In 1849, having perused an album of miniatures Franz Liszt had written for the Goethe centenary, Richard Wagner lost no time in encouraging his friend to take on a different sort of project: ‘What I felt most vividly, after my acquaintance with these compositions, was the desire to know that you were writing an opera or finishing one already begun’. This urging of Liszt to write for the stage is unsurprising, given the nature of Wagner’s own output, as is his belief that bigger is better: ‘Creative power in music surely requires this stimulus no less than does any other great artistic power; a great incitement alone can make it effective.’Footnote 1 However, the phrase ‘finishing [an opera] already begun’ suggests that he was not only aware of Liszt’s general operatic ambitions, but he knew that the Hungarian was actively pursuing these. Liszt’s response to his exiled colleague confirms as much: ‘In the course of the summer my Sardanapale (Italian) will be completely finished’.Footnote 2
Except, of course, it never was finished: Sardanapalo (to give it its Italian title) was quietly abandoned in the early 1850s, seemingly just another entry in the substantial list of abortive operatic projects Liszt at one time contemplated. It had in fact been pursued more thoroughly and over a longer period than most of these other plans, with Byron’s tragedy Sardanapalus first mooted as subject matter for an opera in late 1845. When Liszt entertained hopes of succeeding Donizetti as Kapellmeister in Vienna the following year, he envisaged Sardanapalo serving as supporting material for his candidacy.Footnote 3 However, it was only in August 1848, when he was settled in Weimar, that he was provided with an Italian-language libretto by an unknown poet sourced by his correspondent, Princess Cristina Belgiojoso.Footnote 4 The most authoritative account of the compositional process estimates that ‘the music for Act 1 which Liszt did compose was almost certainly notated between April 1850 and February 1851’.Footnote 5 Within a few years, his ambitions as a composer were redirected, with the orchestra, rather than the stage, becoming his principal focus. One bellwether for this changed direction of his creative energies is a letter he sent six months after the above declaration to Wagner, in which we learn that alongside working on the opera he was also striking out in a new direction:
I am applying myself well to Sardanapale (Italian text in 3 Acts), which ought to be completed by the end of the year, and in the intervals, I am finishing off some of the symphonic works of which I am undertaking a series that will only be ready in its entirety in two or three years.Footnote 6
This ‘series’ of symphonic works refers of course to the symphonic poems, of which he would eventually write a dozen while in Weimar. Over the course of his thirteen years in the city, Liszt the composer would leave his mark on many large genres – the piano cycle, oratorio and, most importantly, programmatic orchestral music – but not opera. As to why work on Sardanapalo was abandoned, many theories have been advanced: these include the jealousy of his mistress Carolyne von Sayn-Wittgenstein, since this opera project had been associated both with Liszt’s former flame (and mother of his three children), Marie d’Agoult, and with Belgiojoso, another possible romantic interest; Liszt’s wish to avoid rivalry with Wagner on the latter’s turf; the ineptitude of the libretto;Footnote 7 or more simply, the delay in receiving the revised libretto of Acts 2 and 3.Footnote 8 The incomplete sketches for this opera languished in the Goethe- und Schiller-Archiv in Weimar, their existence known to specialists, but serving as little more than further evidence of Liszt’s creative uncertainty as he transitioned from performance to composition.
However, more work had been done on Sardanapalo before it was abandoned than most scholars realized. In 2017, David Trippett, a musicologist from the University of Cambridge, studied the extant materials and determined that the 115 pages of piano-vocal score constituted an almost continuous draft for the first act of the planned three-act opera. He reconstructed the music from Liszt’s handwriting, where necessary amplifying the harmonies and patching over the missing portions, and then provided his own orchestral realization. Performances of a few excerpts were released online as teasers in a carefully coordinated publicity campaign, and the entire Act 1 was premiered to much fanfare on 19 August 2018.Footnote 9 A recording with the same artists was released in 2019,Footnote 10 a critical edition of the piano-vocal score followed,Footnote 11 and an orchestrated version was made available for hire.Footnote 12 Further performances have followed, among them the work’s U.S. premiere.Footnote 13
The excitement attendant on this discovery, which one journalist claimed ‘changes music history’, has not yet translated into a proper scholarly assessment of the Act, aside from Trippett’s own account of the gestation and reconstruction of the work, which also analyses certain stylistic features.Footnote 14 The present review essay will provide an overview of the new editions followed by some reflections on the musical style of the reconstructed Act I, situating it in the context of Liszt’s creative career and of nineteenth-century practices more broadly. Why Liszt, despite his lifelong fascination with opera as a genre, should never have completed anything for the stage after Don Sanche (1825) is a complex matter for which there is no single explanation. However, when it comes to Sardanapalo specifically, I will suggest that certain qualities of the source material were incompatible with the nature of opera as Liszt himself would come to understand it.
