Verbalising the Intangible:
An Exploration towards a Performance of Alfred Schnittke’s Second Piano Sonata

Table of contents

DOI: 10.32063/0103

Alessandro Cervino


Active as a concert pianist, Alessandro Cervino recently appeared in important venues and festivals in Belgium and abroad. In May 2012 he obtained a doctoral degree in the arts at the University of Leuven. He is currently an enthusiastic professor of piano at the LUCA School of Arts, campus Lemmensinstituut, Leuven.

by Alessandro Cervino

Music + Practice, Volume 1


Defining the issue

The score of a composition is made up of musical signs that convey performance instructions. In order to be able to follow these instructions, performers have to know the conventions through which these musical signs can be translated into sound, i.e. the rules establishing a correspondence of each element of notation of a given composition with a precise instruction. These rules change with time; a trill sign, for instance, has not the same meaning today as it had three hundred years ago. Similarly, the instrument for which such an old piece was written has changed over time. So, for instance Mozart’s and Sciarrino’s piano sonatas are not written for the same instrument, and performance practices have changed. It might be asked whether a valuable performance of a given work is possible without being aware of or deliberately ignoring these conventions and performance practices. However, this article does not aim to pursue this subject, especially since it focuses on a recent composition. Indeed, if it is true that ‘[t]he pursuit of authenticity is not usually controversial … when the musicians are intimately familiar with the conventions and traditions that the composer presupposes’ (Davies 2001: 208), then to perform Alfred Schnittke’s second Piano Sonata would be a quite straightforward matter. Indeed, its score is clearly notated and follows the rules concerning how musical signs should be translated into sound that are taught at the conservatoires. Furthermore, there is no doubt about the instrument for which the composition was written. Nevertheless, to prepare a performance of this piece, or any other similar work, is not as easy as it may seem. Although, on the one hand, a pianist does not need to research what musical signs indicate, on the other hand, ‘many matters that must be settled in generating a performance are not covered by them [by the composer’s indications], being left to the performer’ (Davies 2001: 209). Sound aspects such as attack, articulation, sound quality, and subtle differences in intensity and duration between notes are not clearly indicated by Schnittke’s score. These aspects are extremely important. They are part of the music, they confer to it a specific character and, furthermore, they are crucial in distinguishing a great performance from a simply good one. As Benson (2003) wrote ‘it is precisely what is not to be found in the score that we often most value’ (84—85). Since these aspects are not to be found in the score, they can be realised by a performer in a multitude of different ways without contradicting the written text.

Starting from this observation, the present article raises the following queries: although the score does not give any precise indication about many sound aspects, is it possible to find criteria for performers’ choices? How can these criteria be put into words? Is such a verbalisation useful for efficient practice?

Looking for answers in the score

Many authors seem to suggest that some of the performance instructions not specified by the score can be inferred by an analysis of it. As Eugene Narmour suggests in his 1988 essay, the score, which can be conceived as ‘a kind of syntactical road map based on a highly efficient but therefore limited symbol system’ (318), can be scrutinized in detail, relations between its elements can be pointed out and the analytical findings can be translated into performance instructions. For instance, Narmour argues, by applying his own analytical method to three bars of Strauss’ Der Rosenkavalier, Act II, Sc. 1 that a melodic movement upward should be accompanied by ‘a further reduction in dynamic’, a syncopated B should be played slightly longer than how it is written, a triplet should be sung ‘more or less rhythmically “straight”’, and an A sharp in a triplet would benefit from a slight lengthening (333—34).

In ‘Attacking a Brahms puzzle’, Edward T. Cone (1995) proposes an interesting analysis of Brahms’ Intermezzo op. 116, no. 4. Although, unlike Narmour, he does not make any explicit suggestion to performers (except about a ppmark which, in his view, applies only to one bar instead of to a whole passage, as the notation indicates), he takes for granted that his analysis can have an influence on performance. Thus he writes: ‘[t]he test of the foregoing analysis lies, of course in performance. To a certain extent that is true of any analysis, but it is especially true of one that, in trying to make sense of a problematic work, implies specific recommendations about interpretation’ (1995: 77).

