Abstract
Musical virtuosity is often studied with an emphasis on the outstanding skills of the performer and their sensory and emotional effects on the listener. However, Franz Liszt, one of the main virtuosos in music history, was convinced that what matters in virtuosity is its specific function within the semantic and communicative processing of music. This article proposes a theoretical framework to further investigate Liszt’s intuitions in light of contemporary research on musical meaning and communication. First, we analyze philological data concerning Liszt’s creative process in the Mazeppa works, a set of works including both his virtuoso and program music. Second, we compare our results with recent research on music semantics based on inferential processes about the composer/performer’s intentions (Antović, Stamenković, & Figar, 2016; Schlenker, 2017). We criticize the associationist hypothesis (the idea that the composer’s intention is that the listener infers an association between the music and a real or fictional state of the world) and we defend a causalist hypothesis (the idea that the composer’s intention is that the listener recognizes the music as being an intentional “rewriting” of the states of the world, mental or physical, that are the causes of the music itself). Third, we suggest that what distinguishes Liszt’s virtuoso and program music rewritings are the visual and gestural components of two listening experiences: the virtuoso music “rewrites” its episodic causes (memories, personal experiences) whereas the program music “rewrites” its genetic causes (the work’s creative path and intertextual relations).
Virtuosity, meaning and communication in Liszt’s aesthetics: A cognitive problem
James Deaville has recently noted that most of the research about virtuosity in classical music, or even in music in general, doesn’t escape from what he calls “the Liszt problem” (Deaville, 2014, p. 292; see also Deaville, 2007; Cvejić, 2016, pp. 7–17), i.e. the fact that Liszt tends to be, if not the center, at least the starting point or the structuring example of the figure of the virtuoso. If it is true that most research about 19th century virtuosity focuses on Liszt (Deaville, 2003; Gooley, 2004; Kregor, 2010; Moysan, 2009; Reynaud, 2006; Samson, 2003), Deaville’s remark is worth being qualified. Indeed, recent years have seen an increasing amount of publications devoted to other virtuosos or other dimensions of virtuosity in which Liszt is not a structuring figure (Barizza & Morabito, 2010; Bösel & Sala, 2004; Carew, 2007; Cvejić, 2016; Gramit, 2008; Howard, 2008; Lilti, 2014; Metzner, 1998; Rasch, 2011; Rybicki, 2014), or at least not the only one (Bartoli & Roudet 2013; Bernstein, 1998; Colin, 2005; Davies, 2015; Morabito, 2014; Moysan, 2013; Penesco, 2007; Stoichita, Grimaud, & Jones, 2011; Thormählen, 2014). In this article, we will try to demonstrate that, due to the quantity of historical and philological data already collected about Liszt by musicologists, it is still relevant to focus on Liszt in order to better understand an important dimension of his virtuosity, namely the relation between the virtuoso dimension of his music and his conception of musical meaning and communication.
Our choice is motivated by the existence of a misunderstanding between the specific conception of virtuosity adopted by Liszt and our intuitive understanding of this phenomenon. The latter is characterized by the primary importance given to the virtuoso’s high level of ability in his/her domain and to the effects of various types (emotional, corporeal or social) aroused by the collective experience of watching, and listening to, a virtuoso performance. In this conception, research on virtuosity tends to concentrate on two points: on the one hand, techniques of observation, measurement, modeling and learning of such a high level of ability and its potential effects on the listener (Altenmüller, Wiesendanger, & Kesserling, 2006; Furuya, Oku, Miyazaki, & Kinoshita, 2015; Gorton & Kanga, 2016; Pachet, 2012; Vidal, 2011; for a more global perspective, see Leman, 2016); on the other hand, historical and/or critical analysis of its social, cultural, ethical and political consequences (see most of the references quoted in the previous paragraph; concerning Liszt’s case, see especially Deaville, 2003; Hennion, 2012; Moysan, 2009).
In both cases, we observe that virtuoso performances are not predominantly conceived as discourse acts aiming to communicate intentional meanings, coming from a single author and having a relatively recognizable referent, i.e. being the public support of a more or less vague representation of a part of the world (for an exception, see Kramer, 2002, pp. 68–99, 2016, pp. 113–140). Yet a referential, communicative and discursive conception of virtuosity was precisely the one Liszt had, as he expressed it many times in his writings (Liszt, 1859, pp. 281–282; 1995, pp. 54–55 [1835], p. 142 [1838], pp. 256–258 [1840], p. 387 [1859]; FLB, vol. 1, pp. 24–26 [1839], vol. 7, p. 408 [1884]). Moreover, working on musical and non-musical primary sources, recent research has highlighted the noteworthy influence of musical rhetoric in the training of musicians and composers of Liszt’s generation (Barth, 1992; Bartoli, 2006; Bartoli & Roudet, 2013, pp. 9–81, 317–347; Bonds, 1991, pp. 132–161; Carew, 2007, pp. 133–165; Cook, 2013, pp. 91–134; Horton, 2014; Kaltenecker, 2010, pp. 19–44, 219–226; Samson, 2003, pp. 35–65; for a general critique of the rhetorical conception of music see Vickers, 1984). They have confirmed the semantic and communicational foundations of the production and the perception of musical virtuosity at that time.
If we accept the validity of the above works about the communicational nature of Liszt’s conception and practice of musical virtuosity, we can derive from their results the following claims providing a preliminary insight on what is a virtuoso performance in a communicational perspective (for more details about the terminology used here and borrowed from the subdomain of linguistics called pragmatics, see Grice, 1989; Searle, 1983; Sperber & Wilson, 1995):
A virtuoso performance is a discourse act structured in two levels: a semantic level and a pragmatic (or communicational) level. The semantic level is concerned with the meaningful dimension of the performance, i.e. its referent, the representational content it possesses about a certain part of the world (the real world but also a fictional world, e.g. in the case of opera). The pragmatic level is concerned with the intentional and practical dimension of the performance. This dimension contains, on one hand, the action that the performer is doing, here a communicational action including necessarily a semantic component, and, on the other hand, the effects, also potentially practical, the performer is provoking in the listener by doing this communicational action.
A virtuoso performance will be a satisfied (or successful) discourse act if the effects it produces on the audience are experienced by it as the intentional communication of a certain representation about the world that the virtuoso performer has in his/her mind. By contrast, a virtuoso performance will be an unsatisfied (or unsuccessful) discourse act if the effects it produces in the audience, however powerful these effects could be, are experienced by the audience as something independent from any form of intentional communication of a representation about the world that the virtuoso performer has in his/her mind.
The main objective of this article is to propose a hypothesis that aims to describe which cognitive function does the virtuoso dimension of a performance play in the listener’s recognition of the performer’s communicative intention, i.e. in the satisfaction of the musical performance conceived as a discourse act containing a certain meaning. We will proceed in three steps.
First (in section 2), we analyze a set of historical and philological data related to a coherent corpus of Liszt’s works (the Mazeppa works) including his virtuoso music as well as his program music. We discuss some recent research about program music and refute the idea that the semantic content of the composer/performer’s communicative intention is limited to create a mental association (or a correspondence) between the music and the extramusical referent of the program or, more generally, with what Genette (1982) calls the paratext (the ensemble of signs surrounding or going with a given text or symbol, such as titles, covers, prefaces, program notes, dedications, epitaphs, illustrations, footnotes, comments). Regarding the conceptual distinction between virtuoso and program music, it is important to note that Liszt’s pianistic virtuoso music is the target of this research, whereas Liszt’s orchestral program music is limited to a heuristic role, a material of comparison with virtuoso music about its discursive and communicational dimensions. Because we assume that all Liszt’s virtuoso music is a discourse act, we assume that it can convey a “programmatic” meaning in a broader sense than the strict addition of a non-musical symbol besides the music, as it is the case in orchestral program music. The heuristic relevance of orchestral program music comes from the fact that Liszt made this programmatic dimension much more visible and systematic than in its virtuoso equivalent.
