MUSICAL VERSIONS AND TRANSCRIPTIONS
Disputing Stephen Davies’ Account
University of Granada
Abstract According to Stephen Davies, versions and transcriptions of musical works have different ontological weight. In his view, transcriptions are new musical works derived from the original, whereas versions are not, and are taken to be the same work as the original. Therefore, transcriptions are ontologically heavier than versions because they count as new entities. Davies holds that transcriptions involve a change of instrumental medium, while versions do not. For Davies, when the medium of a medium-specific work is changed, the result is a new work that is different from the original. In this paper, this difference in ontological weight is rejected, arguing that neither versions nor transcriptions are new musical works. Although Davies’ general account is grounded on strong intuitions concerning musical practices, it will be shown that his view on this point at least is not consistent with the robust intuition that transcribers are not composing new musical works when they transcribe. In first place, we will reject Davies’ distinction in ontological weight between versions and transcriptions as a revisionist view regarding our musical practices. Then we will contend that, in Davies’ account, there are no additional reasons strong enough to support the revisionist idea that a change of instrumental medium is sufficient to generate a new musical work.
1 Davies’ Account of Versions and Transcriptions
Versions and transcriptions of musical works can be undertaken either by the work’s composer or by someone else. For instance, on the one hand, Sibelius modified his Fifth Symphony twice—in 1916 and 1919—after the premiere of the work, seeking to improve the previous versions. On the other hand, Karl Marguerre is said to have penned a performing version of Mozart’s Horn Concerto in D K412/514 from the original manuscript, which was incomplete. In this paper, due to the peculiarities involved, we will put aside this second kind of version. Thus, we will focus only on versions produced by the work’s composer and on transcriptions produced either by the work’s composer or by another person.
Stephen Davies has devoted a significant part of his work in the philosophy of music to the study of versions and transcriptions. On Davies’ account, a common feature of versions and transcriptions is that both can take place only after the completion of the work of which they are versions or transcriptions. In Davies’ words:
“A musician might begin to compose, using a finished work as her source. If she carries the process of re-composition far enough, she writes a new piece. The new work is influenced by the original, and perhaps audible traces of this inspiration remain detectable in quotations or allusions. In a different scenario, the composer does not carry the process very far and she conceives of herself as revising the source rather than going beyond it. The product is what I have called a work version. The practice of transcription lies between these extremes. The audible relation with the original is preserved, as is the sound-structural outline and much else, yet the change in instrumental medium distances the transcription from its model, with the result that a new work is produced.”
Accordingly, versions and transcriptions are musical products that require as a necessary condition that a musical work has been finished. Versions are revisions of a previous work without departing so substantially from it that they constitute a new musical work. By contrast, what is characteristic of transcriptions is, roughly, that they translate a previous work into a different musical medium. For instance, when the original piece is written for orchestra, its medium is the orchestral instruments. Changing the medium of this piece would be transcribing it, for example, for piano, organ, or voices. In contrast to versions, transcriptions are considered by Davies to be new musical works. Thus, in Davies’ view, versions and transcriptions have different ontological weight: the latter but not the former constitute new musical works distinct from the original.
According to Davies, versions involve changes in the constitutive properties of the work. When revising a work in order to produce a new version, the composer can alter the sound structure of the original piece, changing notes, adding new parts and deleting others. However, for Davies these changes in the sound structure are not sufficient to give rise to a new work. On the other hand, transcriptions do not necessarily involve changes in the sound structure. Transcriptions, as Davies points out, aim to preserve “the audible relation with the original” and “the sound-structural outline”. In making a transcription, modifications in the sound structure are required only when the instruments of the new medium are not technically able to play the original sound structure. However, as indicated in the passage above, the change of medium necessarily involved in transcriptions is nonetheless sufficient for a transcription of a medium-specific work to count as a new work. Following Davies’ view, there may be cases in which differences in sound structures are not enough to give rise to new musical works—the case of versions¾, while there may be others—works whose medium has been specified by the composer—in which a change of medium is sufficient to obtain a new musical work. This is the case of transcriptions of medium-specific works.