Jonathan Dunsby (1989) considers analysis as an activity that can help performers solve some of the (mostly small-scale) interpretative problems that a score presents. When he employs the term ‘analysis’, he seems to refer to a systematic and rigorous way of scrutinizing the written text of a composition; in other words, musical analysis as it is taught in the conservatoires or is generally practised by musicologists. Analysis, considered in this sense, may not offer an effective solution to all problems deriving from the underdeterminacy of notation. For example, it ‘seems … little able to capture … that secret of the performer — timing — which subsumes so many factors such as rubato, structural articulation and expressive emphasis, and which is such a powerful element in the presentation of almost any composition’ (14). In order to illustrate this shortcoming, Dunsby considers the very beginning of Berg’s Piano Sonata op. 1 and the appearance of the second theme in Brahms’ third Sonata for Piano and Violin op. 108. Both passages present problems to the performer which cannot be solved through rigorous analysis. However, Dunsby identifies other kinds of un-notated considerations that aim to inform a performance. Whilst he wants to demonstrate that ‘the analyst is powerless’ (15) in such cases, he furnishes examples of how such considerations can be employed as a basis for inferring how non-written aspects of the composition can be realised. This paradox has been pointed out by John Rink (1990: 319).

These are all examples of ‘one-to-one mapping’ (Rink 1990: 320): that is, analytical considerations translated into performance instructions. Rink criticises this kind of approach when he writes:

[a]ttempting to recast the findings of analysis into a performance mould seems to me not unlike translating a book into another language word-by-word, without regard to the second language’s particular idioms, inflections, grammar and syntax. A sentence or two might survive such a process, but generally the result would be stilted, contrived and possibly nonsensical (1990: 320).

This does not mean that all performance instructions derived from analysis should be systematically rejected. They can be accepted only to the extent that a performer ‘believes in them, ‘hears’ them … and considers them appropriate in projecting the work’ (1990: 322). Rink also shows how analysts are generally reluctant to recognize that performers can achieve meaningful renditions of a score guided only by their ‘musicality’. Whilst he does not deny the importance of theoretical and analytical knowledge for performers; he does suggest that a good performance is inspired by the combination of a particular kind of analysis, different from the theoretician’s with ‘informed intuition’. The latter is defined as follows:

[W]hat I call ‘informed intuition’ (or ‘acquired intuition’) … accrues with a broad range of experience and … may exploit theoretical and analytical knowledge … This term acknowledges that musicality is probably not innate … but arises through imitation. One plays ‘musically’ when what has been learned through imitation is made one’s own (1990: 324).

In this article of 1990, Rink only outlines the scope of performers’ analyses, which he says represent ‘a considered study of the score with particular attention to contextual functions and means of projecting them’ (323). A performer’s analysis is concerned with musical shape rather than structure, and ‘tends to be more dynamic through its sensitivity to momentum, climax, and ebb and flow, comprising an outline, a general plan, a set of gestures unfolding in time’ (323). Rink acknowledges shared interests between performers’ and theoreticians’ analyses, such as hierarchy. Indeed, like theoreticians, performers establish a hierarchy among musical elements by choosing what to bring out and what to play down, and all possible gradations between these two extremes. Some years later, Rink (2002) redefined his idea of performers’ analyses and proposed some tools that musicians can employ when learning a piece. Techniques such as ‘identifying formal divisions and basic tonal plan’, ‘graphing tempo’, ‘graphing dynamics’, ‘analysing melodic shape and constituent motifs/ideas’, ‘preparing a rhythmic reduction’, and ‘renotating music’ (Rink 2002: 41) are not, however, meant to infer performance instructions. Rather, they help performers ‘to assimilate terminology and concepts which might heighten their ability to articulate to themselves and others … what is happening in the music’ (41). Performers ‘analyse’ firstly through their informed intuition, and only at a later stage, consider how the functions and positions of the elements of a musical work may take place and influence their way of playing.