Second (in sections 3 and 4), we propose an alternative to this associationist hypothesis. This alternative is a causalist hypothesis. The main idea is that what Liszt wants to communicate is not the semantic association between the music and the extramusical elements exhibited through the program, but the story of this semantic association as it has been experienced by the composer/performer. In other words, Liszt’s communicative intention is not about the semantic relation between the music and the program but about the relation between this semantic relation and the causal sequence of mental and non-mental events that led him to compose/perform the music and include in it the above semantic relation. The condition of satisfaction of this communicative intention requires that the listener infers it directly from the perception of the work. This presupposes that the work contains a recognizable representation of the relevant causal sequence of events including an agentive entity such as the composer/performer. This representation is located in the binary device [program + music] whose two main qualities are, at the same time, to simplify and to simulate both a semantic association and a minimal causal sequence of events. We call rewriting the kind of musical discourse acts in which causal relations and meanings are merged within the musical work in a simplified and unified representation of a particular relation between an individual (here, the composer/performer) and a certain part of the world that s/he has experienced.
Third (in sections 5, 6 and 7), we propose a two-steps model of the cognitive processing of virtuoso rewritings.
In a first step of inferential nature, the listener extracts visual cues (the virtuoso’s attitudes on stage) and auditive cues (the music’s high level of technical difficulty and, consequently, the strong individualization of the performer). Combined with the presupposition that both the listener and the virtuoso share a somewhat similar background knowledge of the program and of the musical syntax of the work, the listener infers that the music is the virtuoso’s public expression of his/her mental states and their emotional correlates about his/her personal experience of the program’s content. In other words, virtuosity is, in this case, a key element enabling the music to be performed and interpreted as a personal testimony, as the narration of a memory, as a public attempt of an individual to share a past experience and to “relive” its intensity.
In a second step of mimetic (non-inferential) nature, the listener simulates in his/her mind what the virtuoso is experiencing and expressing on stage, i.e. rewriting publicly an intense private and past experience. But this mimetic simulation is not a simple mental and private copy of the musical and public testimony. Due to the specific characteristics of virtuosity mentioned above and, moreover, to the strongly differentiated and mutually accepted positions of a virtuoso and a listener in the conventional institution of a recital, this mimetic simulation will be processed through a segmentation of the listening experience. This segmentation will be processed according to the similar (or shared) and different (or non-shared) aspects of the two rewritings (the public one of the virtuoso and the private one of the listener). The similar or shared aspects are emotional ones. They induce an emotional identification between the virtuoso and his/her audience. The different or non-shared aspects are ethical ones. They induce a non-problematic confirmation of the qualitative distance between the virtuoso and the listeners, a notion close to the notion of “exemplary exception” usually defining the genius.
The Lisztian Mazeppas: Historical and philological data
Because our point of view and our sources concern the compositional (or poietic) dimension of music rather than the receptive (or esthesic) one (about this distinction see Nattiez, 2013, pp. 21, 31–35), our method will consist in comparing the processes of composition of Liszt’s virtuoso music with those of his program music (the second main type of music composed by Liszt). We shall focus on the most obvious formal feature shared by most of these two types of works: the existence, in their paratext, of a non-musical (literary or visual) or non-exclusively musical work of art (opera) presented in various forms such as complete reproduction, partial reproduction, paraphrase, mention in the title, dedication to the author of the work, etc. From a heuristic point of view, we consider the presence of another work of art as symptom of the referential and communicative intention previously mentioned. From the point of view of the listener, we consider the presence of another work of art as a conceptual cue facilitating the activation of the communicational phenomenon we aim to describe.
Our comparison between Liszt’s virtuoso music and program music is based on the corpus of the “Mazeppa” works composed before 1854: four etudes for piano and one symphonic poem. Philological data is provided by Johns (1997, pp. 65–68, 150–151), Charnin Mueller & Eckhardt (2001), Samson (2003, pp. 198–212), Liszt (2010), Brussee (2014a), Kaczmarczyk (2012), and Garratt (2013, pp. 203–206) as well as by our own research at the Bibliothèque nationale de France (Paris) and the Goethe-und-Schiller Archiv (Weimar). The genesis of Liszt’s Mazeppian works is far more complex than what is presented here but, for the sake of brevity, this data is summarized in a chronological table (see Table 1) in which each piece of data is numbered (D1, D2, etc.). We have excluded data that are too controversial, redundant or non-relevant in the context of this article. The data contains information coming from musical documents by Liszt’s hand (printed scores, manuscripts), non-musical documents by Liszt or related to Liszt (letters, journal articles) and influential artistic works related to the story of Mazeppa that were familiar to Liszt (on the cultural importance of Mazeppa’s character in Liszt’s time, see Babinski, 1974; Mainardi, 2003, pp. 178–212; Brussee, 2014b; Prymak, 2014; Sibona, 2006).
Liszt’s Mazeppa works.
Note. Historical and philological data describe Liszt’s five Mazeppian works before 1854; D1, D2, etc. indicates the numbering of the data; indicates musical manuscripts and non-musical documents related to these five works;
Antović et al. (2016) have recently published an experimental study about the semantics of program music conceptually based on the distinction between denotation and connotation introduced by Mill (1843). They defend an associationist hypothesis about the role played by the program in the way the listener recognizes an “extramusical” meaning or referent intentionally conceived by the composer. According to their hypothesis, the association between the non-musical program (poem, picture) and the extramusical referent is denotative, whereas the association between the music and the same extramusical referent is connotative (for a study with a similar postulate but focused on Wagnerian leitmotivs, see HaCohen & Wagner, 1997). The main result of their experiment is the following: by way of verbal judgments, the majority of participants attribute strongly similar extramusical connotations when the denotative program of the composer is not provided as well as when it is provided (knowing that, in both cases, the participants aren’t aware of the composer’s program). In return, when the participants listened to the music with a “false” denotative program that contradicts the composer’s one (e.g. the title “Lullaby” on Mazeppa’s music), the authors observed a significant difference in the associative judgments: without the composer’s program, a majority of participants formulated connotative associations in accordance with the denotative content of the composer’s program while, with the “false” program, the proportion of connotative associations in accordance with the composer’s program was half of what it was in the listening without any program.
The authors draw two conclusions from these results. First, they confirm the idea, already expressed by Hanslick (1854/1986, pp. 25–26) and by Liszt himself (1855/2005, p. 56; see also Micznik, 1999, pp. 231–234 and Candoni, 2012, pp. 183–188), that the program and the music are semantically redundant. In other words, the program does not provide a significant amount of complementary information to the musical connotations but, rather, a denotative equivalent of the connotative associations already aroused by the music. Second, they claim that the semantic associations between the music and an extramusical referent, whether they are obtained with or without a program and in denotative or connotative ways, constitute the content of the composer’s intentions. In other words, if such associations occurring in the listener’s mind are effectively caused by the binary device [program + music], the composer’s intentions are satisfied.
The problem of this approach lies in the postulate that it is the composer’s intention that the listener recognize the associations between the music and an “extramusical” referent, however precise or vague the listener’s “extramusical” referent be in relation to the program’s referent. Yet the historical and philological data that we have collected about Mazeppa go against this postulate, at least in Liszt’s case.