In this paper, I argue against the idea that transcriptions are ontologically heavier than versions. There are two ways of rejecting this idea. It could be argued either that both a version and a transcription are new musical works different from the original, or that neither constitutes a new work at all. It is my contention that the second view is correct, since it does more justice to musical practices. That is, I agree with Davies that versions are not new musical works different from the original, but not with his claim that transcriptions constitute new works. As I will show in the second section, this is more consistent with our intuitions regarding musical practices. In the third section, I will discuss, and ultimately reject, Davies’ reasons for regarding transcriptions as new musical works different from the original.
2 Davies’ Revisionism
The thesis defended in this paper, namely that transcriptions are not new musical works different from the original, is grounded on the musical practices involved in transcriptions. In this section, Davies’ hypothesis that transcriptions really constitute new musical works will be assessed. By means of two thought experiments, some pre-theoretical intuitions grounded in our actual musical practices will be highlighted to show why considering transcriptions to be new musical works is problematic, even when the original works are medium-specific. Afterwards, real examples of transcriptions of medium-specific works will be offered to show that Davies’ view is not consistent with musical practices.
It is common practice for singers to rehearse operas from Monteverdi to Puccini with a piano transcription of the orchestral part. Consider Verdi’s Nabucco, an opera from the 19th century that is medium-specific. Let us imagine that Lucia, a soprano, and Hannah, a pianist, are rehearsing in a room. Suppose that Maria enters the room and says: “Lucia, that sounds like the aria ‘Anch’io dischiuso un giorno’ of Verdi’s Nabucco—is it?” What should Lucia’s answer be? If we interpret transcriptions as new musical works differing from the original, Lucia’s answer should be that it is the aria ‘Anch’io dischiuso un giorno’, although not of Verdi’s Nabucco but of Mario Parenti’s Nabucco (Mario Parenti is the author of the transcription for piano). However, this kind of answer would be inconsistent with our intuitions regarding musical practice. Maria knows of only one opera called Nabucco, which was composed by Verdi. It would be surprising for her to learn of another work with the same name, with an aria that has the same melody, rhythms, and harmony, composed by a person other than Verdi. Indeed, notwithstanding the creativity and the interpretive work involved in writing the transcription, no informed person would credit Mario Parenti with composing a musical work for piano and voices called Nabucco.
A second example showing how counterintuitive it is to take transcriptions as new musical works is the following. Imagine that the London Contemporary Music Festival commissions the Argentinian composer Jorge Valdano to compose a new work for a brass quintet for that year’s festival. He is told that the piece should last no more than 10 minutes and that it should evoke the Argentinian musical style. To fulfil the assignment, Valdano decides to make a transcription for brass quintet of Astor Piazzolla’s work Primavera Porteña. Being convinced that a change of medium is enough to give rise to a new musical work, he changes the medium—from bandoneon, violin, bass, piano, and electric guitar—to two trumpets, a French horn, a trombone, and a tuba. He also readjusts the original melodies to the idioms of the new instruments, avoiding certain ornaments and modifying musical articulations that are especially difficult for brass instruments (i.e. Valdano demonstrates creative and interpretative skill). As Valdano is convinced that his transcription is truly a new work, he decides to give it a different name from the original: Bonaerensis. Imagine now that, at the premiere, the original work is performed—Piazzola’s Primavera Porteña—followed by Bonaerensis. How would the audience react? Most of the audience, I suggest, and especially those who bought their tickets expecting the premiere of a new work, would feel cheated. In sharing most of the sound structure, the harmony, and the movement of the voices, Bonaerensis and Primavera Porteña are arguably too close to be felt to count as distinct musical works: to re-present an original work doesn’t seem to be to present a new work, but to present the old work in a different way. It is true that the audience might think that Bonaerensis is an ironical work satirizing the concept of creating a musical work. But even in this case, Bonaerensis wouldn’t, I submit, be considered a new musical work, but a new conceptual work of art. In sum, the idea that a transcription of a musical work is a new musical work is counterintuitive, even in cases where the musician that makes the transcription believes otherwise.
These two hypothetical examples illustrate that it is counterintuitive to take transcriptions as new musical works different from the original, even when the original piece is medium-specific. It is more intuitive to regard transcriptions of a work W as new modes of presentation of W than as constituting new different works (although derived) from W. As a result, Davies’ account should be regarded as a revisionist view of transcriptions.