Relying on my experience as a pianist, I sympathize with Rink’s criticisms of ‘one-to-one mapping’ and with his ideas generally. Nevertheless, I would suggest that his way of proceeding can be counterproductive when applied to piano music of the second half of the twentieth century. Following Rink’s understanding, performers’ informed intuitions are based on their backgrounds: their study experiences, the examples they have heard, the repertoire they have learned, and so on. Whether one likes it or not, pianists’ training is mainly based on eighteenth, nineteenth, and early twentieth century music. Their informed intuition is consequently moulded by a language with a more or less clearly perceivable syntax and tonal harmony. To come to a full understanding of how such informed intuitions react in contact with contemporary works would require psychological research lying outside the aim of this article. However, it can be imagined that to broach a recent composition relying on intuition, and informed by contact with fundamentally different musical styles would not be dissimilar from driving a modern formula-one car relying on our experience of an old runabout: it would be possible, but the potentiality of the new object would be only partially explored.

Looking for answers beyond the score

To return to my previous questions, if the ‘one-to-one mapping approach’ is limited and if intuition is not sufficiently informed to deal with contemporary music, is there another way for defining criteria which can lead performance choices in relation to the aspects of a composition that are underdetermined by the score? I propose to answer this question by adopting another strategy. Instead of trying to make sense — intuitively or deliberately — of what is written in the score, it could be interesting to become aware as much as possible of the several ways a score can be faithfully translated into sound and then try to make sense of what has been discovered. What I am proposing is to explore all possible musical results that are compatible with the composer’s indications and find a way of mapping them. In so doing, it could be easier to find points of references in order to make performance choices.

There are two ways by which one can explore sound results: one is to listen to recordings, the other is to practise the piece. Recordings need not necessarily be by other performers, for artistic researchers can choose to record themselves in order to better and consistently analyse their musical products or their learning process. The distinction between recordings and practice is meant to differentiate between an investigation of the sound result based onlyon listening (recordings) and an inquiry that takes into account a deeper experience of the composition (practice). In both cases, researchers have to face the difficult problem of how to overcome the inadequacy of language. There is no specific lexicon for dealing with such intangible sides of music as timing, sound quality, attack, and articulation. Most of the time performers speak metaphorically and synaesthetically about music, as when musical sounds are defined as clear, soft, warm, heavy, hard, beautiful, ugly, and so on. All these adjectives are imprecise and indicative of a personal judgement when referring to the quality of things that can be appreciated visually or by touch. They are certainly no more precise when employed to define aspects of music. Even if this way of speaking about music has worked successfully and still works among performers, it can hardly be accepted in a scholarly context. Indeed, dissertations or articles like the present one generally require data to be presented in a more rigorous way. It is difficult to imagine an academic text in which the above-mentioned adjectives are employed as the principal tools for sharing knowledge. Artistic researchers who want to focus on intangible aspects of music have, therefore, to find alternative manners of expressing the results of their work. In the past few years a solution to this problem has been found by using software programs that help to analyse many aspects of a recording. For example, the rubato and the dynamics of different performances of the same composition can be graphically represented and easily compared. By simultaneously analysing fluctuations in tempo and dynamics, it is possible to describe the ways in which performers shape a musical phrase. Researchers employing these powerful tools can build up their theories on a solid basis of exact data. This method of inquiry involves recordings as a source of information that can produce results that are relatively easy to disseminate in an academic context.1)Examples of this kind of research can be found on the website of the AHRC Research Centre for the History and Analysis of Recorded Music: .

However, is this the only way of speaking about aspects of music beyond the score? Is a rigorous inquiry in this field possible only by means of technological devices? Or is it possible that performers can research these aspects just by practising and reflecting on what they are doing?

Undoubtedly, data obtained by means of technological devices are easier to manage than reflections on a personal experience. Yet the latter approach has merits which should not be underestimated. Firstly, direct experience can induce performers to reflect in depth on their own practice in order to find aspects of it that can be verbalised in an academically acceptable language. Consequently, they gain awareness and understanding of what they are doing and become able to verbalise their practical knowledge for the benefit of the next generations of performers; in other words, they become better instrumental teachers. Secondly, because performers do not usually practise with an audio analyser device besides their instrument, an inquiry based only on personal reflection would be closer to their everyday concerns.