First, the association between Mazeppa’s wild ride through the Ukrainian steppes and the character of the travelling virtuoso, or more generally stereotypical figures such as the prophetic artist (or genius) with an extraordinary but restless destiny, is a commonplace in Liszt reception, and it appeared several years before the earliest appearances of explicitly “Mazeppian” programmatic indications in Liszt’s works (D7, D8, D10). This seems to have been a common and widespread association then, thanks to a significant number of listening experiences of Liszt in concert and to the vast iconographic production about both Liszt and Mazeppa at that time (drawings, paintings and caricatures). Regardless of the work performed and its eventual verbal program, the visual and listening experience of a Liszt-like virtuoso as such might already have been to be associated with extraordinary and fascinating destiny, embodied in historical and mythical characters such as Mazeppa, Prometheus, Orpheus, Tasso, Dante, Faust, someone fighting for his freedom, or any other characters that were later referred to by Liszt in his programs (Gooley, 2004, pp. 18–116; Davison, 2005).
Second, two aspects of the genesis of the five Mazeppa works demonstrate that it is difficult to maintain that Liszt’s intention was that the listener produces a more or less vague semantic association between his music and the story of Mazeppa.
The first aspect concerns the chronology of the works. Work 1 (D3) precedes any remaining trace of an intention to set the story of Mazeppa in music even if we know that Liszt was already very familiar with this theme (D1–2). Work 2 (D9) was composed during an intense reflection on the myth of Mazeppa, however, without the musical content being associated with the myth (D6–8). In this sense, the late and explicit association with Mazeppa seems to come from an a posteriori awareness by Liszt of the similarity between his music and the Mazeppa story. Indeed, the association was made public and explicit only from Work 3 (D10) (for a similar example concerning the symphonic poem Die Ideale, see Micznik, 1999, pp. 223–226). Thus we can notice that the genesis of the Mazeppa works seems to confirm the phenomenon of semantic redundancy between the program and the music studied by Antović et al. (2016).
The second aspect concerns the choice of the program. If Liszt’s intentions were entirely reducible to create a semantic association with the story of Mazeppa, he could have chosen a poem less hermetic and abstract than the one by Victor Hugo in which the story of Mazeppa is told twice, interlacing historical facts with an allegoric speculation about the figure of the genius (Babinski, 1974, pp. 66–73; Sibona, 2006, pp. 145–225). For example, he could have chosen Byron’s narrative poem (D1) that he was very familiar with and which was no less famous than Hugo’s. He could even have written himself a simpler paraphrase of the poem or a summary in prose of the story, as did Carl Loewe (D5). Not only did he not choose any of these alternatives but he also rejected every potential alternative to Hugo’s poem (D17, D19). Besides, the evolution from Work 3 to Work 5 (D10, D16, D18) shows that the entire paratext (title, preface, dedication, epitaph at the end of the score) is progressively “invaded” by the Hugo poem while, given the philological data previously mentioned, it is clear that this poem is far from being the only work, or even the most important or the first one, through which Liszt had discovered and appreciated the story of Mazeppa. Yet we know that paratextual elements and especially dedications had a crucial importance during the Romantic period (Carew, 2007, pp. 165–166; Green, 2009; Moysan, 2009, pp. 40–53; Samson, 2006). Admittedly, authors’ intentions behind paratextual choices were not only semantic but also sociological (paying tribute to the dedicatee, marking the membership of a social group or the aesthetic movement of the dedicatee, hoping for a future favor or thanking for a previous favor, etc.). However, in the case of the Mazeppa works, this “utilitarian” dimension is not as important as one would be inclined to think, especially in relation to Works 4 and 5. Choosing Hugo could not provide any social advantage to Liszt, who was living in a German milieu, far from the one of his French friend who was in exile in Jersey and with whom he had very little contact at the time (Samson, 2006). We can thus deduce that the persistent will to take the Hugolian version of the Mazeppa story as the only program for his music must have been motivated by reasons related to the poem rather than to its author or to the story of Mazeppa itself.
Rewriting and metareference: When the musical symbol represents the story of its creation
From this investigation of the role of Hugo’s poem in the genesis of the five Mazeppas works we can infer that, at least, the three works with an explicit mention of Mazeppa (Works 3 to 5) have two distinct and equally important referents: the story of Mazeppa in general and its particular version instantiated in Hugo’s poem. There is however a significant difference between these two referents: the story of Mazeppa is only a referent of Liszt’s Works 3 to 5 whereas Hugo’s poem is both a referent of Liszt’s Mazeppa Works 3 to 5 and a verbal representation that refers to the story of Mazeppa. We call metarefent a symbol, such as Hugo’s poem, having this double status in relation with another symbol and we call metareference the relation that links it with the other symbol. In a more formal definition, a symbol A is a metareferent of a symbol B if, and only if, B refers as much to A as to the referent of A.
From this perspective, if we admit that Liszt’s communicative intention is not reducible to the intention of creating a semantic association with the story of Mazeppa, the fact that Hugo’s poem is a metareferent of Liszt’s works in the same way that the story of Mazeppa is a referent of these works implies that Liszt’s communicative intention is not to create a semantic association with Hugo’s poem either. Indeed, in both cases (the semantic association with the story of Mazeppa and the semantic association with Hugo’s poem), the association is already exhibited through the binary device [program + music] in the virtuoso works as well as in the symphonic ones. Moreover, we have seen, in the previous section, that the device is semantically redundant. Thus, the redundant association it contains couldn’t have been the main reason that motivated the composer to integrate a metareferent.
Similar phenomena occur in verbal communication in which a speaker has not only the semantic intention that the addressee associates the words and the utterances that contain them with their referent without any other pragmatic effect than this association. For example, when John asserts to Paul “The cat is on the mat”, what he does is not reducible to the fact of inciting Paul to associate the meaning of the words and utterance of his assertion to a certain referent (a certain cat, a certain mat and their relation syntactically expressed, or any other referential interpretation if we consider metaphorical, ironical and other non-literal meanings). By asserting something to Paul, John does in fact two things: (1) he presupposes a shared conversational background with Paul, i.e. sets of propositions about the world (here, propositions about cats and mats) they mutually accept as true and that enables John that his assertion has a chance to be understood by Paul; (2) he commits himself about a certain part of the world by expressing publicly an epistemic mental state he has about this part of the world (here, the belief that what he thinks about this part of the world is true, i.e. that the cat is truly on the mat).
When Paul understands John’s assertion, what he understands is not only a (true or false) semantic association between John’s words and the corresponding part of the world. The assertion is also the public expression of John’s epistemic mental state about this part of the world. As such, the assertion gives Paul a reason to believe what John says about this part of the world. In other words, the assertion aims to arouse in Paul’s mind a more complex, “relational” and possibly persuasive mental state about this part of the world instead of a strict semantic association. In pragmatic terms (see Austin, 1962; Kissine, 2013; Searle, 1969, 1983), this linguistic phenomenon is called a speech act or illocutionary act. However, except for the general notion of “act” used in the hypothesis of virtuoso performance conceived as a discourse act, we prefer to be careful and not use this specific terminology here (HaCohen & Wagner, 1997, pp. 449, 470, make the parallel between Wagnerian leitmotivs and speech acts but their justifications are quite problematic).
The question is thus to know what semantic function, other than an associative one, the metareferent, in our case Hugo’s poem, fulfills from the pragmatic perspective of musical communication. Between the long genesis of the Mazeppian works and their final result available to the listener in the form of virtuoso piano or symphonic works accompanied by their paratextual program, two elements can clearly be perceived by a listener once they have publicly appeared in Work 3: the semantic association between the music and the story of Mazeppa and the intentional dimension of this association (i.e. the fact that the binary device [program + music] makes it clear to the listener that this association comes from an intention of the composer/performer). By contrast, one element profoundly changed in parallel with the changes in the paratext: the mental representation that the listener is likely to have about the “story” of this semantic association, about its successive causes, its raison d’être, the sequence of events that gave the composer/performer reasons to have this association in his/her mind and to have the intention to communicate it.
