Broadly speaking, the markings fall into two categories: those that convey details of phrasing and registral connections (the what to do) which may well be taken at face value, and such added details of internal balance in chords and pedal markings (the how to achieve it) which are the product of a coaching session with an individual performer and are therefore of a more provisional nature. Given his reluctance to discuss technical issues related to serialism, it seems highly unlikely that Webern would have shared with Peter Stadlen details of the evolution of the work, and indeed the source material remained unavailable for scholarly perusal for some forty years after the composer’s death. However, investigation of this material can help shed additional light on Webern’s annotations, and on their relationship to the formal processes of the work. The sketches reveal that the opening movement was originally conceived in quintuple time, a stage which was eventually superseded by the greater visual clarity of the published triple metre (cf. Reference BaileyBailey 1995; Reference Boynton, Schweiber and UrbanekBoynton 2009). Turning again to Webern’s annotations, the seemingly inconsistent tenuti marks in the opening system of the first movement reveal themselves as a series of articulations within a series of $\begin{array}{c}5\\ 16\end{array}$ bars, subdividing each bar into units of 3 + 2. One cannot help but wonder whether if Webern had shared this information with the young Stadlen, a similar interpretive result might have been achieved, and with it an enhanced understanding of the rhythmic character of the movement – an explanation of the underlying creative impulse rather than a series of injunctions. Further practical insights, unavailable to an earlier generation of performers, are yielded by even a superficial examination of the drafts for the work. For example, remarkably, an early version of the theme of the third movement contains no fewer than seven changes of tempo, with the contrasts between piano and forte phrases accompanied by ritardandi. The effect, along with the reduced note values of the original $\begin{array}{c}3\\ 8\end{array}$ time-signature is to alter one’s perception of the movement as being much more fluid in character than the rather ascetic appearance of the published score with its $\begin{array}{c}3\\ 2\end{array}$ signature might suggest. Why Webern would have omitted these markings from the score is something of a mystery, especially in view of his concern that the swiftly changing character of his music be communicated in a flexible way. The existence of such a secondary source does not undermine the legitimacy of the published version, but it is an illustration of the fact that all scores, however detailed their notational exactitude, are provisional in their status as records of the composer’s intentions. A further effect of the quadrupling of note values in the published edition is the loss of the connecting beams which were present in profusion in the early sketches. For example, a draft of the first variation in the third movement has a series of connections as shown in Figure 18.1. One might note in passing the recasting of b. 14 into two bars in order that the eleven-bar proportions remain consistent throughout the movement, but more significant is the range of new connections articulated by the original beams. If these groupings are articulated in the context of the comments relating to musical character in the Stadlen score, an enhanced interpretive richness and understanding results. Even more remarkable is the comparison between the printed score and a draft of what became the fourth variation (the inconsistency of bar numbering with the published score is a consequence of the subsequent decision to eliminate two variations). The hand-crossing which is feature of the printed score (bb. 45–56) can encourage a disjointed approach to interpreting this passage, as can the late decision to emphasise the syncopated character of the variation by means of rhythmic displacements and the addition of accents. One can well imagine the dialogue which occurred between composer and interpreter as Webern added pedal marks and indications of phrasing to the score, all features which are implicitly present in the groupings and layout of the original draft (see Figure 18.2).

Figure 18.1 Anton Webern, draft of first variation, third movement of Piano Variations op. 27, bb. 12–23

Figure 18.2 Anton Webern, draft of fourth variation, third movement of Piano Variations op. 27, bb. 56–66

Returning to the first movement of the Variations, even more puzzling than the tenuti marks in the opening bars are the dynamic gradations (< >) placed over individual notes. Stadlen offers the following by way of explanation: ‘The unrealisable “vibrato” signs in bars 2 and 3 … which Webern wrote in my copy, give an idea of the cool, passionate lyricism of expression that he envisioned here’ (Reference StadlenStadlen 1972). Again, an awareness of the context provides an additional perspective on these markings. A letter to Hildegard Jone of 2 September 1937 contains the information that Webern had just put the final touches to a movement for String Quartet. Bearing in mind that the coaching sessions with Stadlen took place in the early autumn of that year, it seems highly likely that Webern had the sonority of the quartet movement very much in his mind at the time and was thinking in terms of the expressive qualities of string instruments as he worked with the young pianist. It is noteworthy that the close of the first movement of the String Quartet op. 28 contains precisely the same dynamic markings – in this context, of course, eminently realisable on string instruments. It could even be the case that in revisiting the piano work during these coaching sessions, Webern might well have become more consciously aware of the quartet-like texture of much of the writing in the first movement of the Variations for Piano.