Reconstructing the Assyrian drama
The critical edition of Sardanapalo published by Editio Musica Budapest is subtitled ‘Fragment’. While not untrue (we are, after all, dealing with only the first act of a three-act work), it undersells what Trippett and his collaborators have actually provided. This is no mere diplomatic transcription of Liszt’s sketches, but rather a fully realized piano-vocal score. True, some of the textures are technically unplayable as they stand but there is nothing here that a halfway skilled repetiteur couldn’t navigate with judicious practical adjustments. Liszt was notating with an eye to the eventual orchestral arrangement, and so writing an idiomatic piano part was a secondary consideration, particularly as this was a composition sketch intended for his use only. This explains some of the uncharacteristically awkward and unpianistic figuration, such as repeated note tremolos (e.g. bars 297–306), impossible alternating thirds (bar 1 in the LH), and unplayable large stretches (bars 187–95), as well as oddities of notation like the occasional use of the tenor and alto clefs (deployed successively in bars 426–33). There are also places where the decay of sound inherent to the piano has not been compensated for by other means: for instance, the held notes in bars 819–24 sound far better in the brass and woodwind on the recording than they do on the piano. In the manuscript from which Trippett was working (the autograph held at the Goethe- und Schiller-Archiv in Weimar, Shelfmark 60/N4), there are occasional indications of the instrumentation Liszt had in mind; these have been retained in this edition.Footnote 15
Most importantly, there are no gaps: Act I can be played through continuously from the beginning to where Liszt left off (more on this below). This has necessitated considerable editorial intervention, as the autograph is ‘skeletal at points’, and the composer used ‘frequent shorthand’ that needed decipherment.Footnote 16 Places where Liszt left empty bars or incomplete figuration have been filled in by analogy with earlier textures. Every alteration to the text in the autograph, from added marks of articulation up to passages that needed complete harmonic reconstruction, has been exhaustively catalogued in the critical notes.Footnote 17 Less frequently, there are variant readings in the manuscript. Some of these record places where Liszt changed his mind: the discarded version may be provided in an ossia staff (e.g. bars 15–17), or in a footnote (bars 1015–1018); the original version of bars 851–72, too long for either of these solutions, is provided in Appendix A. In other instances (such as the cadenza in bar 311), Liszt leaves two or more equally viable options. Trippett acknowledges that ‘a different scholar completing the same editorial task may have produced a slightly different result’, something which is particularly true when it comes to the larger gaps in the piano part (bars 700–51 and 786–803; bars 1082–86 and 1089–93; bars 1103–61).Footnote 18
In fact, Trippett’s realizations have themselves evolved over the course of the project: for instance, bars 704–30 were reproduced in his 2018 article, and there are multiple tiny divergences in both the piano figuration and even the harmony between the article example and the critical edition (see Example 1). There are also several differences of varying significance between the critical edition and the orchestral score. In bar 28 of the latter, Trippett has gone with Liszt’s discarded first thoughts (whereby the appoggiatura d5 is resolved before the bar line instead of being suspended into bar 29), a reading which sacrifices the parallelism with bar 24 for a reduction in dissonance. Liszt’s shorthand instruction ‘weg’ [dying away] in bar 56 is interpreted in the critical edition as a continuation of the quaver pattern from the previous bar, but in the orchestral score as a decrescendo across a high violin trill. Bars 217–18 in the critical edition have an added doubling in 3rds in the alto voice, but these notes are not found in the orchestral score, which instead opts to repeat the undoubled inner-line found in both editions in bars 215–16 (see Example 2).