However, Davies’ thesis is not only revisionist concerning our pre-theoretical intuitions involved in musical practices, it is also out of step with our practices of referring to and cataloguing transcriptions. A paradigmatic real case of this phenomenon is Mussorgsky’s Pictures at an Exhibition. The piece, finished in 1874, was originally written for piano and, according to the context of its composition, is a medium-specific work. However, the piece figures among the orchestral repertory due to several transcriptions that have been made: Touchmalov (1815), Wood (1915), Funtek (1922), Leonardi (1924), Ashkenazy (1982), Wilbrandt (1992), and the most famous one by Ravel (1922). When Ravel’s transcription is performed, programmed, or recorded, what it is said to be performed, announced, or recorded is a work by Mussorgsky, not by Ravel. The same phenomenon arises for the other transcriptions of this work. Many other examples can be cited—for instance, Franz Liszt’s transcriptions for organ of Chopin’s piano preludes no. 4 and 9 op. 28, or Debussy’s transcriptions for piano of his own orchestral works La Mer and Prélude à l’après midi d’un faune—in which, despite the original works being medium-specific, the results of these transcriptions were not considered to be new compositions by Liszt or Debussy. Therefore, the thesis that transcriptions are new musical works different from the original is a revisionist one, and it should be rejected if we want to respect our musical practices and pre-theoretical intuitions.
3 Discussing Davies’ Reasons
Despite the conclusion of the last section, we might not be pure descriptivist philosophers, taking musical practices as having the last word in our ontology. Indeed, our actual musical practices and intuitions could be wholly mistaken, or at least require revision in light of other considerations. Consequently, the contention that transcriptions are new musical works should not be rejected exclusively on the basis of apparently fundamental intuitions concerning this topic. In fact, Davies doesn’t intend his distinction to merely capture our intuitions, but offers other reasons for his distinction. In this section, we will consider Davies’ main reasons for holding this view in order to see whether they are strong enough to justify such a revisionist account concerning transcriptions.
Davies’ starting point is that a change of medium is enough to give rise to a new musical work, for medium-specific musical works. Davies claims:
“Usually, a change in medium involves a change in instrumentation (and note changes consequent on this). It is possible to produce a new piece through a change in instrumentation, because most musical works are medium-specific.”
“Where a medium-specific piece is adapted to a new medium the result is a distinct work.”
Therefore, a crucial feature that determines the identity of a medium-specific work seems to be the instrumental medium for which the work was written by its composer. According to Davies, transcriptions involve a change of medium from the original work. If the original work is medium-specific, such a change is sufficient to consider the product a new musical work. Three reasons for this claim can be found in Davies’ work: (1) certain aesthetic properties essential to musical works depend on the specific medium in which they are performed; (2) in the case of medium-specific works, the context in which they were composed determines that the medium specified by their composer is essential to them; and (3) colour (or timbre) is a necessary condition of the structure and content of a musical work. In what follows, we will discuss these three reasons in more depth.
The first reason says that certain aesthetic properties essential to musical works are dependent upon the specific medium in which they are performed. In this regard, Davies claims that “recognizing what is achieved in the work involves consideration of the constraints its media impose”. Davies takes ‘virtuosic’ as an example of those essential properties of musical works that depend on the medium prescribed by the composer. Being virtuosic is an essential property of Beethoven’s Sonata because “it is a characteristic of the work that it is technically demanding of the pianist who would perform it”. The property is essential to the work, and whether the work bears the property depends on the medium; therefore, the medium is essential to the work.