Suppose that, instead of Hugo’s poem, Liszt had chosen as metareferential program an anthology of the Mazeppian works that inspired him or a chronological table, similar to Table 1 in this article, describing the events that led him to associate intentionally his music with the story of Mazeppa. In these two cases, such an “empirical” program would have better matched with the factual genesis of these works. It would have enabled listeners to imagine much more precisely the personal experiences, reflections and encounters that constitute what Liszt called the “affective content” (Liszt, 1855/2005, p. 62) of his Mazeppian works and to associate them with his music. Nevertheless, this sort of program, consistent with the factual truth of the genesis of the work, would have led listeners to form a representation of the causal story of this semantic association as a sequence of contingent and complex events, a sequence only subsequently (or a posteriori) justifiable (as the kind of information about the composer’s life or the work’s genesis and reception that we find in many actual concert programs).
Liszt, like Wagner, wanted to use his music and other arts to communicate contents to his audience. Moreover, through the addition of a paratext, he aimed to create a situation in which the act of musical communication itself persuades his audience of the fact that both the communicative intention and the content of this act were cognitively relevant (for more details about the pragmatic notion of relevance, see Sperber and Wilson, 1995; on its application in music see Downes, 1995 and Stachó, in press). Knowing that a communicated content is more persuasive when it seems more obvious, or “natural”, to its receiver, the fact that Liszt chose a metareferent consistent with the factual truth, but complex and contingent, is quite noteworthy as such a metareferent could have reduced or even cancelled the feeling of obviousness of the semantic association. Consequently, in order to elicit from the listener the feeling of obviousness of the semantic association (which is considered as a part of the intentional content of the communicative act), the story of the intention to make this association also needs to appear obvious and “natural” to the potential listener who is likely to represent it in his/her mind, even if it means to be less “faithful” to the factual truth about the way this intention has emerged in the composer’s mind. This is not to deny this factual truth by replacing it with a fictional and false genesis but to “stage” the factual genesis by representing it differently. In this perspective, the more the form of this representation is simplified, the more its content is likely to look persuasive and obvious. When Liszt eliminates every version of the story of Mazeppa he is familiar with, except for Hugo’s one, and when the paratext of his musical works is progressively “invaded” by the Hugolian metareference, as if it was the only one he was influenced by, he does not represent a false story of his intention to associate his music to the story of Mazeppa. He represents the true story representationally simplified, rewritten by the means of the metareferential property of the program (on the notion of “representational truth”, see Dominicy, 2011, pp. 148–154).
We call rewriting a communicative act by which an individual symbolizes, in the symbol produced for the communicative act, the causal story of the production of this symbol. By the notion of “causal story of a symbol”, we mean a notion close to the one of genesis but including cognitive entities and not only textual ones. It consists in the particular sequence of non-strictly mental events (usually called “facts” and “experiences”) and mental events (intentions, beliefs, desires, affects, memories) at the origin of the mental states that lead someone to produce a given symbol. For example, the simplified causal story of a short verbal symbol such as this sentence is, from our point of view as its producer, a sequence constituted by non-strictly mental events (here, experienced facts such as the Budapest symposium on virtuosity, the letter of submission for this special issue of Musicæ Scientiæ, the writing of the article and the revision of the manuscript) and by mental events (here, our intentions to describe certain data, express certain beliefs and hypotheses in a certain way about a certain content). When we judge that it is necessary to express publicly, i.e. within the sentence, the own causal story of this sentence, we are producing what we call a “rewriting” of this sentence. Coming back to music in this perspective, the function of a rewriting would be to create in the listener’s mind a mental representation of its causal history as simplified and obvious as possible. The main tool of a rewriting is a semantic one, namely the metareferent. Thanks to the fact that the relation between, on the one hand, the metareferent’s double semantic status (at the same time symbol and referent) and, on the other hand, the symbol of which it is the metareferent, a rewriting simulates in a very simple fashion the linear relations of either a referential chain or a causal sequence of events.
In other words, a rewriting is a type of communicative phenomenon that intentionally merges the level of the meanings of a symbol and the level of its causes within the mental representations aroused by this symbol. Hence, the function of the “extramusical” program takes place less at the level of the semantic association between the music and its referent (which, in order to be elicited in the listener’s mind, does not need the program) than at the level of the listener’s mental representation of the history of this semantic association and the intentions that justify its relevance. To use Liszt’s own terms, it is the tool of a “method” of communication adapted for instrumental music and aiming to satisfy a certain type of communicative intention: Literature itself could not represent on stage the passions whose sinuous flow should be followed from its beginning to its end in the chaos of the past. Their interest is much more concentrated on the internal processes than on the actions of the external world. Would perhaps the specifically musical symphony be better suited to such subjects? We doubt it. The conflict between its independent style and the one forced on it by the subject would affect us disagreeably, being without evident or intelligible cause. The composer would cease to conduct our imagination into the regions of an ideal common to all mankind and, without clearly indicating the particular path he wishes to choose, would only lead the listener astray. With the help of a program, however, he indicates the direction of his ideas, the point of view from which he grasps a given subject. (Liszt, 1855/1998, p. 1170. This English translation is incomplete; the first two sentences are borrowed and translated from the French translation in Liszt, 1855/2005, p. 63)
The rewriting between semantics and pragmatics: The condition of satisfaction issue
The functional role of the semantic association in the satisfaction of a rewriting has to be understood on a more abstract level, focused on the question of the representation of causality in a symbol. A particular semantic association within a rewriting is less relevant than a general conceptual association constituted by the principle of semantic association and its embodiment or “fulfilment” in a binary structure like the binary devices [non-musical program + music] or [virtuoso attitudes + music]. Such a conceptual association is indeed one of the simple and intuitive ways to represent or “mime” in an iconic way the linear structure of a sequence of causal events including a single agentive entity (here, the composer/performer) capable of using complex and multimodal modes of expression.
If we accept that what the producer of a rewriting intents to communicate is the causal story of the symbol s/he is communicating and a certain way of communicating this causal history (the question of the ways of rewriting will be handled in the next section), we can consider, independently of the effects this communication produces on the listener, that this agentive causal structure is able to correspond to any particular “fulfilment” of the binary association [program + music]. This does not mean that there are no semantic constraints at all in a particular association [program + music]. It means that these constraints are as concrete as they are weak, in the sense that they necessarily but only concern the level of general characterization of the referent. To borrow an example already used in section 2, Mazeppa’s music would certainly not be titled “Lullaby” by someone who listens to it without knowing its Lisztian paratext, but it could be titled in various ways, even very far from the cultural field of the Cossack hetman (for example: other individual or collective characters with a tragic-heroic destiny, intense and violent natural phenomena, political or moral allegories, generic characters or mythical figures whose main function is to embody values, feelings, concepts and attitudes such as honor, love, courage, sacrifice, fear, faith, victory, revolution, nation, freedom, etc. by “transforming” them in agentive entities involved in causal sequences such as myths or narratives).
The condition of satisfaction of a rewriting is thus weaker in our causalist hypothesis than in the associationist hypothesis focused on a particular semantic association: as soon as the intention to communicate an agentive causal structure has been recognized by the listener in the work and attributed to the composer/performer, the rewriting is satisfied. We think that the weakness (and thus the easiness) of the causalist condition of satisfaction of a rewriting is due to the fact that it occurs independently of two aspects of the listener’s interpretation of the work. Firstly, it occurs independently of the positive or negative judgment made about the particular semantic association that has been the explicit support of the recognition and attribution of the communicative intention of rewriting. Secondly, it occurs independently of the fact that the composer/performer had effectively such a communicative intention to produce a rewriting (in other words, it let the possibility of the listener interpreting a work as a rewriting whereas it has not been produced in that sense).