In view of the information that was not available to Peter Stadlen at the time, his later assertion that, in Webern’s op. 27, ‘an authentic interpretation is impossible without the aid of direct, detailed tradition’ (Reference Stadlen and StadlenStadlen 1979) has to be read in the context of his aversion to the multiple serialism of the post-war period, and (as he saw it) the various – and differing – misinterpretations of Webern’s legacy by the Darmstadt generation (see, inter alia, Reference IddonIddon 2013: 89–100). It is certainly the case that for Stadlen, charged with the responsibility of the first performance, direct access to the composer’s thoughts on interpretation was an important factor in his desire to achieve a performance in accordance with Webern’s wishes – an ‘authentic’ interpretation. Yet the very existence of the Stadlen annotated copy is an admission that the printed score cannot in itself be a comprehensive record of the composer’s intentions, and one might well argue that given the limited performance directions in the published scores, Webern was himself an unwitting contributor to any perceived misinterpretation of his musical legacy. Just as Stadlen’s annotated score offers a counterbalance to an ‘objective’ view of Webern, so musicological research can assist in further clarifying the composer’s intentions – hence assisting in achieving an authentic performance, but one in which the emphasis has subtly shifted from an authenticity based on the composer’s personal intervention to one involving a re-assessment of source material which was unavailable to the first generation of performers. After all, in the final analysis, concern for authenticity is a state of mind in the performer, rather than a checklist of criteria in need of constant updating. As such, the term constantly shifts its focus, as research uncovers fresh evidence relevant to interpretation, and succeeding generations of performers bring their own perspectives and insights to the work in question.

***

Following the resumption of musical life in Germany after the Second World War, Peter Stadlen gave the German premiere of the Variations on 31 July 1948 at the Darmstadt New Music Courses. Within a year of this performance, Olivier Messiaen was invited to Darmstadt to give a performance of his Visions de l’Amen (partnered by Yvonne Loriod), and during his brief visit, he evidently worked on a new piano piece, ‘Mode de valeurs et d’intensités’. Strictly speaking, Messiaen has no part in a discussion of dodecaphonic technique, since however radical a departure it represents, the organisation of the four parameters of pitch, rhythm, attacks, and dynamics in ‘Mode de valeurs’ is based on a free ordering of the component elements, without the constraints imposed by strictly serial procedures. Nonetheless, the piece had a profound influence on the younger generation of serial composers, to the extent that the three projected volumes of Boulez’s Structures were in effect both a homage to, and a critique of, ‘Mode de valeurs’, each book to be based in turn on the three pitch scales of Messiaen’s piece, but employed serially.

It unfortunately remains the case that the works of this brief period of multiple serialism are analysed much more frequently than they are performed, and for the good reason that they present seemingly intractable problems. On one level, performance of a work that specifies a duration, mode of attack, and dynamic for each individual note ought to be relatively unproblematic. After all, the composer’s intentions have been indicated unequivocally in seemingly unsparing detail. Furthermore, Stadlen’s concern for authenticity, rooted in a direct connection with a performing tradition, is taken to a new level by the availability of recordings involving the composer as executant in his own music. All in all, the aspiring performer of contemporary music is seemingly endowed with an unprecedented richness of source material on which to base an interpretation, and yet this plethora of information is a somewhat mixed blessing. Turning to Messiaen’s own recording of ‘Mode de valeurs’, the playing time is a fairly sedate 3′52″, and his performance is a valuable document, not only for his exemplary playing, but for the numerous (almost inevitable) inconsistencies with his own notational exactitude. Equally fascinating are the timings of recordings by two pianists most associated with Messiaen, who worked closely with the composer – Yvonne Loriod (3′25″) and Michel Béroff (2′37″). Leaving aside for a moment the considerable interpretive differences between these performances, notwithstanding the unparalleled precision of the markings, the discrepancies in timing alone are startling – the small-scale equivalent of comparing recordings of a classical symphony of half an hour (Béroff) and forty minutes duration (Messiaen). Given such discrepancies, how is one to approach the music of the period of multiple serialism in pursuit of an ‘authentic’ performance?