However, the first premise can be rejected on the basis that ‘virtuosic’ is not in sensu stricto a property predicated of musical works. Not all aesthetic predicates are applicable both to musical works and their performances in the same way. There are aesthetic predicates, such as ‘atonal’, ‘dissonant’, or ‘melancholic’, that are applicable only to musical works. For instance, we do not say that a performance of Pierrot Lunaire is atonal. Rather, the work itself is atonal, and a performance’s atonality derives from the work. Thus, the atonality of a performance necessarily depends upon the work performed: only if the work is atonal can the performance also be so. In contrast, we commonly apply the adjective ‘virtuosic’ to performances, but it sounds strange to claim that a musical work is virtuosic per se. The musical work possesses the property of being virtuosic in a derivative way from its performances. First, we would not say that a piece is virtuosic if an appropriate rendition of it is not virtuosic. Moreover, pieces that are usually not considered technically demanding of the performer can be performed in a technically challenging way. For instance, consider a hornist playing Franz Strauss’ Nocturno lying on the ground. While the work is not particularly demanding for the performer, the performance can be judged as virtuosic because of the difficulty of maintaining the air column while playing with the body in a horizontal position. Furthermore, there are pieces for which the appropriate renditions were technically challenging to the performer in the beginning, but have become much easier to perform due to technical advances in the instruments. When Beethoven wrote his Sonata for Horn and Piano, true intonation and some fast passages of this piece were very difficult to achieve on the natural horn. When valves and cylinders were implemented on the French horn, the Sonata became much easier to perform in an accurate way. However, although these technical improvements made appropriate performances of the Sonata less technically challenging, the identity of the piece remained the same. Therefore, being virtuosic is not an essential property of the Sonata because the piece can lose this property without altering its identity. The same phenomenon identified in the case of ‘virtuosic’ can be extended to other aesthetic adjectives, such as ‘clean’, ‘blurry’ or ‘accurate’, that primarily apply to performances and only by extension to musical works. Accordingly, properties such as virtuosic that depend on the specific medium in which the work is performed could be counted as essential properties of performances, but not of the work itself.
In empirical support for the thesis that ‘virtuosic’ is an adjective that properly applies to performances, and only in a derivative way to musical works, we can cite research on uses of ‘virtuosic’ in English (http://corpus.byu.edu/coca/ ). We found 205 occurrences of ‘virtuosic’, 195 of them concerning different kinds of arts: 4 for dance, 1 for architecture, 3 for painting, 2 for photography, 2 for sculpture, 7 for cinema, 8 for theatre, 17 for literature, and 151 for music. Among these 151 occurrences in musical contexts, it is significant that only 15 of them are not applied to performances, and 30 of them are applied to performance only in an indirect sense. Therefore, the vast majority of occurrences of the word ‘virtuosic’ are applications to performances in a direct sense: 106 occurrences. This tendency of English speakers supports our thesis that virtuosity is a property that primarily belongs to performances, and only in a derivative way to musical works.
Consequently, in light of the reasons adduced above and of the empirical evidence offered, ‘virtuosic’ is an aesthetic property that cannot play a role in determining the identity of a work. By contrast, Davies (2008), Levinson (2011), and Kivy (1988) take ‘virtuosic’ to be an adjective predicating a property of musical works. For instance, discussing Kivy’s arguments against Levinson’s instrumentalism, Davies claims:
“To say a work is original or virtuosic is to say something about the properties it has when played correctly […]. Surely it is false to assume modern listeners are incapable of recognizing the virtuosic character of Nicolai Paganini’s music when they see and hear it played on a violin.”
The point is that ‘virtuosity’ is a normative property of a work W that must be satisfied by W’s performances in order for them to be performances of W. However, according to Davies, virtuosity has nothing to do with the number of notes, the register, or tempo, but with the difficulty of playing certain passages on a specific instrument. Therefore, in medium-specific pieces, it would seem that “the virtuosic character of a piece rests on the difficulties of playing it correctly on the specified instruments”. However, it is not true that the difficulty of a performance rests exclusively on the instrument played. It also depends on the abilities and skills of the instrumentalist, that is, the particular person who plays the instrument on a particular occasion. Consider Schumann’s work Adagio und Allegro for horn and piano. Let us imagine two horn players: Sarah, who has a special gift for playing fast notes and for playing notes in all registers, and Elizabeth, who finds it difficult to play high notes and very fast arpeggios and scales. In a concert, both play Adagio und Allegro perfectly, without missing a single note. However, if virtuosity were a normative property of the work, it would be a property that the work has when it is played correctly, so that any correct performance of the work should possess this property. Since Sarah plays Adagio und Allegro without effort due to her abilities, Sarah’s performance would not be a correct performance of the work, which seems to be counterintuitive. These kinds of problems do not arise if we think that virtuosity is a property of performances and that any work can have virtuosic and non-virtuosic performances. Therefore, properties such as virtuosity, dependent on the instrumental medium, are not essential properties of musical works, but of performances. Consequently, a change of medium does not determine a change in the work’s identity, resulting in a new work. In sum, Davies’ argument from aesthetic properties in favour of the thesis that a change of medium is enough to produce a new work, if the original is a medium-specific one, seems to be the following:
However, since virtuosity is not a property of musical works, there is no pressure to accept the conclusion of the argument. It is noteworthy that this is only one case –namely, virtuosity –, so Davies might be able to run the above argument with other properties successfully. Nonetheless, as no other such properties are forthcoming, the burden rests on Davies to come up with them.