A case-study: practising the third movement of Alfred Schnittke’s second Piano Sonata

The case-study illustrated in the forthcoming paragraphs shows a solution obtained through practising. In order to avoid any misunderstanding, it is necessary to specify what I mean by the verb ‘to practise’. In this article, I employ it in a usage familiar to performers of Western art music: to do or imagine doing a set of interrelated activities on a musical instrument (the piano in my case) in order to learn to play a composition. Each time I use this verb, I am referring to effort that results in substantial improvement in my way of playing. As can be imagined, such efficient practice is not possible without some kind of reflection. Therefore, the verb ‘to practise’ implies my will to improve my level and a set of activities led by reflection. In this particular inquiry, practice is also nurtured by verbalisation. Indeed, by finding the right words for expressing what goes on in practice, I acquire insights that feed, inspire, and enhance practising. More precisely, the method I apply in this inquiry consists of playing, reflecting, verbalising, reflecting, and playing. The first awareness of the issues the score presents is obtained by playing. Reflection produces a verbalisation of all possible solutions of problems encountered. Further considerations of the verbalised material trigger several small- and large-scale performance choices that are eventually tested by playing. This process may generate an account of the composition’s structure rooted in the individual experience of practising it.

As already said, language is inadequate to convincingly describe sound aspects like, for example, timing, sound quality, attack, and articulation. Therefore, it is not a good idea to define performance choices by describing the results they produce (if these are not reduced in numbers by the appropriate technological devices). If this path cannot be followed, then there should be only one other possibility left: to define performance choices by describing the means by which the aimed result is produced. By ‘means’ I intend nothing but all the movements involved in my playing the composition.

Figure 1: Alfred Schnittke: Piano Sonata No. 2, 3rd movement; ‘Allegro moderato’, bars 1-4

Figure 1 Alfred Schnittke: Piano Sonata No. 2, 3rd movement; ‘Allegro moderato’, bars 1-4

Consider, for example, the first four bars of the third movement of Schnittke’s second Piano Sonata (Figure 1). They can be performed in several ways, such as a succession of brilliant non-legato sounds, an expressive legato melody, an obsessive hammering, and so on. As can be easily imagined, such descriptions, although frequently used by performers, are too imprecise to be employed in a consistent investigation of the musical aspects to which they relate. Therefore, it is necessary to find another manner of referring to the musical results a pianist aims to produce. As suggested above, a valuable possibility could be to describe the means by which what one aims at can be achieved. Indeed, since there is a link between the sound result one wishes to obtain and the sort of movements one has to perform to that end, a description of the latter could be a valuable way of referring to the former. For example, the notes of these bars are more likely to sound non-legato and brilliant if they are played by holding arms and wrists still and by employing the articulation of the fingers (I call this way of playing Movement A; shown in Video 1). If one aims at a more legato sonority and with subtle differentiation between the intensities of notes, it is preferable to hold the fingers in contact with the keyboard and to move the wrist in order to employ and control the weight of the body (Movement B; Video 2). If, instead, the intention is to give a stronger impulse to each note, it is better to press each key by means of a vertical movement of the forearm always holding the fingers in contact with the keyboard (Movement C; Video 3). All these three possibilities are compatible with the indication ‘Allegro moderato’, even though the first two options allow a faster tempo than Movement C.

The third option, i.e. to press each key by means of a vertical movement of the forearm always holding the fingers in contact with the keyboard, is the only one that allows good performance of bar 5 (Figure 2). Otherwise, this sequence of clusters can be performed by holding the wrist still and by moving the forearm up and down (Movement D; Video 4), or by moving the hand up and down and holding the forearm still (Movement E; Video 5). The first four bars could also be played by Movements D and E, but the tempo at which this piece should be performed (even at its acceptably slowest rate), these movements are quite awkward and the musical result cannot be easily controlled.

Figure 2: Alfred Schnittke: Piano Sonata No. 2, 3rd movement; ‘Allegro moderato’, bar 5

Figure 2 Alfred Schnittke: Piano Sonata No. 2, 3rd movement; ‘Allegro moderato’, bar 5

These are five kinds of bodily movements by which the whole third movement of this sonata can be played. They have to be considered as nothing more than means for realising musical ideas and their descriptions are only ways of indirectly referring to otherwise indescribable aspects of the music.