The empirical problem with such a level of independence of the condition of satisfaction of a rewriting is that it becomes difficult to find, in the judgments of the listeners, reliable cues that they recognized the communicative intention of the composer when this intention is attested with other sources. Yet in verbal language, judgments and other kinds of utterances are a relevant material to observe and measure the recognition of a speaker’s communicative intention by an addressee. For example, if I say “Yes please” to the question “Do you want some water?”, my answer is a reliable cue that I understood the speaker’s communicative intention to interact in a certain way with me about a certain part of the world, here the bottle of water, and not only his/her intention to say something about this part of the world. In music, judgments are a relevant material to observe and measure the semantic cognitive processing of a musical work (for empirical studies about semantic processing of Wagnerian leitmotivs and program music, see HaCohen & Wagner, 1997; Pérez-Sobrino, 2014 and, as previously mentioned, Antović et al., 2016). But they are more problematic to observe and measure the recognition of a communicative intention in a musical work through the semantic processing of this work. We don’t have a solution to this empirical problem. Moreover, it goes beyond the scope of this article and the limits of our competences.
Nevertheless, the existence of the communicative phenomenon we try to describe could at least be demonstrated indirectly by using the methods of sketch studies and genetic criticism. These disciplines study all types of traces of the work’s genesis, i.e. the compositional process (sketches, printed scores, letters, interviews, etc.) in order to identify the strategies developed by composers to anticipate the potential “misunderstandings” of the interpretations of their intentions (for a detailed discussion about the heuristic value of sketch studies, see Sallis, 2015). If sufficient and stable cases of such strategies and, more generally, awareness of potential attributions of intentions by listeners are found in the traces of composer’s creative processes, the frequency of such creative behaviours could be interpreted as a consequence of the phenomenon of intentions’ attribution to a musician by listeners, and thus a confirmation of its existence.
Besides the Mazeppa works, a well-documented case is the fifth piece of Schumann’s Fantasiestücke op. 12 for piano. Like the other pieces of the cycle, Schumann gave it a very vague title (In der Nacht) and he did it after its composition. Also after the composition, he confessed to Clara: “I found, to my delight, that it contained the story of ‘Hero and Leander’ in it. […] When I play [In der Nacht], I always think of this image” (Letter to Clara Schumann, 21st of April 1838, quoted in Kregor, 2015, p. 97). But he never inserted the Hero and Leander story in the paratext of In der Nacht for reasons related to the performance context of the piece and to the intentions that the listeners would have attributed to him with such a title. Kregor (2015, p. 97) synthesizes these reasons as follows: The specificity of this recollection explains the reason why Schumann kept it between him and Clara. His version was of two lovers separated, “a beautiful romantic story”. But had he called the movement “Hero and Leander”, it not only would have disrupted the programmatic / characteristic balance of the eight Fantasiestücke, but listeners might have attempted to map more contemporary versions of the story onto music. In Franz Grillparzer’s tragedy Des Meeres und der Liebe Wellen (Waves of the Ocean and of Love) from 1831, Leander dies trying to reach Hero, who in turn dies of grief. Their fate is the same in Schiller’s ballad of 1801, although here Hero drowns herself in the same sea that claimed her lover.
Episodic rewriting and genetic rewriting: An elementary typology based on the visual attention given to the performer’s attitudes
Now that we have proposed a definition of the concept of rewriting and examined some of its general characteristics, we can turn our attention to the specific way it works in the two cases related to our Lisztian corpus: virtuoso and program music. How can a strategy of musical communication such as a rewriting “guide” the listener to what Liszt calls, in the text quoted at the end of section 3, a certain “subject”, i.e. a certain intentional semantic content? Given that Liszt is using the same referent and the same metareferent in the Mazeppian virtuoso works and in the programmatic one, we can hope to understand the differences between these two types of musical works by understanding the differences between the way they are rewritten in the sense we have defined the term “rewriting”. Hence it is necessary to ask which type of causes is simplified in the rewriting. Which part of the history of the musical symbol is likely to be represented and simplified in the listener’s mind during the listening experience?
We propose the following hypothesis based on the distinction between two types of rewritings: Liszt’s virtuoso rewritings are episodic (by reference to what Tulving (1972) calls “episodic memory”, a memory storing the “episodes” of an individual’s life). They symbolize the virtuoso’s memories of his/her experience of the non-musical work denoted in the paratext. The causal history of the virtuoso piece is simplified in the form of a public representation of the history of his/her personal experience about another work functioning as a metareferent. By contrast, programmatic rewritings such as the symphonic poems are genetic (in the philological sense of the word “genesis” which relates to the sources, the influences and the relations with other symbols and artistic works). They symbolize the genesis of the musical work, i.e. the creative path followed by the composer to reach the final state of the work performed in the concert. The causal history of the symphonic poem is simplified in the form of a public representation of the history of its intertextual relations with another work functioning as a metareferent and considered as the most salient cue to understand the creative path followed by the composer.
Following Kramer (2002), Gooley (2004), Davison (2005), Moysan (2009) and Davies (2015), we favor this hypothesis because of the distinctive visual role that the performer(s) of the rewriting play(s) in the prototypical sequence of a virtuoso recital and a symphonic concert.
In the case of the symphonic concert of program music, the visual attention of the listeners is not focused on the attitudes of one musician (not even to the conductor, and even in the case of a technically very demanding symphonic work) as in a virtuoso recital. Thus the perception of the orchestral musicians’ attitudes less probably plays a causal role in the recognition of the composer’s communicative intentions than does the perception of the virtuoso’s attitudes (for 19th century examples and counter-examples of this position see Jensen, 1994 and Kaltenecker, 2010, pp. 240–246, 334–341).
Secondly, a genetic rewriting, as a communicative act, needs fewer multimodal means of expression. The composer does not need to attract the visual attention of the listeners if s/he wants to represent the creative path of the work they are listening. The juxtaposition of the musical flow and its non-musical paratext, supposedly either known in advance or quickly read at the beginning of the concert, is enough to represent, in a simplified way, the story of a multimodal and semantic association between two works supposedly similar and created by similar geniuses. The association between the musical work and its metareferent will be especially “natural” when the juxtaposing form of this association is very simple to grasp.
A third reason why the visual attention is not used in the recognition of the composer’s communicative intention to produce a genetic rewriting is the fact that the cognitive effect of the listener’s satisfactory recognition of this communicative intention intensively uses visual attention at another level than the perceptual one: an evocative and imaginative level (or, in other words, the level of mental imagery). Indeed, the specificity of a genetic rewriting is not that the visual perception of the performing musicians enables the listener to reconstruct the creative path of the composer but, rather, that the auditory perception of the symphonic poem heard and interpreted as a rewriting of this creative path causes in the listener’s mind a set of specific mental states that can be divided in two categories: on one hand, a category of “programmed” mental states constituted by the recognition of the composer’s communicative intention and, on the other hand, a category of particular or “non-programmed” mental states, constituted by the mental images, sensory-motor effects and emotional states that each listener will infer from his/her recognition of the composer’s communicative intention.
It is important to note that the distinction between “programmed” and “non-programmed” mental states involved in a rewriting does not concern directly all the mental states resulting from the cognitive processing of the musical work. It only concerns the mental states of the work as a rewriting, i.e. the mental states involved in the potential satisfaction of the communicative intention of the composer/performer. Whatever their type (percepts, concepts, emotions) and their modal origin (visual, auditive, verbal), mental states resulting from the standard cognitive processing of the work and of the program could cause the recognition of the communicative intention but they could not constitute this recognition. A similar difference between these two types of mental states appears clearly in musical quotations, a pragmatic phenomenon close to the rewriting but more simple. Consider a musical theme T1 quoting a second musical theme T2. The mental states resulting from the cognitive processing of T1 are a cause of the recognition that T1 is a quotation of T2 but it does not constitute the recognition. This recognition is constituted by other mental states such as the retrieval of T2 from long term memory, the perception of formal equivalences between T1 and T2 and the belief that the formal equivalences of T2 with T1 are not caused by chance or by easily predictable stylistic similarities but by an intention of T1’s producer to produce these specific equivalences. All other things being equal, if we compare the notion of quotation with the notion of rewriting, T1 with the rewriting’s music and T2 with its program, we see that the difference between mental states that cause the pragmatic phenomenon and mental states that constitute it, is the same in the phenomenon of quotation compared with the rewriting.
