FRANZ LISZT : Rarities, Curiosities, Album Leaves and Fragments (Leslie Howard) (1998) 4CD / APE (image+.cue), lossless

The fifty-seventh and last volume of this series will present the complete Hungarian Rhapsodies in their final form; this penultimate volume contains some of the earlier versions and alternative texts for some of the Rhapsodies and, as the title suggests, a drawing-together of several other strands of Liszt's work.

By way of an overture to this quadripartite recital, the series of solo versions of symphonic poems continues with Orpheus. As with the earlier publication of Les préludes the impetus was provided by one of Liszt's students, who first made the transcription: Friedrich Spiro. Liszt thoroughly revised the transcription before it was published and, as so often, made some modifications which are not present in any of the other published forms of the work (for orchestra, two pianos, piano duet and piano/violin/cello).

The next group of works comprises five transcriptions in less common guise. The Orgel-Fantasie und Fuge in G minor, transcribed from Bach's mighty original for the organ (BWV542) is offered here in its primary text, without the embellishments which later adorn it (the other version is in Volume 13 of the series). The differences affect only the Fantasy, but the work would be unthinkable without its magnificent Fugue.

Liszt's important arrangements of six of Chopin's Polish Songs appear in Volume 5 of this series; the later French edition of the fifth song, Mes joies (My joys – Moja pieszczotka, op 74/12) contains several varied passages, and a very beautiful additional coda. The simple version of Schumann's Widmung (Liebeslied op 25/1) which follows is in a completely different character from the very popular published version (see Volume 15). The manuscript of the simpler version, which adheres more closely to Schumann's text, is located in the Library of Congress, and has been published in recent times in America.

The concert version of the Marche pour le Sultan Abdul Medjid-Khan (by Gaetano Donizetti's brother Giuseppe) is to be found in Volume 40. The so-called simplified version is quite another view of the same material, more delicate and less ominously rumbling, but building to a stirring climax just the same. The first (completed) version of the transcription of Rossini's La charité (the third of his Trois Choeurs religieux) was intended for publication but was then submitted to much revision before the published version appeared (see Volume 24). The Goethe-Schiller Archive (to whom we are eternally indebted for so much kindness, and for permission to record so much unpublished material in their possession) houses Liszt's original manuscript, with instructions to Conradi to prepare a fair copy, leaving some passages blank for Liszt to rewrite. They also have Conradi's manuscript, with these passages completed, and so this version can be performed. The Conradi manuscript also contains later revisions which pertain to the printed version and are not relevant to this original text.

The first version of Il penseroso – later revised for the second book of the Années de pèlerinage – has not been published, and the manuscript does not actually bear a title, but it does contain the same epigraph – a quatrain by Michelangelo which appears in the published collection (see Volume 43). There are many minor differences between the two versions, but the intense expression of this simply-laid-out work remains constant. The title of the early version of the Canzonetta del Salvator Rosa is extended to read 'per il clavicembalo da F. Liszt', but a performance of this piece on the harpsichord would be a travesty, so essentially pianistic is its conception, despite the Baroque origins of its model. This version is very straightforward, with the opening literally repeated after the interlude (the later version appears in the second book of the Années, in Volume 43).

En mémoire de Maximilian I is the first version of the Marche funèbre from the third book of the Années (see Volume 12). All the catalogues date this piece and its revision in 1867, presumably because the manuscript bears the date '19 juin 1867', but that is merely the date when the Emperor of Mexico was executed. The date of composition cannot be precisely determined, and the revision may well have been made only at the time the published collection was assembled. The most obvious among numerous differences from the final version is the much shorter and less triumphal ending.

Among Liszt's many album-leaves and musical visiting cards are a group of pieces based on a theme familiar to us as the principal theme of the Première Ballade. The Album-Leaf in A flat (Portugal) contains some small harmonic touches not found in other versions. The Album-Leaf (Première Consolation) proceeds exactly as the opening of the first of the Consolations, but concludes with new material.

The first draft of a piano version of the Schnitterchor (Reapers' Chorus) from the Prometheus Choruses is part of a very messy manuscript which continues with the opening of a piano transcription of the following chorus – the Winzerchor. The final version is actually based on a later version of the choruses themselves (see Volume 18) and avoids some of the unnecessary complications of this first effort. The first version of the Künstlerfestzug follows the same design as the printed version (Volume 28) with the usual fascinating array of alternative views of the same fundamental musical material.

The second disc of this collection also opens with a symphonic poem in piano transcription. Festklänge was published as a transcription by Liszt's disciple Ludwig Stark, which had no doubt been made under Liszt's supervision, especially since it deviates from the standard orchestral score by employing all of the Polish-character variants which Liszt composed (presumably to please the Polish Princess Carolyne zu Sayn-Wittgenstein) and which only appear as a supplement to the score. (Liszt's own versions for two pianos and piano duet do not utilize these variants.) This performance was prepared from Liszt's corrected version of Stark's transcription and from a further manuscript containing Liszt's own rewritten transcription of most of the peroration.