Mostly, both hands have the same possibilities in the same passages: There are only eleven bars in which the musical material does not allow for one hand performing some kinds of movements in similar ways to the other. For example, in bar 7 (Figure 3), while the left hand can play by means of all movements, the right hand can employ only Movements C, D, and E: which makes it impossible for the right hand to perform a succession of repeated four-note chords by Movement B and extremely awkward to employ Movement A.

Figure 3: Alfred Schnittke: Piano Sonata No. 2, 3rd movement; ‘Allegro moderato’, bar 7

Figure 3 Alfred Schnittke: Piano Sonata No. 2, 3rd movement; ‘Allegro moderato’, bar 7

Each kind of movement I have described does not univocally correspond to a musical result. Instead, it allows for certain sorts of sounds, making it hard or even impossible to produce others. For instance, by employing the articulation of the fingers (Movement A), I am more likely to produce clear, brilliant and distinct sounds rather than an enveloping legato. To be precise, I could also obtain a legato also by playing with a vertical movement of the fingers, but this would require a lot of practice and give a less than satisfactory result. To obtain a convincing legato, I prefer to hold the fingers in contact with the keyboard and move the wrist in order to shift the weight of the body from one key to the other (Movement B). Yet, this kind of movement alone does not guarantee the quality of the final result. Once I have decided which kind of bodily movement to employ for playing a given passage, I need to practice the definition of each detail. The ways in which this happens are very hard, if not impossible, to describe in words. The subtlest variations in sound quality, attack and articulation are the consequence of minute differences in the position of hands and fingers, and of the weight applied to the keyboard, the point of contact of the fingertips, and so on. The same can also be said about dynamics. Although it is true that some movements make it easier to play louder rather than softer (for example, Movement D), the intensity with which sounds are played depends on actions of different muscles that are too complex to be efficiently described. It is therefore difficult to imagine how a performer could be aware of or reflect on every slight detail of the movements he or she executes. Rather, my experience as a professional pianist makes me think that, when practising, the subtlest aspects of my bodily movements are settled without focussing on them. The idea one bears in mind about how a composition should sound functions as a touchstone against which the sounds actually produced are compared. If these do not match the ideal sound image, one plays the passage in question again and again until all necessary adjustments of bodily movements happen automatically.

At this stage of practice, the distinction between my body and the instrument I am playing seems to disappear. This phenomenon is very similar to that described by Merleau-Ponty (1963) apropos of the way an organist learns to play an instrument he or she has never touched before:

… between the musical essence of the piece as it is shown in the score and the notes which actually sound round the organ, so direct a relation is established that the organist’s body and his instrument are merely the medium of this relationship (Merleau-Ponty 1963: 168).

If one accepts this view, one could also imagine that when pianists practise the subtlest aspects of a given passage, they establish through listening a direct relation between the score’s indications and the musical results they produce. If experienced pianists wish to modify something in their playing, they stay focused on the musical result and their habits enable something in their bodily movements to change without them being aware of it. Therefore, when I illustrated the movements employed when playing the first five bars of the third movement of Schnittke’s second Piano Sonata, I did not give a detailed account of them. Rather, I described their most salient aspects. For example, I wrote that the first four bars can be played by holding arms and wrists still and by employing the articulation of the fingers, but I did not specify how high the fingers should be raised, how fast they should be dropped and whether all notes should be played in the same way. These are details that can be settled only by repeating the passage over and over again. Nonetheless, what I have described is a valuable point of reference for better determining the performance choices at my disposal. Employing a particular movement for playing a given sequence of notes makes it easier to obtain certain results whilst inhibiting other kinds of attack, colours, or articulations. Therefore, if a specific bodily movement does not correspond with a precise sonic outcome, it can nonetheless define a range of possibilities one of which will be the actual musical result. Furthermore, these movements can be easily described in words, and thus be verified by anyone with sufficient competence.