Varying realizations of Sardanapalo bars 722–5: (a) Article;Footnote 19 (b) Critical edition
Realization of Liszt, Sardanapalo bars 216–19: (a) Critical edition; (b) Orchestral edition (reduction)
In the matter of the orchestration, Trippett proves to be a worthy successor to August Conradi and Joachim Raff, who provided initial orchestral versions of Liszt’s works in the early 1850s, even if the final versions were heavily reworked by Liszt himself. The few instrumental clues Liszt provided in his sketches are generally adhered to: ‘ohne pizz’ in bar 37 is interpreted a trifle loosely as ‘without strings’ rather than simply cancelling the previous pizzicato instruction, but the woodwind-only colouring Trippett deploys until bar 43 is very pleasing (and perhaps a nod to the deliberately naive sound world Liszt would evoke in parts of his later oratorios Die Legende von der heiligen Elisabeth and Christus). Other nice touches include the solo violin cantilena over a bed of sustained strings and harp arpeggios accompanying Mirra’s intervention in scene 4 (bars 986 ff), and the prominence of the triangle in the final bars, a call-back to Liszt’s notorious fondness for this instrument (most memorable in his Piano Concerto no. 1). In some places the colours are a trifle crude, but in an echt-Lisztian manner: the added off-beat cymbal crashes in bars 240–3, for instance; or the high-camp harp glissando when the two lovers finally unite to sing the main cabaletta theme in bar 786.
The only lacuna that remains in the critical edition concerns the very ending, as the sketch stops at bar 1256, during the play-out for scene 4. We have to turn to the orchestral performance score published by Schott to find the concluding bars Trippett added so as to enable a complete performance of this first Act. Even these bars 1257–75 were not freely composed, but are heavily based on material from the instrumental conclusion to the scene 3 duet (bars 804–11). Presumably the fact that this ending is entirely conjectural rather than an amplification of Liszt’s existing notation is the reason why this passage was not supplied in the critical score. Still, it might have been usefully included in an appendix, particularly as Schott does not offer a piano reduction as part of (or indeed separate from) the orchestral score.Footnote 20
Matters were just as challenging for Marco Beghelli when it came to reconstructing the libretto text of the unknown Italian poet, the manuscript of which has not survived. Liszt’s command of Italian was demonstrably shaky and there are multiple grammatical and syntactic errors in the text underlay, at times requiring conjectural readings, as well as places where words or phrases were not transcribed in his sketch. It is worth noting that even when we can be sure of Liszt’s intentions, the text setting may sound clunky: the last few syllables in Example 3 feel shoehorned into a pre-formed musical conception (Liszt’s first version of this passage was still more unsingable, although it would have been perfectly idiomatic in a piano paraphrase).Footnote 21 Appendix B of the critical score provides the reconstructed 191 lines of Italian poetry alongside English, German, and Hungarian translations (the last an addition to the trilingual CD liner notes, presumably to conform to the house style of Editio Musica Budapest).
Liszt, Sardanapalo bars 579–80
In this reconstructed form, the layout of Liszt’s 52-minute Act is as follows: a brief Prelude leads directly into scene 1, featuring a chorus of concubines who sing of voluptuous pleasures and acknowledge the pre-eminence among them of the King’s favourite, Mirra (Myrrha in Byron’s drama). They leave, and scene 2 constitutes a multipartite solo aria for Mirra, in which she reveals she has three traits in common with many other Byronic heroines: she’s Greek, a slave, and submissively devoted to her lover (in the play, she proclaims to him: ‘I have no happiness / Save in beholding thine’Footnote 22 ). And yet, despite being sure of his love, this solo scene turns on her unhappiness: she is torn between longing for her lost Ionian homeland, and a complicated mixture of emotions towards Saradanapalo summed up by the ‘ansio rapimento’ (anxious ecstasy) she felt on first meeting him.
