Start Submission

Reading: Interpretive Authenticity: Performances, Versions, and Ontology

Download

A- A+
Alt. Display

Research Article

Interpretive Authenticity: Performances, Versions, and Ontology

Author:

Nemesio G. C. Puy

University of Murcia, ES
X close

Abstract

Winner of the Fabian Dorsch ESA Essay Prize.

Julian Dodd defends the view that, in musical work-performance practice, interpretive authenticity is a more fundamental value than score compliance authenticity. According to him, compliance with a work’s score can be sacrificed in cases where it conflicts with interpretative authenticity. Stephen Davies and Andrew Kania reject this view, arguing that, if a performer intentionally departs from a work’s score, she is not properly instantiating that work and hence not producing an authentic performance of it. I argue that this objection fails. A detailed analysis of work-performance practice reveals, first, that the normative scope of interpretive authenticity encompasses the practice of composing musical versions and that, second, when performers sacrifice score compliance to maximize interpretive authenticity, they are performing the target work by means of performing a version of it. By means of the nested types theory, I then show how performances produced in this way can be properly formed instances, and hence authentic performances, of their target work.

How to Cite: Puy, Nemesio G. C.. “Interpretive Authenticity: Performances, Versions, and Ontology”. Estetika: The European Journal of Aesthetics 59, no. 2 (2022): 135–52. DOI: http://doi.org/10.33134/eeja.327
61
Views
6
Downloads
  Published on 15 Sep 2022
 Accepted on 08 Jul 2022            Submitted on 14 Feb 2022

I. Introduction

Julian Dodd identifies two fundamental performance values in classical work-performance practice:1 score compliance authenticity and interpretive authenticity.2 Both are good-making features that make a performance preferable to performances that lack them, other things being equal. Score compliance authenticity (SCA) and interpretive authenticity (IA) are two ways in which a performance can be faithful to the performed work. A score-compliant authentic performance of a work W is faithful to W by means of accurately rendering W’s score into sound. In turn, an interpretively authentic performance of W is one that is faithful to W by means of displaying a deep or profound understanding of W’s content. Both kinds of authenticity usually go hand in hand. Dodd notes, however, that they sometimes come into conflict. Accurately following W’s score may frustrate ‘the successful development of the point’ or the artistic consistency of W in performance (BT, p. 145). He maintains that, in cases of normative conflict, the sacrifice of score compliance for the sake of producing performances that evince understanding of the performed works is not only sanctioned but also ‘required’ in work-performance (BT, p. 164). The reason he adduces is that, by contrast with SCA, IA is a constitutive norm and the most fundamental value of that practice.

The use of musical scores characterizes work-performance, distinguishing it from other practices that hold, for instance, in the domain of folk, popular, and jazz music. Scores have a central normative import for rendering the performed works. Given the authority traditionally attributed to scores, Dodd’s view on authenticity looks prima facie striking and challenging. Indeed, Stephen Davies and Andrew Kania demur, arguing that SCA is non-negotiable in work-performance.3 They suggest that Dodd’s view on authenticity faces an ontological problem: performers who deliberately depart from W’s score do not properly instantiate W, and hence they do not produce authentic performances of W.

I argue that this objection from ontology fails and, accordingly, does not undermine the idea that compliance with W’s score can be sacrificed to produce interpretively authentic performances of W. In support of this thesis, I proceed as follows. In Section II, I introduce Dodd’s view on authenticity and the paradigmatic examples he supplies in its support. In Section III, I analyse the objection from ontology against Dodd’s view. First, I consider in more detail the criticisms by Davies and Kania and Dodd’s response to them. I note that there is, however, a more subtle way to put the objection that evades Dodd’s response and yet captures Davies’s and Kania’s main intuition. In Section IV, I argue that the objection is nonetheless motivated by a misguided approach to work-performance. A more detailed analysis of work-performance will reveal, first, that the normative scope of IA as a constitutive value encompasses the practice of composing a work’s versions and, second, that when performers sacrifice SCA to evince a profound understanding of W they are performing the target work by means of performing a version of it. Finally, I show by means of the nested types theory how performances that depart in this sense from W’s score can be properly formed instances, and thus interpretive authentic performances, of W.

II. Dodd’s View on Authenticity

Dodd defines the authenticity of a musical performance as a kind of faithfulness to something, as a matter ‘for the performance to be true to something that it purports to be true to’ (BT, p. 2). When the intentional object of this faithfulness is the performed work, we have a case of what Dodd calls work authenticity. He observes that work authenticity has a central place in work-performance. This practice is ‘work-focused’, which means that it is one in which ‘we value the performed work for its own sake, and we evaluate the performance as a performance of the work’ (BT, p. 4). However, performers do not only play the role of being mere deliverers of the performed work. They are also ‘sense makers’ of the performed work and are expected to present a ‘convincing interpretation’ of it (BT, p. 5). SCA and IA arise in this context as two forms of work authenticity. A performance can be faithful to the performed work either by accurately adjusting to its score’s instructions or by presenting it in a revelatory and insightful way.

SCA and IA have a specific normative profile in work-performance, according to Dodd. First, they are values, which means that a performance that is authentic, either in the score-compliant or in the interpretive sense, is preferable to any other performance that is not, other things being equal. And, second, SCA and IA have final value. They occupy a non-consequentialist space insofar as they cannot be negotiated for the sake of maximizing other performance values, like originality, liveliness, novelty, or expressiveness (BT, pp. 144–45). Accordingly, performers have a pro tanto obligation to maximize in performance the accuracy with the work’s score as well as the understanding of the performed work.

Given the work-focused nature of work-performance, the two kinds of authenticities usually go together: an insightful presentation of W in performance usually demands the performer’s full compliance with the instructions encoded in the score written by W’s composer. Dodd notes, however, that the two performance values may sometimes come into conflict. Being accurate in following W’s score may sometimes prevent the performer from offering performances that are revelatory of W’s content. In those cases, Dodd argues, performers ‘often try to achieve greater depth or profundity in understanding the work they perform by virtue of compromising a little on score compliance’ (BT, p. 145). This is an empirical claim that Dodd supports by means of different examples. Let me consider two of them.