Figure 4: Alfred Schnittke: Piano Sonata No. 2, 3rd movement; ‘Allegro moderato’, bars 1-3 and 17-19

Figure 4 Alfred Schnittke: Piano Sonata No. 2, 3rd movement; ‘Allegro moderato’, bars 1-3 and 17-19

However, to be aware of all the kinds of movements that can be employed for any one passage is only a starting point. In order to define performance choices efficiently, closer examination of the musical work is needed. By practising the composition, it is relatively easy to notice passages that, although similar in some important musical aspects, present different technical problems. A clear example of this is the passage from bar 17 to 19 in relation to the first three bars (Figure 4). All these bars start with the same motif in the right hand immediately repeated by the left (mirrored in bars 17—19). Melodically, these two passages have similar material. They can be considered as two sequences in which the same motif is repeated three times each at a higher pitch. However, while in the first three bars and in bars 17 and 18 both hands play a succession of single notes, in bar 19 they play a series of two-note chords. Consequently, the most appropriate movements they can employ differ from the one passage to the other. All bars except bar 19 can be played mainly by Movements A, B, or C (as I wrote above, Movements D and E are possible but awkward at the tempo of this movement). Both are also uncomfortable in bar 19, but here, they are preferable to Movements A and B. Indeed, by holding the arm and wrist still and moving only the fingers up and down (Movement A), it would not be easy to achieve perfect control of all the notes of this passage; some would tend to be slightly longer than others (especially those played by the thumb and the forefinger). Moreover, Movement B is impossible as such. Because the disposition of the notes does not allow for perfect legato, this movement should be wisely combined with a vertical movement of the arm (Movement D; ‘combined’ since the wrist should move horizontally in Movement B and vertically in Movement D, ‘wisely’ because Movement B can easily lose its peculiarities when combined with Movement D). Therefore in these bars, due to the disposition of the notes, there are different kinds of movement which seem to be most appropriate and give better results (Movements A, B, and C for bars 1—3 and 17—18, and Movements C, D, and E for bar 19) and others that, although possible, require effort in order to respect the musical text (Movements D and E for bars 1—3 and 17—18, and Movements A and B for bar 19). As a performer I can choose. Although all movements are compatible with what is indicated in the score, my choice will determine the shape I give to the musical phrase. If, for instance, I chose to play each passage by performing the most comfortable movements, I would play two similar sequences in two different ways. In bars 1—3 (which I would play by employing Movement B), the melody would sound legato and each note would have a rounded attack. The same would happen in bars 17—18, but the sequence bars 17—19 would differ from that of the beginning since the notes in bar 19 (when played by Movement E) would be separated rather than legato and have a more brilliant and distinct attack (Video 6).

At this point, it is important to stress again that the differences I am writing about concern aspects of the music that are not determined by the score. These passages would undoubtedly be described differently by formal analysis. Although in both cases the same motif is repeated three times, the pitches are different, the ways in which the left hand imitates the right hand are different and, above all, bar 19 presents four voices instead of two and, after the appearance of the motif, continues differently from the other bars. The differences that can be pointed out through an analysis of the score are qualitatively distinct from those that concern me here. Indeed, while the former are given by the score and, obviously, cannot be changed or erased, the latter depend on the way a pianist performs the composition. If, for instance, I played these passages in the way described above, I would confer a particular structural importance to bar 19. The motif would have for the first time another colour, another articulation, and consequently would sound different than all previous instances of the same. In this case, I would consider the passage from bar 17 to 19 as a reappearance of bar 1 to 3 with an important swerve at the end (bar 19). However, I could also make other performance choices. I could play bars 1—3 by still employing Movement B, but play bars 17—19 by using only Movement C (Video 7). In so doing, I would consider the whole passage from bar 17 to 19 as a repetition of bars 1 to 3, though with a completely different character. Then again, I could make an effort in bar 19 to play both passages employing only Movement B. In this case, I would emphasise the similarities and mask the differences between them. Another way of stressing their commonalities would be to play bars 1, 2, 17, and 18 by means of Movement B and bars 3 and 19 by Movement C. Therefore, several combinations are possible, each one influencing my interpretation. Since similar passages recur at many places in the composition, performance choices can be made on the basis of large-scale considerations. Each passage in which the main motif appears may be played by using a different kind of movement aiming at different sound results, or the gestural variations could be distributed throughout the piece following certain criteria.