The word ‘pointillist’ is frequently (and misleadingly) appropriated as a generic description for the instrumental works of the post-war period, prompted by the visual appearance of scores comprised of a succession of seemingly self-sufficient sounds coalescing into a whole. The paintings of the pointillist group are notable for their dazzling use of timbre, with a constantly fluctuating use of tiny variations of colour as the tones gradually merge. An equivalent representation of such variations of timbre is impossible within the boundaries of musical notation, and this limitation is both a reminder of the provisional status of a written score, and an important factor to bear in mind when seeking to come to terms with scores seemingly already overloaded with performance indications. In the case of ‘Mode de valeurs’, the opening note, E♭, appears in the top part with its register, dynamic (ppp), and mode of attack (legato) unchanged throughout the entire piece. Yet the context in which this note is heard is constantly changing and, although Messiaen for the most part avoids direct clashes, the pitch inevitably takes on a different character and an individual pianistic colour according to its context: in other words, the performer may perfectly justifiably feel the necessity to adjust the dynamic level according to context – when, for example, it appears against the background of a resonating fff B♭ in the bass on the second page, or is sounded directly against a forte C♮ in the middle register at the top of the penultimate page. Rather than being censured as transgressions, such adjustments, whether of dynamics, methods of attack, or almost undetectable changes of timing, are a means of realising the expressive character of the music, and as such are in the tradition of Webern’s annotations in the Variations for Piano. That the piece is subject to quite varied approaches, even by those artists who have a claim to work within a direct and authentic tradition, is no more than an illustration of the role of the individual interpreter in observing the spirit, rather than the strict letter communicated by the score.

On 4 May 1952, Messiaen joined with the composer to give the premiere of the first piece of Boulez’s two-piano work, Structures – a public demonstration of rapprochement, following a period of cool artistic and personal relations between master and former student. In an interview some two years before his death in 2016, Boulez was asked about his approach as a performer to Structure Ia, and his reply began: ‘Well, let me say that it should be as anonymous as possible’ (Reference O’HaganO’Hagan 2017: 329). Although it is likely that Boulez was thinking of the French anonyme in the sense of ‘impersonal’ rather than the exact transliteration of the word, his words fail to address the notational challenges posed by the score when one attempts an interpretation of the piece. The strict application of serial technique to all the musical parameters, with comparatively restricted creative intervention by the composer, results in a series of conundrums for the performer. Even assuming that it is possible to apply twelve distinct dynamic gradations consistently throughout the piece, the operation of serial processes results in a series of notational contradictions. Thus, the extreme dynamic ppp appears throughout one section (bb. 86–97) with the value of a semiquaver and articulated by means of an accent and staccato dot. Whilst it might well be possible to observe the duration and dynamic level for each note, it is virtually impossible to realise these elements in combination with the articulation as marked: because of the wide difference in register between the various pitches, a note in the low bass register will inevitably have much greater resonance than one in the upper treble. The problems with a literal interpretation are multiplied when one considers the ensemble aspects of the piece: at bar 94 in the same passage, the ppp bass A♮ in Piano II coincides with a forte bass F♮ in Piano I. Clearly, for both pitches to be audible, some adjustments will need to be made in the interests of balance. A striking characteristic of the piece, despite its mechanistic elements, is the extent to which the two pianos engage in dialogue, with numerous instances of repetition of pitches and echo effects, especially in the central Lent section. The constant interplay of tritones between the two instruments is an invitation to engage spontaneously in performance with these spatial effects – and notwithstanding the exigencies of notational exactitude, the precise timing and dynamic level of these exchanges is likely to vary in different acoustics and according to the resonating characteristics of the instruments available. An eminently practical musician himself, Boulez as performer was certainly aware of such variables, evidenced by his own performances of Structures, notable for their fidelity to the spirit of the music, if not always for their textual exactness.