Karajan and other conductors replace bassoons by horns in the recapitulation (mm. 303–7) of the first movement of Beethoven’s Symphony No. 5, disobeying the norms for performing-means indicated in Beethoven’s score. According to Dodd, this decision is motivated by a profound understanding of the work’s internal logic: keeping the fanfare function of this motif just before the smooth character of the B theme’s melody. Natural horns of Beethoven’s time were not able to perform in natural sounds the motif in the tonality in which it appears in the recapitulation. With the implementation of the valve mechanism, modern horns can play the motif in the recapitulation just as well as in the exposition. Dodd argues that, for Karajan, ‘truthfulness to the motif’s musical point overrides fidelity to an instruction that is rooted in historical contingency concerning the evolution of the horn’ (BT, p. 148).

A second example is Andreas Staier’s recording of Mozart’s ‘Rondo alla Turca’, in which he intentionally departs from Mozart’s score in some melodic, harmonic, and rhythmical aspects. Dodd notes that Staier’s departures are motivated by displaying understanding of the work’s content for a contemporary audience. The ‘Rondo’’s point is that of a ‘little ethnic showpiece’ that aims to ‘caricature the Turkish style through stylistic exaggeration’ (BT, p. 149). In Dodd’s view, to succeed in evincing the ‘Rondo’’s point to a contemporary audience, Staier stressed in his performance the ‘cartoonish character’ of this work by means of the sort of departures he carried out from Mozart’s score.

Dodd maintains that, in these cases of normative conflict, the sacrifice of score compliance for the sake of producing performances that evince understanding of the performed works is not only sanctioned but also ‘required’ by work-performance (BT, p. 164). As said, both SCA and IA are final values: they are valued for their own sake and performers have a pro tanto obligation to maximize them. However, according to Dodd, only IA is a constitutive norm of that practice: the value of IA directly derives from the ‘telos’ of the practice (that is, to facilitate the understanding of the performed work), whereas SCA is only a final value because it tends to secure that telos (BT, p. 158). Since failing to satisfy a practice’s constitutive norm entails failing to properly participate in that practice, it follows that we should sacrifice SCA to benefit IA when the two values come into conflict.

III. The Objection from Ontology

As noted above, Davies and Kania have questioned Dodd’s position. They maintain that, in work-performance, compliance with the score written by W’s composer is non-negotiable to authentically perform W. Apart from discussing Dodd’s examples, the core of their objection is ontological. In a nutshell, a performer who intentionally disregards the score written by W’s composer is not properly instantiating W; consequently, their performance is not true to W and it is not an authentic performance of W.

According to Davies, compliance with W’s score is not an interpretative option but an ‘ontological requirement’ to perform W.4 Davies argues that, for a performance to be an authentic performance of W, it is not enough for it to be revelatory of W; it must succeed in instantiating W. That is, the performance must satisfy ‘the composer’s work-determinative instructions’, being a subset of those ‘recorded in the work’s score’.5 In Davies’s view, although intentional departures from W’s score can be made in order to interpret W, such departures prevent a performance from satisfying W’s determinative properties. To intentionally disregard the instructions encoded in W’s score ‘is to fail to perform the work in question’, and hence to perform it authentically.6 According to Davies, this is what happens with Staier’s performance owing to his significant departures from Mozart’s score. Davies accepts that the performance may be revelatory of Mozart’s ‘Rondo’, but ‘without succeeding in instancing the work in question’.7

Expanding Davies’s line of reasoning, Kania argues that ‘to perform a work is to (reasonably) intend to instantiate all the work-determinative properties mandated by the composer in the score’.8 He notes that deliberate departures from the score belong to a different category from mistakes in performing a work. In his view, a condition to perform a work in work-performance is that the performer must intend to fully comply with the composer’s instructions written in the score (HC, p. 133). The reason is that a work’s identity and meaning are determined in a strong sense by its composer’s score (HC, p. 136). Thus, even if Staier’s and Karajan’s performances are revelatory of their target works, they do not faithfully convey those works’ content because of their departures from their composers’ scores. Kania offers two alternative ways of thinking about those cases.

Kania defends Karajan’s and Staier’s performances as being not of Beethoven’s Symphony No. 5 and Mozart’s ‘Rondo’, respectively, but of ‘a very minimal transcription or arrangement’ of them.9 According to Kania, ‘performances of arrangements of, fantasias or improvisations based upon, or works influenced by a given work might reveal things to listeners about their target work without being performances of it’.10 Insofar as the performances of Staier and Karajan intentionally depart from the scores of their target works, they do not instantiate the determinative properties of those works and are not performances of them. They are performances of versions or arrangements of their target works, which Kania considers to be different works of music.

In a more recent article, Kania provides a slightly different way of seeing Karajan’s and Staier’s cases. He concedes that their performances may reveal important aspects of the content of their target works. However, since Karajan and Staier intentionally depart from the works’ scores, they perform those works not in toto but only partially. Their performances exemplify ‘most of the properties of those works without (strictly speaking) being performances of them in their entirety, because of their intentional deviations from the work-determinative requirements of their scores’ (HC, p. 133). By means of playing notes different from those written in Mozart’s score of the ‘Rondo’, Staier is ‘producing an inaccurate instance of it’ (HC, p. 136). The same consequence follows for Karajan’s performance.

In short, Kania concludes that IA is constrained by SCA. In his view, there is only ‘a single constitutive norm at the heart of classical work-performance practice, namely to evince an understanding of the work through complying with its score’ (HC, p. 127).11 Dodd has mistakenly divided this norm into two and considered one of them – namely, IA – more fundamental than the other, SCA.

Dodd has replied to some of these criticisms (BT, pp. 155, 178). Concerning the ontological issue, Dodd claims that a performance being less than fully compliant with W’s score does not prevent it from being a performance of W (BT, p. 151). He notes that, for a performance to count as a performance of W, it must only satisfy the minimal level of compliance with W’s score that enables W to be recognized in that performance (BT, p. 161).12 If a performance satisfies this minimal level of compliance with W’s score and also evinces understanding of W’s content, that performance is an authentic performance of W in the interpretive sense. This is how we should understand the performances of Karajan and Staier considered above.

This response is consistent with Dodd’s ontology of musical works as norm-types of sound-sequence events.13 According to this view, a musical work is a type individuated by the condition that must be satisfied by a sound-sequence event in order to be a properly formed token of that type. This condition is a set of normative properties, and hence the exemplification of those properties admits of gradation. This means that a musical work qua type may have properly and improperly formed tokens, the latter being those sound-sequence events that do not exemplify all the properties that individuate the work qua type.14 Accordingly, for a sound-sequence event to be a token of the work at all, it must only exemplify a sufficient number of those normative properties that individuate the work qua type, not all of them.15 This is the case with Karajan’s and Staier’s performances, which comply with a reasonable number of their scores’ prescriptions. Accordingly, following Dodd’s ontology, there seems to be no ontological obstacle preventing them from being authentic performances of those works.16

We may concede that Dodd’s response is sufficient concerning the charge posed by Davies and Kania that Karajan’s and Staier’s performances are not performances of their target works. However, we may have more doubts about its soundness concerning Kania’s view that they are not performances of their target works in their entirety. Indeed, there is an alternative way to pose the ontological objection that is not undermined by Dodd’s response and that, to my view, captures Davies’s and Kania’s main concern.