Also the positions of passages in which the main motif does not appear or which present other important musical material can be taken into account in order to choose the sort of movement to employ, and thus the kind of sound aimed at. Consider, for instance, bar 5 (see Figure 2), which is completely different from the first four bars (see Figure 1). This succession of clusters has nothing in common with the preceding bars, or at least, nothing that is indicated in the score. However, by focussing on those aspects of the musical work that cannot be notated, it is possible to find, or more exactly, to create unexpected relations. Imagine playing the first four bars by employing Movement B and bar 5 by means of Movement D. This would be a way of emphasising the contrast between four polyphonic bars (bars 1—4) and the violent succession of clusters in the following bar. Nevertheless, it is also possible to perform all notes of these five bars by moving the hand up and down whilst and holding the arm still (by Movement E, even if, as I said before, this is not the easiest option for the first four bars). In so doing, bar 5 would be integrated in the preceding passage and would sound as a continuation of it. This example illustrates how relations can be created between elements that are completely different on paper and reveals the great freedom a pianist has to shape the musical unfolding, even by faithfully following the indications of the score.

Figure 5: Alfred Schnittke’s Piano Sonata No. 2, 3rd movement; ‘Allegro moderato’, bars 1-19

Figure 5 Alfred Schnittke’s Piano Sonata No. 2, 3rd movement; ‘Allegro moderato’, bars 1-19

If I write down the movements that can be employed in each passage, I obtain a ‘map’ of the piece by which I can more easily determine my path. This path, which is to say my choices about the bodily movements perform, can also be described by means of words. Since, as written above, there is some relation between the movement I employ and its musical result, a verbalisation of my performance choices can stand as a particular description of the shape of the composition. Imagine I decided to perform the first 19 bars as in Figure 5, i.e. by playing all bars as more or less melodically related to the first four bars by employing Movement B, and then playing the remaining 6 bars by means of the contrasting Movement E. In this case, the description of this passage would identify a main musical stream (that characterised by the use of Movement B) interrupted by ‘windows’ of contrasting elements. Such performance choices would produce an account that would probably share commonalities with a desk analysis of these bars. For instance, differences in kind of employed movements coincide with differences in melodic and rhythmic elements, which can be easily identified in the score.

However, I could also choose to employ the same movement in order to play contrasting elements. For example, I could play bar 1 to 8 by using Movement C. In this way, I would give the notes in bar 5, 7, and 8 a kind of attack, articulation and sound quality similar to the more melodic bars 1 to 4 and 6. Further on, I could play bars 9—14 by employing Movement B and bars 15—20 by means of Movement C again. These performance choices would point up a shape comprising three sections (bars 1—8, 9—14, and 15—20), the third section beginning similarly to how the first one ends. What holds the elements of a section together is not written in the score but, with more or less subtlety, is determined by my way of playing.

As is easy to imagine, such kind of descriptions can be made of the whole movement. They generate from a particular perspective, that of the performer. In their own way, they take into account the elements notated in the score, the bodily movements necessary to play the composition and the musical result. Obviously, I do not pretend that all performers employ the same movements. There are many different ways of playing piano. What I have described is just my technique, as I have developed it during many years of practice and by applying the suggestions of my teachers. Certainly, I am not the only one who plays in this way and what I write is not new. Furthermore, I do not claim that my way of playing is the only one through which is it possible to achieve a valuable performance. The aim of this article is not to propose a piano technique but, rather, an approach to an issue. At the beginning I formulated questions, acknowledged a difficulty in answering them and proposed a method by means of which to overcome this difficulty. The method I have proposed can be employed by other performers even if, since their way of playing can be different from mine, their descriptions of the movements employed can be different. However, by applying my approach, they may also describe the means by which these sounds are produced.