In scene 3, her royal master enters, and a rather static love duet ensues. In the play, Sardanapalus is a self-confessed hedonist (‘Eat, drink, and love; the rest’s not worth a fillip’, 20) who embraces pacifism and resists comparisons to his more bellicose ancestors, Nimrod and Semiramis. His refusal to wage war abroad and root out opposition at home is raised to the level of a philosophy (‘Must I consume my life—this little life— / In guarding against all may make it less? / It is not worth so much! It were to die /Before my hour, to live in dread of death, /Tracing revolt: suspecting all about me’, 27), but shorn of the sweep of Byron’s verse, the operatic Sardanapalo appears as little more than a besotted lover.Footnote 23 Even when, in response to his question if she loves him, Mirra responds ‘Would that I could not! Heaven has sealed my fate’, the heedless monarch entirely ignores her interior struggle.Footnote 24 Further aspects of Mirra’s emotional quandary are revealed here: the ‘shame and grief’ attendant upon their ‘ill-fated flame’ arise from the disdain of courtiers and the jealous rage of Sardanapalo’s neglected wife (an important figure later in the play, but otherwise nowhere alluded to in the operatic first act).Footnote 25
In the equivalent colloquy in the play, the two reveal contrasting attitudes to governance (MYRRHA: ‘for a king, / ‘Tis sometimes better to be fear’d than loved’; SARDANAPALUS: ‘And I have never sought but for the last’, 34), but political issues only emerge in scene 4 of the opera with the entry of a new character, Beleso. This priest and elder statesman angrily reproaches the king for indulging himself in wanton pleasures when an uprising led by Arbace is breaking out, although in the play, Beleses is himself one of the main motivators behind the rebellion of Arbaces (Sardanapalus calls him the ‘master mover of his warlike puppet’, 165). In the opera, he effectively takes the place of the King’s brother-in-law Salamenes, a military commander torn between loyalty to the King and disapproval of his lifestyle. In response to Beleso’s advocacy of swift action to put down the unrest, Sardanapalo states ‘I hope to win over [the rebels] with mercy and with forgiveness’, which is clearly inadequate to the situation.Footnote 26 His pacifist credo (‘Every glory is a lie if it must be bought with the weeping of afflicted humankind’)Footnote 27 frustrates Mirra, and she intervenes, begging him to ‘show the world [his] valour’ if he loves her.Footnote 28 Her plea instantly wins over Sardanapalo, and the act ends with martial music, as the royal troops prepare to depart for battle.
Liszt’s stylistic mixture
The few scholars to have examined the music of Sardanapalo in detail have usually invoked comparisons with better-known works and styles to give a sense of what Liszt achieved here. Kenneth Hamilton refutes the ‘common description of the style of Sardanapale as a combination of Bellini and Meyerbeer’ as misleading, but concedes that Liszt uses a ‘melodic style that is undoubtedly of the Bellini-Donizetti type’.Footnote 29 In his liner notes to the CD recording, Trippett also invokes other composers’ works in an attempt to convey the ‘myriad shades of musical influence [present in the opera] – from the opening Verdian chorus and bass motif from Nabucco, to the proto-Wagnerian harmonies and Tannhäuser textures, Bellinian melodies, quasi-Palestrinian falso bordone (cf. Lamentations), and massed sonorities after Berlioz’.Footnote 30 Other comparisons are possible: for instance, the very opening phrase of the Prelude has an airiness that recalls Mendelssohn’s fantastical fairy textures, and in fact uses the exact same harmonic schema as a phrase from ‘Ye spotted snakes’ in the incidental music to A Midsummer Night’s Dream. Footnote 31 I also hear echoes of the start of Act II of Lohengrin in Beleso’s ‘O di regi d’Assiria’ (bars 851–74), a resemblance furthered by Trippett utilizing the bass clarinet here in a fashion directly comparable to Wagner. Nor was Wagner the only touchpoint for the orchestration: the playout after the conclusion of the scene 4 trio (bars 1174 ff) begins in the vein of Mussorgsky’s Night on Bald Mountain.