Consider a performance p that is not fully compliant with the score of a work W. The core idea of the reformulated objection is that, if musical works are norm-types, being a mere token of W is not sufficient for p to be an authentic performance of W: p must also be a properly formed token of W. Or, in terms of the performer’s normativity, maximizing IA in performing W does not merely require tokening W; it implies maximizing the proper formation of W’s tokens. Let me expand on this point.

The tokens of a type T play a crucial role in our epistemic access to T. As Dodd himself notes, we have perceptual experience of a type passing through its tokens.17 The reason is that, first, the tokens of T are those things that satisfy the set of properties that individuates T, and, second, the things that we predicate of T are transmitted from its tokens.18 The transmission of predicates between tokens and their types is a feature that distinguishes types from other sorts of abstract entities. For instance, whereas the property of being red is not itself red, the type of the Union Jack and its properly formed tokens are red, white, and blue. Crucially, we have access to the Union Jack’s formal configuration and colours by means of experiencing its properly formed tokens. We cannot do this by means of seeing an improperly formed token of it, for instance one that has curvy pale grey lines. If I point to that improperly formed token as a reply to a child’s question of what the Union Jack is, the child will acquire a mistaken belief of the formal configuration of the Union Jack, because the token I have used does not transmit some of the properties that individuate the Union Jack qua type. Crucially, based on her mistaken belief, the child will not arrive at an adequate understanding of the Union Jack’s appearance.

I might tell the child that the Union Jack really has straight white lines instead of curvy grey ones. With this additional piece of information, the child might form a correct belief about the Union Jack’s formal configuration and colours. In general, we might accept that a subject can have epistemic access to the Union Jack on the basis of an accurate description of its formal configuration and colours. However, this procedure is not suitable for knowing and understanding a musical work’s content if we follow Dodd’s account. He maintains that the content of a musical work W is ‘specific’ and ‘inextricably tied to its sonic profile’.19 Accordingly, W’s meaning resists a paraphrase in words because any paraphrase leaves W’s meaning unspecified. The only way to know and grasp W’s content in an appropriate way is by means of hearing a performance of it. However, not any performance can play this role. Given the epistemic role that tokens play with respect to their types, it follows that only performances that are W’s properly formed tokens convey W’s content in a faithful way because they instantiate all the properties that individuate W qua type.

Recall that an interpretive authentic performance is not only revelatory of W’s content. It must also reveal W’s content in a faithful way, a condition included in the concept of authenticity; only this way the performance is true to W.20 The problem is that, if the normative properties that individuate W qua type are totally or partially represented in the score written by W’s composer (as Dodd’s ontology states),21 disobeying W’s score results in improperly tokening W, which disqualifies those performances as authentic performances of W. Authentic performances of W, even in the interpretive sense, can be only found among its properly formed tokens. A performance that departs from W’s score does not possess all the properties that individuate W qua type (that is, it is not a properly formed token of W), and hence it does not convey W’s content in an appropriate way. It is a performance that instantiates a different set of properties, thus providing a misleading epistemic access to W’s properties. Consequently, such a performance cannot provide an overall faithful aesthetic vision of W, which prevents it from facilitating a better understanding of W’s content.

Accordingly, the alternative way to put the objection from ontology against Dodd’s view on authenticity can be summarized as follows:

  1. If IA is the most fundamental performance value, performers ought to maximize IA whenever it clashes with SCA.
  2. Ought implies can: performers are under such obligation only if a performance that faithfully presents W’s content can be obtained by means of sacrificing compliance with W’s score.
  3. But a performance that departs from W’s score does not faithfully present W’s content because it is not a properly formed token of W.
  4. Therefore, compliance with W’s score is non-negotiable to authentically perform W in the interpretive sense, and hence IA simpliciter is not the most fundamental performance value.

This way of putting the objection is immune to Dodd’s replies introduced above. Hence, if it is right, it shows that compliance with W’s score is non-negotiable for producing authentic performances of W. It thus captures the issue addressed by Davies and Kania. In addition, premise 3 captures Kania’s way of seeing Karajan’s and Staier’s performances. After all, Kania argues that they are not performances of their target works in their entirety. They exemplify most of their properties but not all of them, which amounts to saying that they are not properly formed tokens of those works.22

IV. Returning to Musical Practice: Interpretive Authenticity and Versions

In Section V, I will argue that premise 3 of the objection from ontology is false and hence that the objection fails. Before that, I will reanalyse some aspects of work-performance that will play a crucial role in order to show why 3 is false.23 In particular, I will suggest that the objection is motivated by a misguided approach to work-performance that takes versions apart from the phenomenon of IA. As we shall see, there are good reasons to think that the normative scope of IA encompasses the practice of composing versions of a musical work.

To show this requires that we endorse Dodd’s distinction between SCA and IA as two different norms of work-performance. In other words, it requires rejecting Kania’s view that there is just a single constitutive norm of work-performance: to evince understanding of the performed work by means of complying with its score. Kania’s reason seems to be that the score includes the work-determinative instructions of the original work’s composer to perform the work. Thus, for Kania to be right, it must be true that, for the works performed in work-performance, there is a score that includes the work-determinative instructions of their original composers to perform them. Otherwise performers cannot comply with this norm and participate in a proper way in that practice.