The musical communication in the episodic rewriting of a virtuoso performance
In this section, we would like to go deeper into the description of the cognitive processing of a virtuoso rewriting, with an emphasis on the function of the virtuoso skills of the performer and their multimodal perception by the listener. In a virtuoso performance, the recognition and the appreciation of the musician’s skills rely as much on the auditory perception of the music as on the visual perception of the musician. Thus his/her virtuoso attitudes have a higher probability to play a causal role in the listening experience, what seems intuitively obvious (about the importance of the visual perception of Liszt’s hands, see Davies, 2015, pp. 152–179; see also Cook, 2013, pp. 316–336), and also, in a less obvious way, in the inferences aroused by the recognition of the composer’s communicative intentions (even when the virtuoso is not him/herself the composer of the piece s/he is performing). The cognitive process underlying such a causal role of the virtuoso attitudes can be described in two steps: first, an inference following the visual perception of the virtuoso attitudes and, second, a sensory and emotional experience consisting of “reliving” by mimicry a supposed episode of the virtuoso’s mental and/or concrete life.
Let’s look first at the inferential step. We have differentiated an episodic rewriting from a genetic one by the fact that the former does not intend to communicate to the listeners a representation of the creative path of the work they are listening to, but a representation of the creator’s personal experience of this creative path. In the episodic rewriting, the juxtaposition of the music and its non-musical paratext, even if it is known in advance, is a necessary but not sufficient condition to represent, in a simplified way, the history of a multimodal semantic association as it has been experienced by the composer. Listeners also need perceptual cues that enable them to infer that the composer’s intention is to communicate not so much the particular content of this semantic association between the music and its metareferent, but rather a personal way of experiencing it. In other words, these cues should enable listeners to represent the supposed mental states of an individual about this semantic association. Yet in the situation of a virtuoso recital, such cues are essentially visual and concentrated on the person of the virtuoso itself. The metareferent, through its explicit presence in the paratext, is perceptually as salient as are the virtuoso’s attitudes during the concert. Thus the listener will tend to associate his/her knowledge of the written paratext and his/her visual perception of the virtuoso’s attitudes. S/he will do it by categorizing such attitudes as being the public expression of the mental representation that the virtuoso has of the metareferent exhibited by the paratext.
If the listener infers the existence of the virtuoso’s mental representation of the metareferent, s/he must also infer that the virtuoso necessarily knew the work of the metareferent before the concert. Because the exceptional skills of the virtuoso create a strong individuation of the emotional effects s/he gives to the audience, this mental representation will probably tend to be understood as a representation of the virtuoso’s personal experience of the metareferential work. As to the music s/he is performing, it will be perceived simultaneously as an attempt to publicly express this personal experience as being the narration of a memory or a testimony and, due to the music’s exceptional difficulty and its emotional power, as a direct and “logical” consequence of the intensity of this personal experience. It is partly for these reasons that many of Liszt’s virtuoso fantaisies, based on operas by Auber, Meyerbeer, Bellini or Donizetti are called réminiscences, a French word that the dictionary Littré defines as “the act through which one seeks to grasp an incomplete memory”. This word is closer to the idea of an intentional communication of a preexisting mental content than to the idea of a semantic association between preexisting non-mental contents (Kregor, 2010; Moysan, 2009).
In the next section, we will propose a concrete, although problematic, example of the perceptual cues that enable the listener to infer the performer’s communicative intention. Before doing that, though, we will look at the mimetic step. Given that the work used as a metareferent is supposed to be known by the listeners and that, consequently, they have the same disposition towards the reminiscence as the virtuoso on stage, the categorization of the virtuoso’s attitudes on stage as being the expression of a personal experience could, by mimicry, activate a similar process in the listener’s mind. However, contrary to the inferential step, it is not a matter of causing a set of affective and epistemic reactions from the understanding of a creative path rewritten and already past. It is a matter of experiencing personally a similar rewriting, “private” and non-musical, of the rewriting that the virtuoso is communicating, in a public and musical way, of the memories of his/her own experience of the metareferent. In this mimetic process, three characteristics of musical virtuosity are likely to play a role: an auditory characteristic (the high difficulty of the musical works and the skills required to perform them), a visual one (the fascination caused by the contemplation of the virtuoso, especially in a performing situation) and a social one (the fact of considering the emotional and gestural attitudes expressed by the virtuoso, even the most outstanding ones, as normal adaptations to the conventional position s/he has in the situation of a recital, compared to the emotional and gestural attitudes of a non-virtuoso individual facing an equivalent situation in his/her life).
Following the Aristotelian conception of mimicry, conceived as a relation of similarity rather than a literal copy of the imitated object, and the applications of this conception in the psychology of music (Cambouropoulos, 2009; Deliège, 2007; Margulis, 2014), we can argue that the above three characteristics of musical virtuosity are the triggers of the listener’s mimetic effect. As the notion of mimicry is an aesthetic equivalent of the cognitive notion of similarity, the mimetic effect that these three characteristics of virtuosity cause consists in what could be called a “segmentation of the experiential sequence of a virtuoso performance”. Auditory, visual, verbal and other modal features of this experiential sequence are segmented, grouped and categorized according to a distinction between what the listener perceives to be similar or non-similar features of his/her experience (a distinction that, in the case of the perception of musical form, Deliège [2007] calls the segmentation of the musical flow based on the principles of similarity and difference).
We suggest that the similarities concern primarily the emotional aspects of the listening experience, inducing a form of emotional identification between the listener and the virtuoso. It can be explained by the social characteristics of virtuosity described above: the particular situation in which the virtuoso’s musical communication takes place makes it possible to consider the performer’s attitudes on stage and the intense emotions s/he is expressing musically to be a particularly exceptional and powerful public example of the same types of emotions that the listener is likely to feel and share with him/her.
In turn, we suggest that the differences are primarily related to the ethical aspects of the listening experience (personal identity, agency, practical judgment, motivation, empowerment), inducing a form of individual and non-problematic distantiation between the listener and the virtuoso. It could be explained by the fact that the listener’s consciousness of his/her inability to reproduce the exceptional and powerful public representation of a certain type of shared emotions indicates a personal difference between him/her and the virtuoso. However, this is a non-problematic difference because of its “normal” or “natural” character in the specific context of a recital (or in a context which is able to be experienced as similar to a recital). If one would put into a stereotyped phrase of a listener what is like to “feel” the mimetic effect of an episodic rewriting in the case of a virtuoso performance, it could be “What I feel is like what s/he feels even if what I am is not like what s/he is” or “I feel what s/he feels even if I am not like him/her” (for a very close hypothesis about virtuosity elaborated in an important anthropological theory of art, see Gell, 1998, pp. 71–72). In summary, the experiential mimetic effect produced by the virtuoso is close to the concept of “exemplary exception” through which many historians have described the notion of “genius” since the Antiquity (Mortier, 1982; Perras, 2015; Zilsel, 1926/1993).