Next, Davies’ second reason for the arguing that a change of medium is a sufficient condition for obtaining a new work will be considered. It has to do with the role played in the individuation of musical works by the musico-historical context in which they are composed. According to Davies, the features relevant for a work’s identity are determined by the musico-social context in which the work is composed. The conventions governing musical practices in the context of the composition of a work determine the meaning of the composer’s notations in the score, and hence the properties that are normative for performances of that work. Consequently, it is the context of composition that determines whether instrumentation is part of the identity of a given piece. If the conventions implicated in the context in which W is composed determine that the medium is essential to W, since transcriptions always involve a change of medium, a transcription of W should be counted as a new musical work different from W. However, Davies does not think the conventions always determine that instrumentation is essential to the identity of a work. Hence, he rejects an absolute instrumentalist account for the individuation of musical works; that is, he rejects the thesis that a properly formed performance of a work must not only exhibit the timbral qualities intended by the composer, but also be performed by the instruments prescribed by the composer, regardless of the conventions involved in the context of composition. According to Davies, musical works began to be more and more medium-specific from the 18th century, and he claims that it is only since the 19th century that instrumentation (colour) has been mandated by composers. In short, instrumentalism only holds in those contexts in which conventions and musical practices mean that the composer’s indication of the work’s performance-means are to be taken as directions concerning the identity of the piece. The practices of the Renaissance do not determine that the instruments prescribed by the composer are essential to her work, for example. Gabrieli’s Canzona per sonare no. 1 performed either by an organ or by a brass ensemble would count as a properly formed instance of that piece. By contrast, according to Davies, the relevant practices in Tchaikovsky’s context exclude any performance not played by the instruments mandated by Tchaikovsky as a properly formed instance of his Second Symphony. The latter is a context in which the work’s instrumentation is taken to be mandated by the work’s composer. Pieces composed in these contexts are thus medium-specific works.
We have seen that, despite the claim that the context of composition settles the properties that belong to the work’s identity, it does not follow from this that a change in instrumentation results in a new work. Contrary to what Davies thinks, when a transcription of a medium-specific work is made, sometimes the result is not a new musical work. Even in 19th-century contexts, where practices determined that being performed by the instruments prescribed by the composer was a normative property of musical works, the practice of transcription was common and transcriptions were not usually counted as new compositions. From the idea that instrumentation is a normative property for properly formed performances of a musical work (in those contexts), it does not follow that it is an essential property of that work (in those contexts). What follows is the weaker claim that it would be more or less preferable for the work to be performed by the original instrumentation. Thus, this opens the possibility of having new instrumentations of a medium-specific work without creating new musical works different from the original. Let us recall the example of Mussorgsky’s Pictures at an Exhibition. The first transcription of this work was made in 1886—five years after Mussorgsky’s death—by Mikhail Tushmalov, a student of Mussorgsky’s friend Rimsky-Korsakov. The relevant practices in the context of composition of this transcription are the same as those of Mussorgsky. However, the transcription was neither premiered nor identified as a new musical work composed by Tushmalov—a work different from Mussorgsky’s Pictures at an Exhibition. Moreover, this transcription is usually programmed as a work from Mussorgsky, and when people go to a concert to listen to the transcription, or reproduce a recording of it, what they intend to listen to is a work by Mussorgsky, despite its being orchestrated by Tushmalov. The same can be said of Ravel’s transcription. If, as Davies claims, transcriptions are new musical works, different from the original, our practices of referring to this work would be mistaken. But we are successful in our referring practices concerning this work. Consequently, Pictures at an Exhibition illustrates how a medium-specific work does not have its medium as an essential component, since transcriptions of it are never considered new musical works different from the original piece written by Mussorgsky. Although the context in which Pictures at an Exhibition was composed determines the medium-specificity of this work, changing its medium is not enough to give rise to a new work.