The problem is that there is no such score for a significant number of works performed in work-performance. This is the case for Mozart’s Sonata K. 331. It is supposed that Mozart’s original manuscript had nine pages, but until 2014 all pages were lost except the last one, which only includes the coda of the ‘Rondo alla Turca’. In 2014, four more pages were discovered. However, as observed by Jonathan Del Mar, it is not guaranteed that the composer’s original manuscript captures her final instructions, for often modifications are made even after publication. What has become the canonical score of many works is usually the product of an editor. Even urtext editions do not always capture the work-determinative properties intended by the composer. As Del Mar notes, ‘some are so badly done, with so little understanding of the music or of the tradition into which the music was written’ (my italics).24

Two important consequences follow from this. First, often (and especially for works prior to the twentieth century) we do not have a document that tells us what exactly the work-determinative properties mandated by the composer are. Consequently, given that a constitutive norm of a practice cannot demand something that participants cannot fulfil in a significant number of cases, Kania’s claim does not fit with the actual practice. What does it mean to comply with Mozart’s score of the Sonata K. 331? It makes little sense to demand that a performer comply with Mozart’s score in order to properly perform this piece if there is no such score. Second, as revealed by Del Mar’s analysis, the norm of understanding the work guides the editor’s elaboration of the canonical score of that work. It is true that works performed in work-performance typically have a canonical score. But this score is elaborated upon understanding of those works. Thus, since following the norm of understanding a work is required for making its canonical score, this norm is prior and required for SCA and, hence, different from it.

Kania offers a second formulation of what he considers to be this single constitutive norm of work-performance: ‘to evince an understanding of the work by performing it’ (HC, p. 127). He thinks that this formulation is equivalent to the former one. However, we have good reasons to think that it is not. Work-performance enables cases of evincing understanding of a work by performing it that do not require complying with its canonical or original score. To show this requires taking into consideration the phenomenon of musical versions.

As noted above, Kania argues against Dodd’s view that, when a performer intentionally departs from the score written by W’s composer, she is not performing the target work but a version of it. Dodd rejects this connection between IA and the phenomenon of musical versions, claiming that the goals of the performer and of the version’s composer are radically different: the performer is work-focused, aiming to faithfully present W in a revelatory way in performance, whereas the version’s composer seeks ‘to produce a new artistic content based on that of the source work’ (BT, p. 99). My contention is that there is something right and something misleading in the views defended by Kania and Dodd. Kania is right, pace Dodd, that Staier is performing a version of Mozart’s ‘Rondo’ and Karajan is performing a version of Beethoven’s Symphony No. 5. Yet, pace Kania, Dodd is right that Staier’s aim is not to articulate a different artistic content from Mozart’s ‘Rondo’ and that he is not performing a different work. The same can be said of Karajan concerning Beethoven’s Symphony No. 5.

Kania and Dodd confuse the phenomenon of versions with that of inspired works. Whereas an inspired work (for example, Brahms’s Variations on a Theme by Haydn) is treated in work-performance as a different work from the source work (Haydn’s St Anthony Choral), a work’s versions (for example, the 1887 and the 1890 versions of Bruckner’s Symphony No. 8) are not regarded as different works but as different ways of presenting the same work in performance.25 This difference in the way in which inspired works and a work’s versions are treated in work-performance also reveals relevant differences in their composers’ goals.26 Brahms aimed to articulate a new artistic content different from, but based on, Haydn’s St Anthony Choral (the source work), with the purpose of being appreciated as a different work. Bruckner’s goal, in turn, was not that the 1890 version ought to be appreciated as a work different from his Symphony No. 8 but rather to improve the presentation of this work in performance by means of writing a score slightly different from the score he wrote in 1887. Bruckner was focused on the versioned work and did not aim to articulate a different artistic content.

This distinction, firmly rooted in work-performance, suggests that the normative scope of IA is wider than Dodd takes it to be, governing not only the practice of performing musical works but also the practice of composing versions of them. If IA is a value in the practice of composing musical versions in the same sense as it is in the practice of performing musical works, it must have the same normative profile: (i) it must be a value, which means that ceteris paribus a version is better for evincing a profound understanding of the versioned work; (ii) it must be a final value, and hence that it cannot be negotiated for the sake of maximizing other instrumental values (for example, originality or the composer’s artistic personality); and (iii) it must be a more fundamental value than SCA because of being a constitutive norm of that practice.

One of Bruckner’s main motivations for composing the 1890 version was the negative opinion of the 1887 version formed by the famous conductor Herman Levi. He considered it impossible to perform owing to its technical demands and judged it too complicated for the audience’s taste, recommending to Bruckner that he write a new version.27 The influence of Levi’s criticism on Bruckner’s composition of the 1890 version motivated an important debate about its authenticity. For instance, the editor Robert Haas argued that the 1890 version is inauthentic insofar as the work’s ‘organic life-essentials’ were compromised for the sake of making it easier to perform the work.28 By contrast, Benjamin Korstvedt considers the 1890 version to be authentic, and even preferable to the 1887 version, because the changes introduced by Bruckner ‘exhibit a compelling musical logic and it is wholly understandable that Bruckner felt that they improved the work’, for instance ‘solidifying the tonal framework’, or ‘emphasizing’ and ‘clarifying’ the preparation of the Trio of the third movement.29

The two parties to this debate about the authenticity of the 1890 version presuppose that a version is better ceteris paribus for evincing a profound understanding of the versioned work. This means that condition (i) is met. The parties also assume that the norm of understanding a work in composing a version of it cannot be compromised for the sake of maximizing other values, such as making it easier to perform or satisfying the audience’s taste, which meets condition (ii). Condition (iii) is also satisfied insofar as the composition of the 1890 version involves intentional departures from the work’s original score (that of the 1887 version) intended to facilitate the understanding of the versioned work. Bruckner would not have been involved in the practice of composing a version of his Symphony No. 8 if he had not followed the rule of being true to that work’s content. He would have been involved in a different practice: didactic, commercial, or even in the practice of composing an inspired work.

Two additional examples will illustrate the connection of the phenomenon of versions with Dodd’s paradigmatic cases of Staier and Karajan. Mozart was commissioned to compose a version of Handel’s Messiah with the aim of facilitating the understanding of this work to the Viennese audience.30 As Rachel Cowgill notes, given the evolution of musical language from Handel’s time, ‘for late-eighteenth-century audiences the services of an arranger were helpful in making sense of by then unfamiliar modes of expression’.31 Mozart increased the original orchestration with the inclusion of wind and brass parts, and enhanced the original textures, adding counterpoints and motives for the new instruments. However, he introduced those changes neither to exhibit his own artistic personality (indeed, Mozart abandoned his own compositional style to embrace Handel’s in light of new orchestration techniques) nor to articulate a new artistic content (he was ‘respectful of the integrity’ of the target work ‘as a whole’).32 Instead, the changes were intended to enrich the expression of the work’s climatic points of grief, joy, or sadness, and to articulate the end of sections, in ‘an attempt to clarify for contemporary audiences the structural outlines of the Baroque oratorio’.33