Like the evocative and imaginative states aroused by the genetic rewriting, the mimetic experience aroused by an episodic rewriting and, above all, its visual dimensions, arouses mental states that fall into two categories. On one hand, collective or shared mental states are “programmed” by the intentional similarity established between the extraordinary personal experience, expressed by the virtuoso, and the ordinary personal experience “relived” by the listener; on the other hand, individual or “non-programmed” mental states result from the difference of the modes of public expression of this personal experience (musical or non-musical) and from the previous relation with the metareferent. Finally, if further studies would confirm (1) that this mimetic experience is caused by the categorization of the performer’s virtuoso attitudes as being the public expression of his/her experience with the metareferent; (2) that this categorization is an inference caused by the perception of the binary device [metarefent/program + music], then the occurrence of the mimetic experience would satisfy the communicative intention of the virtuoso and, therefore, the act of episodic rewriting.
Empirical illustration and the issue of testability
There would be two complementary ways of testing our rewriting model of a virtuoso performance: an experimental test with contemporary subjects and a historical test with past sources. Because they go beyond our field of competence, we will not detail the issues concerning experimental tests with contemporary subjects. For a theoretical and empirical research, close in its aims to the one presented in this article, and based on the GERMS model of communication in music performance developed by Juslin (2003) and Juslin and Timmers (2010), see Stachó (2012, in press). We limit our comment to a general point of comparison. We aim to bring to light what seems to be, ultimately, a strong and stimulating philosophical obstacle to the possibility of testing our model in the experimental framework of the psychology of music.
Although simpler and narrower in its scope compared to Juslin’s psychological theory of communication in musical performance, the rewriting model agrees with some fundamental claims of Juslin’s theory, which makes it, at least, conceivable to test it within this experimental framework: (1) “performance expression [in music] is […] a multi-dimensional phenomenon that can be decomposed into subcomponents that make distinct contributions to the aesthetic impact of a performance” (Juslin & Timmers, 2010, p. 454); (2) the multi-dimensional subcomponents causing the impact of a musical performance on a listener are psychological mechanisms that are not unique to music (Juslin & Västfjäll, 2008); (3) within the process of musical communication between a performer and a listener, there is a crucial distinction between information perception and information causation. Information perception refers to a kind of information (emotion, stylistic feature, visual imagery, conceptual content) that the listener perceives in the musical performance, for instance, in the way the performer is playing and/or in his/her body attitudes. Information causation refers to a kind of information that is caused in the listener’s mind/body by the musical performance (Sloboda & Juslin, 2010). The information causation (or induction) corresponds to what the listener “feels” and how s/he reacts (emotionally and/or hermeneutically) to the communicational act. Our rewriting model tries to articulate these two dimensions of musical communication. The inferential step belongs to information perception whereas the mimetic step belongs to information causation.
By contrast, the rewriting model differs from Juslin’s theory on one essential point directly connected to the question of its testability within an experimental framework: we don’t aim to describe a general or common way by which a virtuoso performance communicates and is listened to. We aim to describe a particular way by which a virtuoso musical performance could communicate and be listened to, consecutively to a particular way by which this virtuoso music has been produced and performed. The word “particular” means here “historically contingent” rather than “subjectively determined” or “culturally determined”. In other words, the rewriting model is located neither at a purely observational level (how music is listened to, e.g., in the culturally determined everyday life of an individual, or an aesthetic tendency endorsed by a group of individuals) nor at a normative level (how music should be listened to, e.g., according to some concrete aesthetic models and/or aesthetic theories) but at a dispositional level (how music could be listened to, among the existent possibilities within the general empirical constraints of human bodies and minds). Juslin (2013, p. 283) makes the distinction between these three modal categories of empirical objects. However, he groups the dispositional one (the could) with the normative one (the should) and he opposes both of them to the observational one (the is). In other words, he adopts a different conception of empiricism than the one we adopt. Here we hit on the philosophical obstacle mentioned above.
In the field of humanities, what we call the dispositional level of objects and phenomena is not a strict equivalent of the logical concept of possibility. It refers to a contextual and empirical dimension of objects and phenomena, provided that the words “contextual” and “empirical” are understood in the contingent sense assumed by an historian or a philologist rather than in the subjective, cognitive and/or cultural sense assumed by a psychologist. As we have tried to do in this article, such a dispositional level of objects and phenomena can be described with observational data (such as scores, genetic processes, testimonies and other kinds of data available because the disposition of listening to virtuoso music in this certain way has been concretely realized during at least one period of history) and what could be called “normative data”. By this expression, we mean a set of devices increasing the probability that a certain potential way of listening to virtuoso music (a particular “disposition” to listen to virtuoso music) will be actualized during a certain duration. These devices can be formal ones (e.g., the [program + music] device) but also socio-institutional ones (e.g., the standard classical concert device, the system of copyrights), and ideological ones (e.g., the Romantic conception of art as a substitute for religion, the meritocratic conception of social life in which conspicuous talent and genius are criteria for belonging to the elite).
This actualization of a dispositional way of performing virtuoso music can be explained as a consequence of the fact that, at a particular moment and for reasons that are neither entirely subjective nor entirely cultural, (1) some people (here, Liszt) thought that it was good to actualize such a way of performing and listening to virtuoso music, (2) this way of listening belongs to the possibilities provided within the general empirical constraints of listening to a musical performance.
Because of this philosophical opposition between two conceptions of empiricism and modality, an opposition probably not insuperable but still quite resistant, the rewriting model would be at first more easily testable with historical material. Such a historical test would consist in observational research that can be described as follows: (1) collecting testimonies written by people that have attended Liszt’s virtuoso performances of his pianistic program music; (2) analyzing the testimonies to observe if, beyond their individual characteristics, the authors of these written testimonies describe their listening experience of Liszt’s pianistic fantasies and réminiscences in a way that corresponds to the way we describe the cognitive processing of an episodic rewriting. Within the limitations of this article, we will give only one example of such a testimony (a similar example about the Réminiscences de Don Juan has been analyzed by Gooley, 2004, pp. 28–29, 42–48). It comes from a review published in the Diario de Barcelona by the Spanish critic Antonio Fargas y Soler, after a concert performed by Liszt on the 7th of April 1845 in Barcelona (the pieces evoked are the piano transcription of Rossini’s Guillaume Tell overture and the Fantaisie sur des motifs de la Sonnambula, based on Bellini’s opera): If Rossini achieved to translate the great poem of Schiller in the dignified and admirable way of his own language, we might say that maybe Liszt is the only one that has interpreted the short but no less admirable digest created by that composer. In the fantaisie on motives from la Sonnambula, Mr Liszt exhibited a miracle of strength and hability against the difficulties he had overcome; a strong and wide melody played with the fourth and fifth fingers of the right hand stood out against the luminous and tender cantabile of Bellini . […] The prodigious performance of Mr Liszt cannot be explained, because even if one sees it, it cannot be understood, unless if one is being stunned and in ecstasy. […] He never usurped less the justified name of king of pianists when we contemplate his look during the wonderful performance; because if in the steps of expression and feeling he was bringing us tears of tenderness, in spite of the smile dawning in his lips, in the moments of bravura it was worth seeing how such a bright and serene forehead was flooded by artistic passion and enthusiasm whose electric fluid was spreading through the extremity of his fingers, torturing those who were listening and watching it. (Fargas y Soler, 1845, Diario de Barcelona, n° 100, pp. 1369–1370, translated from Spanish, quoted in Seibold, 2017, pp. 157–158)
Although the inferential and mimetic steps of the rewriting model are observable in this testimony, which also contains an emphasis on the visual component of the rewriting and some precise examples of the corresponding perceptual cues (fingers, smile, body gestures), such sources should be taken warily when used as a test material. The example we have selected is a “good example”. It concentrates in a few lines what an example has to exemplify. Among the thousands of reviews of Liszt’s performances published during his life, few of them articulates so clearly the role of the visual perception of Liszt’s attitudes on stage and the role of the listener’s knowledge of the non-musical program. These two aspects were often treated separately, as is the case in the research on Lisztian virtuosity that we have used for this article (with the notable exception of Moysan, 2009). The most striking example is the complete partitioning between Kramer (2002), analyzing Liszt’s visual impact on his audience, and Kramer (2016), analyzing the impact of Liszt’s literary programs on his audience—the same audience that saw Liszt and that Kramer had studied in his first article.