A third reason given by Davies for considering transcriptions new musical works is that colour (or timbre) is generally a condition of the structure and content of a musical work. Since transcriptions involve a change of medium, and a change of medium (usually) entails a change of colour, a transcription of a work W would usually modify W’s original structure and content, resulting in a new and different piece. Davies defends this point using an analogy between music and painting. Concerning painting, Davies claims:
“The colours of paintings often make a vital contribution to organizing the represented space or revealing its contents in other ways […]. Other structurally relevant spatial effects are generated via interactions between the relative area, contrast, complementarity, saturation, hue, and brightness of the colours used.” 
And in the case of music, an analogous phenomenon takes place. Davies claims:
“A work’s instrumental colour often makes a vital contribution to structural and other features. It helps delineate form and can add expressive and depictive qualities that are central to the work’s character and identity.”
For Davies, the key of the analogy is that “in music, the equivalent of colour is timbre”. However, although we ordinarily use the word ‘colour’ to refer to the instrumentation of musical works, I argue that we do so in a metaphorical way, and that the analogy with painting is essentially mistaken. In the case of painting, we talk about harmony between colours. Chromatic harmony is a specific term that refers to the “correct proportion and correspondence between colorations or between combinations of colours”. For instance, we say that Monet’s Impression, soleil levant has a harmonic combination of blues and greys. However, we do not talk about harmony between timbres in the case of music. For example, we do not say that the orchestration of Debussy’s Sirenes is harmonic. We would say that it is balanced, brilliant, or smooth—but not harmonic. Strictly speaking, harmony in music refers to the proportion between pitch heights, and this point illustrates the disanalogy between colour in painting and timbre in music. This disanalogy is grounded on the physics of colour and sound. Colour depends on wave frequency. Higher wave frequencies tend towards the ultraviolet spectrum, while lower wave frequencies tend towards the infrared spectrum. On the other hand, height of pitch or tone also depends on the frequency of the fundamental pitch. If wave frequency is higher, the pitch is more acute, and if wave frequency is lower, the pitch is less acute. However, timbre does not depend on wave frequency but on the intensity of the harmonic pitches associated with a fundamental pitch. Intensity is related to wave amplitude, which is independent of wave frequency. Therefore, if we want to draw an analogy between music and painting, height of pitch in music is a better candidate than timbre as an analogue of colour in painting. According to this analogy, a sound structure would require the existence of a sequence of pitches. Sound structures thus would depend on heights of sounds and not on timbre. It follows, then, that the form or structure of a musical work would not depend on timbre. Since a change of medium entails a change of timbre, and since the structure of a musical work would not depend on timbre, a change of medium of a work W would not entail a change in W’s structure, and hence in W’s identity. The result of this change of medium would not be a musical work different from W. Therefore, Davies’ third reason does not go through, because an accurate analogy between music and painting would show that a change in timbre is insufficient for a change in a work’s identity.
Consequently, none of the reasons in Davies’ account are strong enough to warrant the revisionist idea that transcriptions are new musical works different from the original. Even for those pieces that are medium-specific, we thus conclude that a change of medium is not a sufficient condition to designate a musical work as being new. Therefore, if we follow our common intuitions about musical practices, and there are no good reasons to reject them, transcriptions should not be counted as new musical works different from the originals.
According to Davies, transcriptions—but not versions—are new and different musical works from the original. In this paper, this thesis has been shown to have unacceptable consequences for our intuitions concerning musical practices, and thus it has been considered revisionist. We then examined whether in Davies’ account there are additional reasons that could lead us to accept such a revisionist claim. Davies’ main point is that a change of medium is sufficient for obtaining a new work in cases where the original is a medium-specific one. Three possible reasons supporting this claim have been examined and rejected here. First, we rejected Davies’ assertion that certain aesthetic properties essential to musical works depend on the specific medium in which they are performed by pointing out that those properties do not pertain primarily to the works themselves, but to their performances. Second, according to Davies, the context of composition determines whether a piece is medium-specific, and if a piece is medium-specific, instrumentation or colour is essential to it. Against this, we argued that even in contexts that determine that musical works are medium-specific, their transcriptions are not taken to be new works different from the original. Finally, we rejected Davies’ idea that colour (or timbre) is a necessary condition of the structure and content of a musical work by showing the physical disanalogy between colour in painting and colour (or timbre) in music.