In his version of Handel’s Messiah, Mozart pursued the same ultimate goal as Staier in his performance of Mozart’s ‘Rondo’ – namely, to facilitate a profound understanding of the target work for their contemporary audiences. In addition, Mozart’s departures from the Messiah’s canonical score and Staier’s departures from the ‘Rondo’’s canonical score are musically of the same sort, enhancing harmonies and textures, and evincing the target work’s structure and expressive character. With his version, Mozart indicated a way of performing the target work with understanding. He intentionally departed from the canonical score of the original version of Handel’s Messiah by means of a compositional process that resulted in writing different norms for performance in a new score. Crucially, this compositional process involved an interpretation of Handel’s Messiah, an overall expressive and structural vision of that work.34 This vision guided and constrained Mozart in composing his version of the target work. In turn, Staier aimed to produce a performance that facilitated a better understanding of his target work. Crucially, Staier intentionally departed in performance from the canonical score of his target work after a process of composition of the same sort as that undertaken by Mozart in the Messiah’s case. Presumably, since Staier’s performance is for solo piano, he did not need to write those departures in a new score to communicate them to other performers.

The connection of the phenomenon of versions with Karajan’s case is evident if we consider Wagner’s version of Beethoven’s Symphony No. 9, which was motivated by his experience as a conductor of Beethoven’s work. In this version, Wagner tried to solve some problems in balance, projection, and expression that he had found when conducting Beethoven’s original version.35

A nice example offered by Wagner concerns the brass and wind fanfare at the beginning of the final movement. He describes this passage as a ‘chaotic outburst of wild despair’.36 This passage serves, in Wagner’s view, as a preparation of the recitative of cellos and double basses that comes next.37 Wagner notes two problems that prevent one from achieving the point of this passage if performed as written by Beethoven: the difficulty in following the wind’s theme and the impossibility of being free of the ‘tyranny of beat’ that makes a ‘laughable’ rather than a tragic effect.38 Wagner identified the origin of those problems with Beethoven’s intermittent use of trumpets in this passage. Beethoven was forced to write trumpets’ parts this way given the limitation of natural trumpets to perform the main melody. This limitation went away in Wagner’s time with the development of modern chromatic trumpets, able to play the full range of tones along more than three octaves. Wagner resolved to rewrite the trumpets’ parts, making them play the fanfare’s full theme in unison with winds.

The parallel between Wagner’s case and Karajan’s departure from the score in the case of the fanfare of Beethoven’s Symphony No. 5 is straightforward. In both cases, the aim is to do justice to the internal logic of their target works, overcoming historical contingencies concerning the evolution of brass instruments in particular passages. In addition, they are departures of the same sort – namely, of the kind that the musicologist Denis McCaldin calls ‘restitution of parts’, consisting in implementing notes in brass instruments that Beethoven had to omit due to their unavailability for the instruments of his time.39 With his version, Wagner indicated a way of performing the target work, Beethoven’s Symphony No. 9, that aims to facilitate its understanding He intentionally departed from Beethoven’s original score by means of a compositional process that resulted in writing different norms for performance in a new score. Wagner’s compositional process was guided and constrained by an interpretation of Beethoven’s Symphony No. 9 (indeed, Wagner abandoned his own personal style of composition for the sake of evincing the logic of Beethoven’s work). For his part, Karajan aimed to produce a performance that facilitated a better understanding of his target work, Beethoven’s Symphony No. 5. Crucially, Karajan intentionally departed in performance from the score of his target work after a process of musical composition of the same sort as that undertaken by Wagner. Karajan probably altered directly the orchestral parts of the musicians, a procedure followed by other conductors as Strauss or Furtwängler in their versions of Beethoven’s Symphony No. 9.40

The analysis of work-performance developed in this section reveals that Staier is performing Mozart’s ‘Rondo alla Turca’. However, he does it by means of performing not Mozart’s original version (represented by the canonical score of this work that is taken to collect most of Mozart’s instructions) but another version of the same work, in the same way as Charles Mackerras’s 1974 performance of Handel’s Messiah is a performance not of Handel’s original version (represented by the Messiah’s canonical score) but of Mozart’s version of that work.41 It also reveals that Karajan is not performing the original version of Beethoven’s Symphony No. 5 (represented by the score as written by Beethoven) but rather another version of the same work, in the same sense in which Wagner’s performance of Symphony No. 9 was not a performance of Beethoven’s original version. In both Staier’s and Karajan’s cases, the target work is performed by means of performing a version of it that is not the version indicated by the original work’s composer.

V. The Objection from Ontology Defeated

The analysis of work-performance presented in the previous section allows us to see that Davies’s and Kania’s objections to Dodd’s view on authenticity are not well grounded in practice. It is true that Karajan and Staier perform a version of the works: Beethoven’s Symphony No. 5 and Mozart’s ‘Rondo’, respectively. However, this does not imply that they are not performing their target works. Karajan is performing Beethoven’s Symphony No. 5 and Staier is performing Mozart’s ‘Rondo’. What Karajan and Staier are not performing are the original versions indicated by Beethoven and Mozart. Consequently, since Karajan’s performance is a performance of Beethoven’s Symphony No. 5, and his performance evinces a deep understanding of the internal logic of this work, it can be fully considered as an authentic performance of Beethoven’s Symphony No. 5. The same applies to Staier’s performance.

As we have seen, based on Kania’s view, there is nonetheless a more pressing way to put the objection from ontology to Dodd’s view on authenticity. According to premise 3, if a performance that departs from W’s score is not a properly formed token of W qua type, it does not convey W’s content in an appropriate way and it cannot be an authentic performance of W in the interpretive sense. Crucially, the analysis of work-performance developed above offers the grounds to show that premise 3 is false, and hence that the objection from ontology against Dodd’s view on authenticity fails.

The nested types theory explains how the properly formed performances of different versions of the same work can be properly formed tokens of that work. In contrast to traditional type/token theories, the nested types theory is a multiple-level type/token theory that distinguishes between more than two levels of objects linked by relations of instantiation. Under this approach, a musical work is a higher-order type that admits of instantiation in lower-order types. A work’s versions are lower-order types (in which the work qua higher-order type is instantiated) that admit of instantiation in musical performances, which are tokens of those lower-order types. In 1872 and 1879 Tchaikovsky wrote two versions of his Symphony No. 2. According to the nested types theory, Tchaikovsky’s Symphony No. 2 is a higher-order type T that is instantiated in two types of lower order: T1, the 1872 version, and T2, the 1879 version. Geoffrey Simon’s performance with the London Symphony Orchestra of the 1872 version recorded by the label Chandos, p1, is a properly formed token of T1 insofar as it accurately fulfils the instructions encoded in the score of that version. In turn, Riccardo Muti’s performance with the Philharmonia Orchestra of the 1879 version recorded by the label EMI Classics, p2, is a properly formed token of T2 because it accurately renders into sound the instructions of the score of this version.