Just two months prior to the premiere of Structure Ia, Boulez received a visit from Karlheinz Stockhausen, the beginning of a friendship which endured throughout the coming decade, despite an increasing divergence of artistic aims. The first and fourth pieces of Stockhausen’s Klavierstücke I–IV were written in the aftermath of this visit and, in them, he takes rhythmic proportions to a new level of complexity. These instrumental works of the post-war decade are inextricably linked with the early development of electronic music, with its possibilities for precise measurement of each of the musical components. Unfortunately, the transfer of such precision to the field of instrumental writing can lead to intractable performance problems. No less a figure than Boulez himself expressed reservations about the practicality of such rhythmic complexity – a comment perhaps influenced by the necessity for a revised version of the vertiginously complex rhythms of the second piece of his Structures (Ib) prior to its first performance. By the following year, Stockhausen’s rigorous approach was beginning to shift as he gained experience in the practicalities of sound projection:

This heightened awareness of the practicalities of sound production is apparent in the next group of instrumental works, beginning with the six piano pieces, Klavierstücke V–X, started in 1954 but not finally completed until 1961. Despite the extended timescale of composition and their varied character, all the pieces were originally generated from a single series, with the various musical parameters derived from it. Whilst the detailed markings of dynamics and attack remain formidably demanding for the performer, the complexity of rhythmic proportions found in the first set of piano works has been considerably reduced. Although the absence of bar lines (except in Klavierstück IX) is discouraging at first sight, in some of the pieces in the cycle the rhythms can be grouped into quaver units for the purpose of learning, and a regular pulse felt throughout – especially so in numbers V and VIII. A notational innovation in Klavierstück VI is the addition of a scale of graduations of tempo notated above each system of the score, replacing the notational rigours of Klavierstücke I and IV with a more practical approach to the minute adjustments of tempo which occur spontaneously in every performance. In the case of Klavierstück X, the procedure is taken a stage further with the use of ascending and descending beams as a means of indicating flexibility of tempo. The fundamental importance of these indications of rhythmic ebb and flow to the conception of the piece is confirmed by the existence of a sketch which draws a complete map of the durations and rhythmic patterns complete with connecting beams, as in the final score, but without the accompanying pitches. Since the composer stipulates a fluctuation within the ratio 3:2 for the subordinate duration values, the performer has a degree of latitude within the basic tempo. This freedom appears to be diminished by the daunting injunction to play literally ‘as fast as possible’. However, a basic tempo can be discovered within the piece, and one possible solution is found on page 28 of the published score: a rapidly repeated D♮ is notated as which equals the overall value of a crotchet. Taking a cue from the opening demisemiquaver, the eight rapid repetitions of this first group provide a basic crotchet pulse, which can be modified during the course of the piece within the limits of the 3:2 ratio. Before leaving Klavierstück X, mention should be made of one of the most striking features of the work, the employment of clusters of varying densities – a characteristic which links it more closely than any of the preceding Klavierstücke with the sound-world of the electronic studio and the recently completed masterpiece, Kontakte. Far from being an invitation to produce a cacophony of indeterminate sounds, the clusters are calculated in their density and exactly notated in range: clusters in fourths, played by the wrist, and frequently accompanied by glissandi; ninths, employing the full span of the hand; and three densities which require the use of the arm – an octave-and-a-half, two octaves, and two-and-a-half octaves. Whilst it is the case that there is a degree of imprecision built into the notation, in that it is physically impossible to sound every single note in a large cluster, nonetheless, they can be balanced, and even given a melodic shape. Stockhausen gives each cluster or cluster-group a dynamic marking, often exploiting the delicate effects possible at the p to pp level. The long resonating pauses which punctuate the piece invite the listener to enter the interior of the sound, with the gradual emergence of the next section as the harmonics fade into silence. Klavierstück X marks the end of an era in Stockhausen’s career, being virtually the last fully notated work for a decade, and his last solo piano composition until the group of pieces derived from the operatic cycle, Licht, began to appear during the 1980s.

Meanwhile, any hegemony developing around the post-war serial composers was being undermined by both internal and external forces. Boulez’s continued expansion of serial possibilities by means of the technique of chord multiplication eventually led him to generate chordal aggregates of the entire chromatic field, to the extent that any fundamental relation to Schoenbergian principles of dodecaphonic ordering of the series was lost – in effect, a return to free atonality, epitomised by the freedom of the writing in the second piece of the second book of Structures (1961). This crisis of serialism occurred just a few years after the indeterminate compositions of John Cage and his followers began to receive serious attention in European musical circles – the invitations to Donaueschingen in 1954 and Darmstadt in 1958 being landmark occasions in the gradual integration of Cage’s innovations into the European musical mainstream. Whilst indeterminacy as a compositional philosophy is on one level at the opposite pole from multiple serialism, each composer would react in his own characteristic way to this encounter. Stockhausen’s Klavierstück XI, apparently conceived in 1956, but not completed until the following year, represents his embrace of the challenge, with the published score consisting of a random ordering of the nineteen groups – although the evidence from the sketches suggests that some of the groups were conceived in a numerical sequence. Details of tempo, attacks, and dynamic levels are indicated at the end of each group and are applied to the next group, likewise chosen at random. The practicalities of realising this in performance are considerable, and a strategy which involves a pre-performance shuffling of the groups both preserves the concept of random ordering and allows precious preparation time for the adjustments to the various musical parameters, as dictated by the sequence of groups.