In this article, I have given an example of how insights can be gained through practising. That this activity is a way of accessing some kind of knowledge seems self-evident, at least to music practitioners. It should not be difficult to accept that by practising a given composition, one not only learns to play it but can also form an opinion about its structure, character, emotional content, and so on. However, it is not so easy to present such kind of knowledge in a way that goes beyond personal opinion, that is to distil from it information which can be employed by others, trigger discussions or feed ongoing scholarly debates. Not all insights gained through practising can be put or are worth putting into words. Furthermore, not all ways of verbalising a given insight function in any context. Everyday musicians’ talk about the issue considered in this article, the sonic aspects of a musical work underdetermined by the score, mainly focus on the result, described through adjectives and metaphors. If, on the one hand, this kind of verbalisation is traditionally accepted among practitioners as a way of sharing thoughts, on the other hand, it can become a target for scorn if employed in a scholarly article. Adjectives and metaphors related to sound cannot be verified, nor argued or supported and, therefore, cannot be the subject of constructive discussion. The alternative I propose is to describe the means necessary to produce a particular musical result. These means take the form of sequences of bodily movements that can be reproduced, and therefore verified, by anybody possessing the necessary instrumental skills. By becoming aware of, defining and describing all the movements that can be employed in order to play a given passage, I can draw up a useful compositional map to assist performance choices. Each choice has an influence on the shape I give to the musical unfolding. By employing the same movement for playing different musical elements, I can emphasise their similarities or show unexpected relations between them. The other way round, performing similar passages by means of different movements, sheds diverse lights on analogous melodic and/or rhythmic elements. The relation between employed movement and obtained sound is not one-to-one. To choose a particular kind of bodily movement does not imply a particular result. Once the movement employed in order to obtaining a particular sound has been decided, I have to find ways to articulate all the musical minutiae. The process by which this happens involves so many subtle actions, mostly habitual, that are almost impossible to describe in words. However, a categorization of the sorts of movement which can be employed in order to play a given passage remains an efficient way for defining performance choices relatively to sound aspects underdetermined by the score. Indeed, each kind of movement, even broadly considered, has always an influence on the sound result since it makes easy to produce certain sorts of articulations, attacks and timbres, and difficult, if not impossible, to generate other ones. The choice of the movements to employ can be inspired by small- as well as large-scale considerations. The sequence of bodily movements chosen for playing a composition can then be investigated in order to generate a description of the musical work’s shape from a performer’s perspective.

The investigation I illustrate starts from precise questions and is conducted by practising. Since craftsmanship plays a crucial role in musical practice, it necessarily occupies a prominent place in my account. However, I hope my detailed descriptions of elements of piano technique do not overshadow what I consider to be the core of this article: namely, the way in which the inadequacy of language for talking about certain intangible aspects of music can be overcome, and, consequently, questions important to a performer can be consistently answered. In this article, I show how practising can lead to results others than a performance of a composition: results which can be written down, read, and discussed. I hope to induce other musicians to develop their techniques for verbalizing the insights they get by practising. Only the collective efforts of many performers and composers may confer to musical practice a dignity equal to that of more established academic disciplines.


Benson, Bruce E.. 2003. The Improvisation of Musical Dialogue: A Phenomenology of Music (Cambridge: Cambridge University Press)
Cone, Edward T.. 1995. ‘Attacking a Brahms puzzle’ in The Musical Times, 136 (1824), pp. 72-77
Davies, Stephen. 2001. Musical Works and Performances: A Philosophical Exploration (Oxford: Clarendon Press)
Dunsby, Jonathan. 1989. ‘Performance and analysis of music’ in Music Analysis, 8/1-2, pp. 5-20< Merleau-Ponty, Maurice. 1962. Phenomenology of perception [translation of Phénoménologie de la perception. Paris, 1945] (London: Routledge and Kegan Paul Ltd.)
Narmour, Eugene. 1988. ‘On the relationship of analytical theory to performance and interpretation’ in Explorations in music, the arts, and ideas. Essays in honor of Leonard B. Meyer, ed. by E. Narmour and R. A. Solie (Stuyvesant, N.Y.: Pendragon Press), pp. 317-340
Rink, John. 1990. Review of Wallace Berry Musical structure and performance in Music Analysis, 9/3, pp. 319-39
Rink, John. 2002. ‘Analysis and (or?) performance’ in Musical Performance: A Guide to Understanding ed. by J. Rink (Cambridge: Cambridge University Press), pp. 35-58

Music examples

Alfred Schnittke: Sonata No. 2. for piano (SIK1876). All music examples have been reproduced with kind permission from © MUSIKVERLAG HANS SIKORSKI, Hamburg



1 Examples of this kind of research can be found on the website of the AHRC Research Centre for the History and Analysis of Recorded Music: .