But whatever has been written about specific points of comparison, the idea of stylistic mélange is accurate. Hamilton sees the work as a ‘battleground between Liszt the musical radical, and Liszt the tiro [sic] Italian opera composer’, which ‘produc[es], on occasion, the effect of a work written by two different composers’.Footnote 32 The Italianate influence is particularly pronounced in the structure of the separate numbers, which draw heavily on the practices standardized in the so-called ‘Code Rossini’.Footnote 33 Mirra’s solo aria in scene 2 follows the double-aria layout, with an introductory scena (bars 277–332) preceding the andante first part of the aria (‘Giù pel piano’, bars 333–63), followed by a transitional tempo di mezzo (bars 363–452) and the concluding cabaletta (‘Ahi! nell’ansio rapimento’, bars 453–511). Similarly, the duet in scene 3 is structured around a slower first cantabile section (‘Parla! parla!’, bars 564–629) and a quicker cabaletta to finish (‘Amiam finche ne invitano’, bars 700–804). Scene 4 begins with a tempo d’attacco, in which Beleso and Sardanapalo declaim in turn while the orchestra makes two passes through a collection of chromatically inflected themes (bars 812–73, bars 894–956).Footnote 34 Mirra then sings the first truly melodic portion of this number (‘Oh perché’, bars 986–1015), in effect a solo substituting for the usual pezzo concertato. The melting appoggiaturas and lush textures (Liszt wrote unusually lavish arpeggios in his sketch here, well translated in the orchestration) all serve to mark her as an instance of the ‘oriental seductress’ trope.Footnote 35 The Italianate character of this passage is cemented by the written-out cadenza, in which Liszt borrows heavily from the best-known aria of another operatic femme fatale: Norma’s ‘Casta diva’, also in F major (see Example 4). Further exchanges among the three characters in a more declamatory vein follow, before the final stretta (‘Diletta vergine’, bars 1103–74) and playout conclude the Act.
(a) Liszt, Sardanapalo, bars 1012–15; (b) Bellini, Norma ‘Casta diva’, bars 27–30
Thus far, so conventional: the cabalette, in particular, are clearly recognizable as such. And yet, in a host of ways, Liszt departs from Italian stereotypes. Most obviously, each number transitions directly into the next without a break, bearing out the composer’s implied criticism of the formulaic conclusions of earlier opera composers (as reported by Robert Schumann): ‘Rossini and Co. always close with I remain your very humble servant’.Footnote 36 Scene 1 avoids a concluding PAC (perfect authentic cadence) in A major entirely, with the E major dominant chord repeatedly deflected onto F major in the choral peroration (bars 267–72). Scenes 2 and 3 do reach more satisfying cadential points, but the orchestral playouts in both cases extend the material into new tonal areas before the texture shifts abruptly with the entry of new characters at the start of the next scene (a horn fanfare hailing Sardanapalo in bar 525, the low brass/bassoons with Beleso in bar 814). Fusion of disparate movements into an unbroken whole is a hallmark of Liszt’s instrumental music in the 1850s—one need only think of the symphonic poems or the B minor sonata—and it is possible to see these scene transitions as another manifestation of this desire for organic connectedness.
Other idiosyncratic matters concern the recontextualizing of material when it is brought back. The scene 3 cabaletta ‘Amiam finche ne invitano’ is launched by the amorously inclined Sardanapalo, the martial C major bespeaking his confidence and single-mindedness. When an anxious Mirra takes over in verse 2, Liszt reharmonises this same melody in A minor, neatly encapsulating the disparity in how they view their situation (see Example 5). In one sense this builds on the idea of the ‘ “dissimilar” duet’ which is frequently employed by Verdi (one example which predates Sardanapalo is the Carlo-Elvira duet in Act I of Ernani [1844]).Footnote 37 However, where Verdi uses different melodies and textures (and parallel modes) to amplify the contrast between the lascivious monarch (Carlo) and unwilling woman (Elvira), Liszt’s procedure of using the same melody in a different tonal context points to a more complex relationship between his two characters.
Sardanapalo (a) bars 705–8; (b) bars 732–5 (piano textures simplified)
Another example of thematic recycling is when the lyrical theme associated with Mirra’s ‘Oh perché’ (bars 986 ff) recurs as the melody for the driving cabaletta ‘Diletta vergine’ (bars 1103 ff) in the same scene. This strategy, which in effect amounts to a simple instance of thematic transformation, was not prompted by a desire for economy alone: rather, Liszt was implying a dramatic connection between the two points, just as he would by similar means in his symphonic poems and Faust Symphony. In the present case, the shared tune was first heard in Mirra’s aside, when she inveighs against Sardanapalo’s seeming pusillanimity and resolves to change his mind; when her lover takes up this tune (now a pulsating Allegro deciso rather than the languid Adagio mesto e nobilmente of before), it confirms that he has been roused to military action solely thanks to her urgings.