Boulez’s more cautious response in the Third Sonata (1957–63) introduced the principle of performer choice – ‘plugging the performer back into the creative circuit’, as he put it (Reference Boulez and ThéveninBoulez 1991j: 37) – whilst maintaining a considerable degree of control over the formal parameters. A surprising feature of the two published movements of this work is that, far from the introduction of some flexibility of structure being paralleled by a relaxation of serial principles, the techniques are applied with as much rigour as in the early works, with the central movement ‘Constellation’ involving all forty-eight forms of the series. Even the use of the sustaining pedal is subject to serial principles, with three methods of attack identified and rotated – sustained, staccato, and one producing harmonics, each of which may in turn be modified by pedal. The first generation of performers who worked with Boulez noted his emphasis on these effects of resonance rather than any commentary on the issue of performer choice (personal communication with Leonard Stein and Charles Rosen). It may be observed here that Boulez’s fanatically precise notation has to be interpreted in spirit rather than strictly according to the letter: needless to say, individual pianos differ considerably in the play of the sustaining pedal, and the effects of resonance are subject to such variables as quality of piano and acoustic properties of the venue – bearing in mind that the acoustics themselves can be disconcertingly altered by the presence of an audience. As always, a degree of practical adjustment in performance is perfectly reasonable, especially within the ‘Points’ sections of ‘Constellation’, where the resonances of individual harmonics can be lost without some discreet adjustment. More problematic is the introduction of a degree of performer choice in the ordering of the sections, all of which must nonetheless be played. This clearly presented Boulez himself with challenges as a performer – not dissimilar in practical terms to those of Klavierstück XI – and in fact the three privately available recordings of Boulez himself playing ‘Constellation’ maintain the same ordering of sections in each case, paralleling David Tudor’s practice of using an identical realisation of Cage’s score from performance to performance. Even more problematic is the choice available to the performer in the other published movement, ‘Trope’. Here, Boulez varied the ordering of the four sections in his recorded performances but played all of the optional passages in the two sections ‘Parenthèse’ and ‘Commentaire’. As I have shown elsewhere, the logic of this lies in the highly sophisticated use of related series to act as inserted commentaries throughout these two sections (Reference O’HaganO’Hagan 2017: 211–15). It could thus be argued that whilst an authentic performance could involve an omission of some (or even all) of the parenthetical sections in accordance with the composer’s instructions, an awareness of the richness of the compositional process would inhibit one from doing so. Ultimately then, it is the performer’s decision, but a decision hopefully based on an informed choice, bearing in mind the role of the commentaries within the framework of the overall structure. The novel published format of these and some other works of this period disguises the fact that the same interpretive principles apply to them as to the other music of the post-war era: that it is the interpreter’s role to choose between the available options in order to present the most convincing interpretation of the composer’s intentions, whilst remaining mindful of the paradox that one route to greater interpretive freedom lies in an enhanced understanding of the complexities of the compositional process.

The frame of this discussion is a mere forty years or so, encompassing the ‘classical’ phase of dodecaphonic technique. Yet within that limited period, the stylistic range is enormous, and the challenge, as always, remains that of interpreting the intentions of the composer with fresh insight within the context of a performing tradition. That interpretations of the same work can vary so much not only between individuals, but from performance to performance by the same individual, is a reflection not only of the elusiveness of the pursuit of an authentic interpretation, but a continuing affirmation of the multifaceted nature of a work of art, and its constant capacity for self-renewal as the performer aspires to an ever-deepening understanding of its musical essence. As Boulez expressed it: ‘The great works, happily, never cease to reimburse the inviolable darkness of their perfection’ (Reference Boulez and ThéveninBoulez 1991d: 145).