Unlike data extracted from questionnaires, surveys and laboratory environments in experimental psychology or qualitative social sciences, historical testimonies have not been produced to help researchers study psychological mechanisms, even when it is obvious that they describe them. Because this anteriority is an empirical constraint, it is also a methodological advantage in the sense that, taken as data for testing a hypothesis, they precede the hypothesis one wants to test with them. Moreover, like corpus-based studies in linguistics, they reduce partly the risk of confirmation bias if the hypothesis is not elaborated mainly on their basis. Nevertheless, it would be wise not to use them as test material in a purely statistical and literal way, such as by focusing on the frequency of certain keywords or topics. Such historical sources need to be sampled after having established a detailed hypothesis about the phenomenon we want to test. Then the sources have to be carefully contextualized with the tools of philology, cultural and intellectual history. Finally, they can be used to test the hypothesis with the tools of textual analysis and, if they are materially accessible and usable by non-professional digital humanists, with the tools of text mining and digital archives.
One of the main inconveniences of the historical test is not to find comparative sources but to ensure the possibility to control them, that is, to be able to include in the test similar historical documents that can function as a case-control for the documents used to test the hypothesis. For example, even though they were culturally contextualized and textually analyzed by Gooley (2004, pp. 18–77), historical testimonies about Thalberg’s performances of his own rewritings, based on famous operas similar as did Liszt for his own rewritings, could not be used as a case-control to test the validity of the model developed for Liszt’s rewritings.
As suggested by Nattiez (2013, p. 3132), a possible solution for this methodological problem can consist in applying a double historical test to the target case (Liszt) and the control case (Thalberg) of a given hypothesis: the first one based on the analysis of poietic historical sources (the composer’s writings, description of the creative processes, analysis of the genetic material) and the second one based on the analysis of esthesic historical sources (testimonies about the performances, like the one reproduced above). By comparison, these two tests could be qualified as a semiotic and historical equivalent of what experimental scientists call retrospective study and case-control study, respectively. If the results of the poietic and esthesic tests turn out to be similar, it can reinforce the probability that the hypothetical model has an explicative value for the communicational phenomenon whose poietic and esthesic sources are two types of traces. More generally, we can hope that the kind of musicological issues studied here and the methodological problems they arouse could stimulate convergence of interests and concrete collaborations, hopefully not as difficult to reach as one could think at first sight, between psychology of music, historical musicology and traditional humanities.
Concluding remarks
As we were completing this article, linguist Philippe Schlenker published in Music Perception a vast programmatic article aiming to establish the foundations of a new formal theory of music semantics (Schlenker, 2017; for a longer version see Schlenker, in press). The main idea of Schlenker is that music semantics, like the semantics of other symbolic systems such as images, natural or sign languages, is primarily referential. This means that music fully allows someone to communicate intentionally something about the external world. Two immediate consequences of this postulate are that it should be possible to evaluate, to a certain extent, whether what the music says about the external world is true or false and, in a pragmatic perspective, to evaluate the interface between the meanings of a music, the contextual factors and the practical effects of the cognitive processing of these meanings and their potential intentional nature.
In this framework, Schlenker’s working hypothesis is that the meaning of a musical work does not reside in its form or in the mental representation of its form but in the inferences made by the listener about the source, real or virtual, of a certain number of formal features of the musical percept. Influenced by the principles of Gestaltpsychologie and by the classical work of Lerdahl and Jackendoff (1983), Schlenker proposes to demonstrate that it is possible to formalize his hypothesis by elaborating a structured set of rules derived from the cognitive processing of the general features of any type of sound (e.g., intensity, speed, timbre) and those of a given musical language (e.g., the tonal language). This structured set of rules is thought to be able to describe how, from the perceived structure of sonic events, the listener infers a possible sequence of the real or virtual causes of these events. In Schlenker’s view, the inferred cause of the sonic event is (or constitutes) what the event refers to. This relation of reference manifests itself by a certain form of correspondence, or similarity, between the structure of the two sequences of events. Because this causal sequence of events to which the music refers to is a supposed ensemble of states of the world, or “facts”, either real (e.g., the genesis of the work, the mental states of the composer) or fictional (e.g., the mental states attributed to a character such as Mazeppa), this ensemble of states of the world constitutes the meaning of the music. It is what the music “tells” us about the external world and about which it is possible to evaluate the music’s truth, its relevance and/or its power of persuasion.
Schlenker’s causalist hypothesis is quite close to our notion of rewriting conceived as a discourse act producing a representation of the causes of the symbol by which the discourse act is done. As such, this article can be interpreted as an application of Schlenker’s theory to a particular aspect of music semantics, i.e. the interface with pragmatics and the issue of communication, and to two particular cases of musical performance: Liszt’s virtuosity and program music (for similar approaches, see Gell, 1998 and Hatten, 2004, pp. 217–234). However, it is clear that the causalist hypothesis elaborated in this article and in Schlenker’s one raise at least three problems:
How can we elaborate empirical procedures allowing to isolate and manipulate, in documents about musical listenings or during musical listenings, mental states such as recognition of communicative intentions and inferences on the causal history of the sonic percept? This point had been addressed in the last section.
What are the past elaborations of this causalist hypothesis in science, in art and in philosophy? (for some insight outside the scientific field, see Corbin, 1994, pp. 459–467 and Kaltenecker, 2010, pp. 197–207). From the specific point of view of Liszt’s compositional evolution, what are the intellectual and contextual reasons that could explain why he changed from episodic rewritings to genetic ones in his career, i.e. from virtuoso piano music to orchestral program music? (for an hypothesis in the framework of the history of public concert and the evolution of the musical work concept, see Samson, 2003 and Hamilton, 2008).
The conceptual framework of our causalist hypothesis should be confronted with the classical objections addressed to similar musicological and philosophical hypotheses about meaning and reference. We think in particular of the theory of causal reference by Kripke (1981) and Putnam (1975) and of the iconic or “similarity-based” conception of the relation between the structure of a language and the structure of the world defended by Wittgenstein (1922/1993), but going back to the ancient idea of Ut pictura poesis. With regard to the question of the sonic source’s vision and its causal role in music perception, we think in particular of the concept of acousmatic listening in which the listener does not see the source of the sounds. This listening situation lets open the mental reconstruction of a tangible source called a “sound body” or “sound object”. The reconstruction process is executed by way of informations extracted from the listening environment of the sonic percept (Schaeffer, 1966; Kane, 2014).
We did not want to overcome these problems concerning musical communication but rather to show in which way virtuosity can be a part of them. More generally, we tried to illustrate the importance of making the distinction between the performative dimension of music (here embodied by virtuosity) and the pragmatic dimension of music (here embodied by communication) with the semantic or “meaning problem” located at their interface. The fact that Liszt himself consciously constructed both his relation to performance and virtuosity on one hand, and the justification of his evolution to orchestral program music on the other hand, in a semantic and communicational perspective rather than an exclusively performative one, could indicate to us that escaping from the “Liszt problem” in the study of virtuosity is not as important as Deaville thinks, precisely because it is not the problem.
Footnotes
Acknowledgements
We would like to thank László Stachó, Michael Walsh, Marc Dominicy, Manuel Couvreur and the two anonymous referees for their insightful comments.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.