When we focus in on details rather than grand designs, the disparity between Liszt and his Italian precursors grows. Trippett has drawn attention to Liszt’s increased deployment of a declamatory vocal style, associated in the Hungarian’s mind with a progressive operatic practice.Footnote 38 The musical language Liszt deploys in these more dramatic moments is appropriately forward-looking. In one case, we can see directly into the composer’s own future (see Example 6): a motif heard early in scene 2 is a pre-echo of the instrumental recitative in the Dante Symphony, later associated with Francesca’s famous lament ‘Nessun maggior dolore’. (That Trippett gave the melody to the cor anglais establishes a further sonic connection between the two passages, in which female protagonists are tormented by memories of a happier past.)
(a) Sardanapalo, bars 312–15Footnote 39 ; (b) Dante Symphony I, bars 312–318
Hamilton has noted that ‘Liszt’s tendency is to marry Italianate melody with very un-Italianate chromatic harmony, the latter usually, but not always, in the passages for orchestra alone’.Footnote 40 There is no space here to do full justice to Liszt’s chromatic musical language, which anticipates many of the features extensively catalogued in his better-known works of the 1850s, so a few illustrations will have to suffice. There are many instances of sequential modulation by symmetrical divisions of the octave (e.g. major-third cycle E–A♭–C, bars 167–83; minor-third cycle D♭–E–G–B♭, bars 836–9; tritone alternation E♭–A, bars 1221–3), as well as other less easily classified modulations (e.g. the slippage from A major to G major across bars 1–8 brought about by the relentless chromatic voice-leading in various parts). This use of surface chromaticism to destabilize the tonality in this last case anticipates what will happen several times later in scene 1, with the motif shown in Example 7a often acting as the agent of destabilization. Through its association with the concubines who are singing at this point, the motif (an elaborated chromatic wedge progression) initially serves as another marker of the oriental couleur locale, a passing effect between the stable B♭ chords that bound it fore and aft. However, the penultimate chord (a dim-seventh) is harmonically pluripotent and can be redirected so that the phrase ends a minor third higher in D♭ major (see Example 7b).
Sardanapalo (a) bars 85–87; (b) 89–91
The opera that never could be?
With all the excitement over the rediscovery of Liszt’s ‘lost opera,’ there’s an underappreciated irony in the fact that the way it has been presented in performances to date makes it more akin to a rediscovered oratorio than an opera (setting to one side the subject matter). This is not meant as a criticism; rather, the accident of it being heard only in concert mitigates the issues that would have arisen had it ever graced the stage. In a nutshell, the material in Act I comes across as essentially undramatic. This may seem an unfair verdict, given that Italian opera of the era has not always set a high bar when it comes to gripping stage action. Moreover, two thirds of Sardanapalo remain unwritten, and Liszt planned a big spectacular finale that would ‘set the audience alight’ to match the conflagration in which Sardanapalo, Mirra, and all his possessions were immolated rather than pass into the hands of the victorious rebels, a scenario famously illustrated by Delacroix.Footnote 41 But as it stands this first Act has little to offer the viewer. The opening chorus of courtesans could be staged sumptuously as a kind of Venusberg in the Orient, but thereafter the underwhelming nature of the action until the sudden decision to go to war is a big disadvantage. Nor is it easy to warm to Sardanapalo himself, either in the original play or the operatic version. At least in Byron’s version we get sonorous speeches that flesh out his reluctance to take decisive action and his hedonist philosophy, but simplified down to the level of an operatic tenor lead, he becomes paper-thin.