About the Author Nemesio García-Carril Puy is a Graduate Teaching Assistant and PhD candidate at the Department of Philosophy I of the University of Granada. He is member of the research project Naturalismo, Expresivismo y Normatividad (FFI2013-44836-P) and of the research group Filosofía y Análisis (FYA) HUM-975. He is currently writing a thesis on the ontology of music and teaching epistemology and aesthetics. He has also a BA and MA in music, and is a freelance French horn player. He has recorded several CDs, three of them with Proemium Metals Brass Quintet, and he has played in Spain, Germany, Portugal, and the United Kingdom, both as a soloist and as a member of different orchestras and chamber music groups. He is currently a member of Proemium Metals Brass Quintet and of the Cambridge University Musical Society.
Davies, David. (2004). Art as Performance. Oxford: Blackwell Publishing.
Davies, Stephen. (2001). Works and Performances. A Philosophical Exploration. Oxford: Clarendon Press.
—. (2003). Themes in the Philosophy of Music. Oxford: Oxford University Press.
—. (2007). “Versions of Musical Works and Literary Translations”. In K. Stock, Philosophers on Music: Experience, Meaning and Work. Oxford: Oxford University Press.
—. (2008). “Musical Works and Orchestral Colour”. British Journal of Aesthetics 48 (4), pp. 363–375.
Gallego, Rosa & Sanz, Juan. C. (2001). Diccionario Akal del Color. Madrid: Akal.
Kivy, Peter. (1988). “Orchestrating Platonism”. In: Aesthetic Distinction. Ed. Thomas Anderberg, Tore Nilstun, and Ingmar Persson. Lund: Lund University Press, pp. 42–55.
Levinson, Jerrold. (2011). Music, Art and Metaphysics. Essays in Philosophical Aesthetics. Oxford: Oxford University Press.
 Davies 2007, p. 87 (my emphasis).
 Davies 2007, p. 86.
 Davies 2003, p. 48.
 Davies 2007, p. 86.
 Davies 2008, p. 365.
 Ibid., p. 369.
 I am very grateful to David Bordonaba Plou for his contribution and help on this point.
 Davies 2001, p. 63.
 Ibid., pp. 58, 74, 76.
 Davies 2008, pp. 374–75.
 Ibid., p. 374.
 Davies 2007, pp. 86–87; 2003, p. 48.
 This idea of the complex contribution of the context to the individuation of a piece is also shared and developed by David Davies, who claims that the identity of a piece is work-relative (cf. Davies 2004, pp. 103–126).
 Cases in which a change of medium does not entail a change of colour are logically possible. If we play Beethoven’s 5th Symphony on a Perfect Timbral Synthesizer, ex-hypothesi there is a change of medium without a change of timbre. However, in practice, there have not been—and nor does it appear that there will be any time soon—transcriptions that change the medium without altering the colour of a piece. It is hard to see how such cases would satisfy our practical interests in transcriptions (cf. Davies 2003, pp. 51–54). Accordingly, I will assume for the rest of the paper that, concerning transcriptions, a change of timbre follows from a change of medium, although one can imagine that one does not necessarily follow from the other.
 Davies 2008, p. 363.
 Ibid., p. 365.
 Ibid., p. 364.
 Sanz and Gallego 2001, p. 96.
 This paper was funded by the Ministry of Education, Culture and Sport of the Government of Spain by means of the Programa Nacional de Formación del Profesorado Universitario (FPU13/00070), and by the research project Naturalismo, Expresivismo y Normatividad (FFI2013-44836-P). I would like to express my sincere thanks to my PhD supervisors, Juan José Acero and Neftalí Villanueva, to Derek Matravers, to David Bordonaba Plou, and the members of the research project FFI2013-44836-P, to the anonymous peer reviewer and to the editors of the DiA, Ryan Doran and Shelby Moser, whose comments and suggestions were crucial for the improvement of this paper.