According to the nested types theory, p1 and p2 are not only properly formed tokens of T1 and T2, respectively, but also of T. Tchaikovsky’s Symphony No. 2 is a higher-order type individuated by a sound structure that is more generic than the sound structures that individuate the lower-order types T1 and T2.42T has variables that do not specify all the details for its instantiation by lower-order types, and hence it can be properly instantiated by structurally different lower-order types. In addition, predicates and perceptibility are transmitted from tokens of lower-order types to the higher-order type.43 Some of the predicates that are true of p>1 are also true of T1: they are transmitted from p1 to T1. They are the properties that individuate T1 and that are instantiated by p1 in order to be one of its properly formed tokens. One of them is, for instance, ‘having syncopes at the beginning of the third movement’. A subset of those predicates is also true of T, and hence those predicates are transmitted from p1 to T. They are the properties that individuate T and that are instantiated by p1 in order to be one of its properly formed tokens. The predicate ‘having syncopes at the beginning of the third movement’ is not transmitted from p1 to T because this is not a property that individuates T. However, other predicates like ‘being tonal music’ or ‘containing Ukrainian folk tunes’ are transmitted between p1 and T because they are properties that individuate T.

Since Staier is performing a version of Mozart’s ‘Rondo’, his performance is subject to the analysis provided by the nested types theory. The work, Mozart’s ‘Rondo alla Turca’,44 is a higher-order type, K, that is instantiated in two types of lower order: Mozart’s original version, K1, and Staier’s version, K2. K is individuated by a sound structure that is more generic than the ones that individuate K1 and K2 (or, in other words, it is individuated by a less demanding condition than those that individuate its versions). It has variables that allow K to be instantiated in structurally different lower-order types.

K1 and K2 are properly formed instances of K insofar as they satisfy the set of properties that individuate K, but they fill K’s variables in different ways. Among the properties that individuate K we can count those of ‘being tonal music’, ‘having a rondo form’, ‘being an ethnic showpiece’, ‘being in 2/4’, ‘starting with a texture of accompanied melody’, and ‘starting with a melody that unfolds the A minor chord in an ascendant arpeggio with some embellishments’. A variable of K is, for instance, the way in which the embellishments of the basic melodic line are implemented when the first eight measures are repeated. In the case of K1, they are implemented by means of ascendant and descendent regular flourishes of semiquavers (see fig. 1). This is a property, F, that individuates K1 but not K. In the case of K2, they are implemented by means of a mordent as an appoggiatura of a quaver (the note of the arpeggio of A minor) and two semiquavers, the first of them as a bottom flourish (see fig. 2). This is a property, G, that individuates K2, but not K.

Canonical Score of Mozart’s Sonata K. 331, 3rd Movement
Figure 1 

Mozart’s original version according to its canonical score (mm. 1–4).

Transcription of Adreas Staier’s version of Mozart’s Sonata K. 331, 3rd Movement
Figure 2 

Staier’s version (mm. 1–4).

Staier’s performance does not have the property F but the property G. This fact disqualifies it as a properly formed token of Mozart’s original version, K1. However, this does not disqualify it as properly formed token of the work Mozart’s ‘Rondo alla Turca’, K. F and G are variable properties of K that specify different ways of implementing its more generic sound structure. Accordingly, certain properties of Staier’s performance, including G, are transmitted to K2; they are the properties that individuate K2 and it is by virtue of possessing them that Staier’s performance is a properly formed token of K2. A subset of those properties, which excludes G, are also transmitted to K: they are the properties that individuate K and it is by virtue of possessing them that Staier’s performance is a properly formed token of K. Having those properties (for example, ‘being tonal music’, ‘having a rondo form’, ‘being an ethnic showpiece’, ‘being in 2/4’, and ‘starting with a melody that unfolds the A minor chord in an ascendant arpeggio with some embellishments’) is what makes Staier’s performance one that appropriately conveys the content of the work Mozart’s ‘Rondo alla Turca’ and, hence, one in which we can experience and know this work in a faithful way. The fact that Staier’s performance is a properly formed token of its target work shows that premise 3 of the objection from ontology is false, and hence that the objection fails.

Before ending, I will consider two reasonable worries about the solution provided by the nested types theory. First, if all possible versions of a work were explained via the nested types theory, musical works would be too thin (have very few properties) so that there would be little of interest to say about them.45 As a response,46 let me note that, on the one hand, the debate about the authenticity of Bruckner’s versions suggests that there are good and bad versions of a work in the same sense as there are good and bad performances of a work. Bad versions are those that do not provide an adequate insight into the work’s content. And, on the other hand, there are alternative equally good versions of a work in the same sense as there are alternative equally good performances of a work. Those good versions provide an adequate insight into the work’s content, but some of them may facilitate a better understanding of the work. The nested types theory accounts for these two cases in different ways.47 The bad versions are improperly formed instances of the work, that is, lower-order types that fail to instantiate some of the properties that individuate the higher-order type. The alternative equally good versions are properly formed instances of the work: they instantiate the properties that individuate the higher-order type and realize its variables in different ways (some of those ways may contribute to a better understanding of the work’s content, and the variables’ scope is limited by the work’s point and musical medium). Therefore, the fact that a work may have infinite versions does not imply that the work is extremely thin: among those infinite versions, several of them are bad ones that fail to instantiate some of the properties that individuate the higher-order type.

A second, related worry is that the nested types theory operates something akin to a Schenkerian reduction of musical works, in this case of the ‘Rondo alla Turca’.48 To put aside this worry, let me highlight a couple of things. First, the nested types theory includes within the properties that individuate a musical work qua higher-order type ones that are not considered by a Schenkerian analysis (for example, ‘being an ethnic showpiece’). Second, what the nested types theory says is that many of the properties that individuate the higher-order type are determinable properties, which are determined at the level of its versions and performances. Consider, for instance, the property of ‘starting with a melody that unfolds the A minor chord in an ascendant arpeggio with some embellishments’. A Schenkerian reduction would eliminate (or obviate) the fact of the melody having some embellishments. The point of the nested types theory is a different one. It only notes that the work admits to realizing in sound those embellishments in different ways. It admits of the embellishments being performed in the way they are written in the canonical score of Mozart’s ‘Rondo’ and also in the way Staier implements them.