At this point, it is worth recalling that Byron’s Sardanapalus: A Tragedy, despite having the form of a dramatic play, was, according to its author’s preface, ‘not composed with the most remote view to the stage’.Footnote 42 Admittedly, Byron’s wish that this be treated as a closet drama has not always been observed. Murray Biggs, the director of a 1990 Yale production of Sardanapalus, has argued that despite Byron’s assertions to the contrary, his plays ‘inevitably suggest stage business even as we read them. […] Their “stage” directions alone specify une scène that is in some sense spectacular.’ Moreover, he draws attention to changing tastes in theatrical performances: whereas in Byron’s day stage plays were often occasions for grand spectacle, today’s audiences ‘are now quite accustomed to theatrical pieces in which characters talk rather than act.’Footnote 43 (It might be remarked in passing that for this 1990 production Biggs used incidental music in a fashion which largely avoided ‘an “operatic” view of the play’, although he conceded that ‘there is something operatic about its climax; I might add grandly so.’Footnote 44 )
Given Liszt’s decision to embark on the composition of Sardanapalo as an opera, he seems to have agreed with Biggs, at least initially. It was not even the first Byronic subject matter he hoped to mount on the musical stage: earlier operatic plans included Le Corsaire and Manfred. Footnote 45 And yet, even these earlier ventures had given him pause: he stopped work on Manfred after composing a chorus, acknowledging that ‘it is much harder than I thought because there is a certain monotony, and it’s difficult to change that.’Footnote 46 By 1855 at the latest (i.e. four years after he ceased work on Sardanapalo), Liszt had come around to the view that Byron’s works were inherently unstageable, and it is plausible that his experiences with Sardanapalo had a major role in effecting this change of mind. In his essay ‘Berlioz and His Harold Symphony’, Liszt stated that ‘one does violence to the stage […] when one seeks to impose constructions on it that have taken root and flowered in other fields of poetry and literature’.Footnote 47 He cited Byron’s Cain and Manfred alongside Goethe’s Faust and Mickiewicz’s Dziady as instances of a modern type of poetry which he called the philosophical epopoeia, ‘a kind of poetry unknown to antiquity and owing its existence to a characteristically modern way of feeling – the poem ordinarily written in dialogue form that adapts itself even less readily than the epos to dramatic performance’ (864). He went on to define the constituent features of this literary genre:
the action and the event lose their importance […]. It has become far more important to show what the hero thinks than how he acts, and for this reason a limited concurrence of facts suffices to demonstrate how predominantly this or that feeling affects him. […] the modern hero often typifies rare and abnormal impulses, little familiar to the human heart (865–6).Footnote 48
While such heroes can be represented in music, the operatic stage is not seen as a suitable vehicle:
Is music unsuited to cause such natures to speak its language? To represent their origin and metamorphosis, their glorious powers of ascent or downfall, their morbid outbreaks and redeeming powers, to portray their inspiring or awesome end? But could music do this in the drama? Scarcely. Literature itself cannot present upon the stage passions whose meandrine [sic] progress must be followed from their source to their disappearance in the eddies of the past. The interest which they arouse attaches itself far more to inner events than to actions related to the outer world. (867)
Needless to say, Liszt’s argument in the ‘Berlioz’ essay was crafted with the purpose of launching his own series of orchestral works, as well as defending Berlioz’s purely instrumental response to Byron’s Childe Harold. And understandably, one should at least be cautious about back-projecting his 1855 attitudes onto the Liszt of five years earlier.Footnote 49 Yet the fact remains: Liszt abandoned Sardanapalo despite the comparatively advanced stage which work on this operatic project had reached, and I believe that the intractable nature of the material was at least an influence on this decision. His affinity for Byron would ultimately find more adequate (if oblique) creative outlets in instrumental music: in the symphonic poem Tasso: lamento e trionfo (pub. 1856) and the first volume of the Années de Pèlerinage: Suisse (pub. 1858).Footnote 50
But does this discovery of Sardanapalo really change music history? The honest answer would probably have to be ‘no’. Needless to say, an operatic fragment never heard at the time can have had no impact, and thus in a way stands outside history. The inherent quality of the music may be high, but it certainly does not eclipse Liszt’s other major works of the era. True, it does shed new light on our picture of Liszt, and it is possible that future narratives of the composer’s story will pay more attention to his lifelong fascination with opera, now that we have a substantial portion of a mature work to engage with. But the Liszt of the 1850s will continue to be defined by the compositions that were actually completed and disseminated at the time: his orchestral and piano music, in particular, which had an enormous impact on his contemporaries and successors. For a complex of reasons, Liszt’s vocal music has never made the same impression: despite occasional pleadings, his oratorios continue to languish in relative obscurity, despite being of the highest quality.Footnote 51 Nonetheless, Sardanapalo is a fascinating new discovery well worth studying as a crucible of his development, but also for its own sake as a signpost to a road ultimately not taken.