In sum, the nested types theory shows that intentional departures from a work’s canonical score do not prevent performances from being properly formed tokens of that work, thus defeating the objection from ontology. As a consequence, SCA is an ontological requirement for authentically performing (in the interpretive sense) the work’s version represented by its canonical score, but not for authentically performing (in the interpretive sense) the work. Being compliant with Beethoven’s score of his Symphony No. 5 is ontologically required for something to be a properly formed token of Beethoven’s original version. However, this is not required for something to be a properly formed token of the work this version is of. Therefore, the objection from ontology does not undermine the view that IA is the most fundamental performance value and that SCA can be sacrificed to maximize that value.

Notes

1In what follows, I use ‘work-performance’ to refer to ‘classical work-performance practice’. 

2Julian Dodd, ‘Performing Works of Music Authentically’, European Journal of Philosophy 23 (2012): 485–508; Being True to Works of Music (New York: Oxford University Press, 2020), hereafter: BT. 

3Stephen Davies, ‘Performing Musical Works Authentically: A Response to Dodd’, British Journal of Aesthetics 53 (2013): 71–75; Andrew Kania, Philosophy of Western Music: A Contemporary Introduction (New York: Routledge, 2020); ‘The Heart of Classical Work-Performance’, British Journal of Aesthetics 62 (2022): 125–41, hereafter: HC. 

4Davies, ‘Performing Musical Works’, 74; Stephen Davies, Musical Works and Performances: A Philosophical Exploration (Oxford: Oxford University Press, 2001), 207–8. 

5Davies, ‘Performing Musical Works’, 71. 

6Stephen Davies, Musical Understandings and Other Essays on the Philosophy of Music (Oxford: Oxford University Press, 2011), 108. 

7Davies, ‘Performing Musical Works’, 74. 

8Kania, Philosophy of Western Music, 191. 

9Ibid. 

10Ibid., 195. 

11See also Kania, Philosophy of Western Music, 196. 

12Dodd borrows this criterion of recognizability from Davies’s theory of performance (Davies, Musical Works as Performances, 160–61). Kania rejects the suitability of this criterion for being, in his view, too lax. He thinks that it can be satisfied by cases such as this: once the students of a middle-school orchestra have accurately played the first two measures of Beethoven’s Symphony No. 5, ‘they can hack through the rest of the piece however they like without any fear that a knowledgeable audience will fail to recognize the work they claim to be playing’ (HC, p. 132). This is to my view a misunderstanding of the recognizability criterion. What the criterion demands is not the target work to be recognized from the performance, but to be recognised in the performance. To this extent, as Davies notes, the criterion demands that the work’s ‘higher semantic structures remain sufficiently intact’ in the performance (Davies, Musical Works and Performances, 161). Thus, the criterion seems to be much more demanding than Kania interprets it. 

13See Julian Dodd, Works of Music: An Essay in Ontology (New York: Oxford University Press, 2007). 

14Ibid., 32. 

15Ibid., 34. 

16Kania has suggested another point of tension between Dodd’s views about musical ontology and musical authenticity. He argues that his view on authenticity suits better a contextualist ontology of musical works, which he takes to be incompatible with musical Platonism. However, this seems not to be a deep problem. Platonism is, strictly speaking, a thesis about the existence conditions of musical works qua types, but not about their identity conditions. Thus, it is possible to maintain that musical works are eternal types individuated by (relational) contextual properties. For more details, see Nemesio G. C. Puy, ‘Contextualizing Platonism and Decontextualizing Aristotelianism in the Ontology of Music’, Journal of Aesthetics and Art Criticism 78 (2020): 183–96. 

17Dodd, Works of Music, 11. 

18Ibid., 17. 

19Ibid., 127. 

20As we shall see in the next section, this idea is neutral with respect to the one that W is totally or partially individuated by the normative properties specified by the composer in the score. Indeed, I will argue that we have good reasons to retain the former and reject the latter. If those ideas were not independent, a careful reader might think that I would be begging the question against Dodd’s view. I am very grateful to Kania for a remark on this point. 

21Dodd, Works of Music, 112. 

22Kania (private communication) expresses his doubts about endorsing 3. He considers that the problem is not that, say, Karajan’s performance is not a properly formed token of Beethoven’s Symphony No. 5. In his view, the problem is rather that this performance is not intended to be a properly formed token of Beethoven’s Symphony No. 5. But I should note that this way of putting things is problematic for his view. Suppose that Karajan goes to perform the symphony with a different orchestra, which has always performed the work in a fully score-compliant sense. During the only rehearsal, Karajan asks horns to play the bassoons part in the recapitulation of the first movement. However, when it comes to the live concert, bassoons and horns forget Karajan’s indication due to the force of habit, and they perform the symphony as it is written. The performance is fully compliant with Beethoven’s work-determinative prescriptions, and it evinces understanding of the work through complying with its score. However, Karajan’s intention was not to comply with Beethoven’s score, and hence, in Kania’s terms, the performance was not intended to be a properly formed token of the work. Consequently, this performance should not be regarded as an authentic performance of Beethoven’s Symphony No. 5, which contradicts Kania’s view on authenticity. This shows that Kania should endorse 3. To offer a consistent account, Kania cannot reject as a work’s authentic performances solely those that are not intended to be properly formed tokens of it; he must reject those that are not intended to be properly formed tokens of the work and that succeed in failing to be properly formed tokens of it. 

23I am very grateful to Luis Otero for suggesting some of the examples in this section. 

24Jonathan Del Mar, ‘The Plain Truth about Urtext, Part 1’, interview, Classics Today, 6 April 2017, https://www.classicstoday.com/plain-truth-urtext-part-1. 

25Nemesio G. C. Puy, ‘The Ontology of Musical Versions: Introducing the Hypothesis of Nested Types’, Journal of Aesthetics and Art Criticism 77 (2019): 242–45. 

26See Nemesio G. C. Puy, ‘Nested Types and Musical Versions: Replying to Davies’ Challenges’, Journal of Aesthetics and Art Criticism 79 (2021): 256–60. 

27Benjamin M. Korstvedt, Bruckner: Symphony No. 8 (Cambridge: Cambridge University Press, 2000), 17–18. 

28Cited in ibid., 69. 

29Ibid., 72–76. 

30Rachel Cowgill, ‘An Unknown Handel Arrangement by Mozart? The Halifax Judas’, Musical Times 143 (2002): 19. 

31Ibid. 

32Ibid., 30. 

33Ibid. 

34See Davies, Musical Understandings, 110, for a definition of interpretation. 

35See Denis McCaldin, ‘Mahler and Beethoven’s Ninth Symphony’, Proceedings of the Royal Musical Association 107 (1981): 107–8. 

36Richard Wagner, ‘The Rendering of Beethoven’s Ninth Symphony’, in Richard Wagner’s Prose Works, vol. 5, ed. and trans. William Ashton Ellis (New York: Broude, 1966), 239. 

37Ibid., 241. 

38Ibid., 239–40. 

39McCaldin, ‘Mahler and Beethoven’s Ninth Symphony’. 

40See Raymond Holden, ‘The Iconic Symphony: Performing Beethoven’s Ninth Wagner’s Way’, Musical Times 152 (2011): 4. 

41George Frideric Handel, Der Messias/Messiah, orchestrated by Wolfgang Amadeus Mozart, Austrian Radio Symphony Orchestra, Sir Charles Mackerass, recorded February 1974, Deutsche Grammophon, 1991, 2 CDs. 

42Puy, ‘Ontology of Musical Versions’, 250. 

43Ibid., 248–49. 

44To simplify the presentation, I consider the ‘Rondo alla Turca’ as a work, although it is strictly speaking a movement of a work, Mozart’s Sonata K. 331. This does not affect the conclusions that will be achieved. 

45I am very grateful to Kania for this remark. 

46For a complementary response see Puy, ‘Nested Types and Musical Versions’, 258. 

47See Puy, ‘Ontology of Musical Versions’, 250. 

48I am very grateful to Jerrold Levinson for this observation. 

Acknowledgements

This research has been supported by a Juan de la Cierva Fellowship (FJC2019-041091-I) of Ministerio de Ciencia e Innovacion del Gobierno de España, and by the research project Normative Aspects of Aesthetic Appreciation (PID2019-106351GB-I00). I have benefitted from fruitful discussions of previous drafts of this paper at ARESMUR Permanent Seminar (University of Murcia), the European Society for Aesthetics Conference 2021, and the 79th Annual Meeting of the American Society for Aesthetics. I am indebted to Luis Otero and Andrew Kania for numerous comments and discussions on this article. I am also very grateful to María José Alcaraz, Hanne Appelqvist, Alessandro Bertinetto, Matilde Carrasco, Vítor Guerreiro, Jerrold Levinson, Vítor Moura, Jonathan Neufeld, Francisca Pérez Carreño, and Salvador Rubio for very useful comments that helped improve the paper.

Competing Interests

The author has no competing interests to declare.

References

  1. Cowgill, Rachel. ‘An Unknown Handel Arrangement by Mozart? The Halifax Judas.’ Musical Times 143 (2002): 19–36. DOI: https://doi.org/10.2307/1004420 

  2. Davies, Stephen. Musical Understandings and Other Essays on the Philosophy of Music. Oxford: Oxford University Press, 2011. DOI: https://doi.org/10.1093/acprof:oso/9780199608775.001.0001 

  3. Davies, Stephen. Musical Works and Performances: A Philosophical Exploration. Oxford: Oxford University Press, 2001. 

  4. Davies, Stephen. ‘Performing Musical Works Authentically: A Response to Dodd.’ British Journal of Aesthetics 53 (2013): 71–75. DOI: https://doi.org/10.1093/aesthj/ays068 

  5. Del Mar, Jonathan. ‘The Plain Truth about Urtext, Part 1.’ Interview. Classics Today. 6 April 2017. https://www.classicstoday.com/plain-truth-urtext-part-1. 

  6. Dodd, Julian. Being True to Works of Music. New York: Oxford University Press, 2020. DOI: https://doi.org/10.1093/oso/9780198859482.001.0001 

  7. Dodd, Julian. ‘Performing Works of Music Authentically.’ European Journal of Philosophy 23 (2012): 485–508. DOI: https://doi.org/10.1111/j.1468-0378.2012.00546.x 

  8. Dodd, Julian. Works of Music: An Essay in Ontology. New York: Oxford University Press, 2007. 

  9. Handel, George Frideric. Der Messias/Messiah. Orchestrated by Wolfgang Amadeus Mozart. Austrian Radio Symphony Orchestra. Sir Charles Mackerass. Recorded February 1974. Deutsche Grammophon, 1991. 2 CDs. 

  10. Holden, Raymond. ‘The Iconic Symphony: Performing Beethoven’s Ninth Wagner’s Way.’ Musical Times 152 (2011): 3–14. 

  11. Kania, Andrew. ‘The Heart of Classical Work-Performance.’ British Journal of Aesthetics 62 (2022): 125–41. DOI: https://doi.org/10.1093/aesthj/ayab031 

  12. Kania, Andrew. Philosophy of Western Music: A Contemporary Introduction. Routledge, 2020. 

  13. Korstvedt, Benjamin M. Bruckner: Symphony No. 8. Cambridge: Cambridge University Press, 2000. DOI: https://doi.org/10.1017/CBO9780511605918 

  14. McCaldin, Denis. ‘Mahler and Beethoven’s Ninth Symphony.’ Proceedings of the Royal Musical Association 107 (1981): 101–10. DOI: https://doi.org/10.1093/jrma/107.1.101 

  15. Puy, Nemesio G. C. ‘Contextualizing Platonism and Decontextualizing Aristotelianism in the Ontology of Music.’ Journal of Aesthetics and Art Criticism 78 (2020): 183–96. DOI: https://doi.org/10.1111/jaac.12714 

  16. Puy, Nemesio G. C. ‘Nested Types and Musical Versions: Replying to Davies’ Challenges.’ Journal of Aesthetics and Art Criticism 79 (2021): 256–60. DOI: https://doi.org/10.1093/jaac/kpab012 

  17. Puy, Nemesio G. C. ‘The Ontology of Musical Versions: Introducing the Hypothesis of Nested Types.’ Journal of Aesthetics and Art Criticism 77 (2019): 241–54. DOI: https://doi.org/10.1111/jaac.12651 

  18. Wagner, Richard. ‘The Rendering of Beethoven’s Ninth Symphony.’ In Richard Wagner’s Prose Works, vol. 5, edited and translated by William Ashton Ellis, 229–54. New York: Broude, 1966. 

comments powered by Disqus