Three pieces connected with Rossini follow: La Caritá is a further, simplified version of La charité, whose unpublished manuscript is marked (but not in Liszt's hand) 'harmonium' – an absurd proposal for a piece so utterly suited to the piano, and quite unplayable on the harmonium without a great deal of rewriting. The manuscript bears a quotation from the first epistle general of St. John, 4/16: 'Deus charitas est, et qui manet in charitas in Deo manet, et Deus in es' ('God is love; and he that dwelleth in love dwelleth in God, and God in him'). La serenate e l'orgia – Grande Fantaisie sur des motifs des Soirées musicales de Rossini is the first version of what was later the Première Grande fantaisie (see Volume 21), and although the shape of the work is not substantially different, several passages were rethought for the second edition; the coda, for example, is a much shorter headlong rush for the final bar. The Introduction des Variations sur une marche du Siège de Corinthe is a problem piece: the manuscript is signed and dedicated, but there is no trace of the actual March or the Variations, and we do not know if they were only improvised, or that they were ever composed, or even if the young Liszt's introduction was intended to precede someone else's variations. (There is also a missing fantasy on Rossini's Maometto, the earlier version of the same opera, but this may have nothing to do with the present work. We are indebted to the painstaking labours of Nancy Reich for our knowledge of this work.) Since the introduction ends on the dominant seventh of F major, it was thought best that the next piece in this recital should be in that key, even if there be no other reason to connect the two works. The Morceau en fa majeur is always listed as unfinished, and in the sense that there is a great deal of unworked-out revision in the (untitled) manuscript, this is true. But ignoring all of the additions one can retrieve the basic original version of this very lovely piece, which was composed at Nonnenwerth in the summer of 1843. There is something faintly operatic about the melody but, unless and until someone can identify a prior source for the melody, we must presume that the theme is Liszt's. A further perplexing manuscript is that of the Fantasie über englische Themen. The present writer has put the four large extant fragments of this work together with the minimum of addition, just in order to be able to present the material in performance. What we have to begin with, then, is the Allegro from the overture to Handel's Messiah. The opening bars of this section are missing from the manuscript, but were easily reinstated. We do not know if Liszt had transcribed the first, slow section of the overture, but since he does not include its reprise at the end of the Allegro, it was thought best to omit it. The Allegro is then extended and leads to variations on See, the conquering hero comes from Handel's Judas Maccabaeus. The last bars of this section were also missing. The rest of the manuscript is fragmentary and is based on Rule, Britannia (from Arne's Alfred) and God Save the Queen. Liszt may have cannibalized the manuscript to produce his fantasies on each of these two melodies. The Rule, Britannia piece has never been found, but the God Save the Queen fantasy – which briefly refers to Arne's melody towards the end (see Volume 27) – provided some otherwise missing material, as well as a suitable coda.

The Anfang einer Jugensonate was the old Liszt's recollection of a work of his youth, which he wrote down from memory for his biographer Lina Ramann.

The programme continues with eighteen Album-Leaves. In the course of this series we hope to have discovered and recorded the majority of the surviving album-leaves, and would also note that the few that are not recorded here are identical to some of those that are – Liszt was particularly fond of one or two of these fragments and would write and write them again as little dedications for many of his friends and admirers over his long career. The circumstances of their composition is usually not known, and many of them are in anonymous private hands and cannot be seen, but they are known from their publication in facsimile in various auctioneers' catalogues. It is for this reason that we lack the last fifteen bars from the verso of the manuscript of the early Andantino in E flat, only the recto of which was published. The present whereabouts of the manuscript are unknown, and this musical material is known only from this leaf.

Ah, vous dirai-je, maman is apparently Liszt's only use of this popular tune. The Pressburg leaf is a melody from the Grande Valse de bravoure (see below and Volumes. 1, 26), whilst the Vienna leaf is a melody from the Valse mélancolique (see Volumes. 1, 55) – and thanks are due to Dr. Mária Eckhardt for making it available. The Leipzig leaf is unknown from other sources, but the beginning of the Exeter Preludio is instantly recognizable as the introduction to the Petite valse favorite (Volumes. 28, 55). The Detmold piece is also an unknown theme; the Magyar theme we know from the tenth of the Magyar Dalok (Volume 29), and the leaf on the Rákóczi-Marsch is but a further example of a real obsession (see Volumes. 27, 28, 29, 35, 51, 55, 57 and below). The E major piece of 1843 is unknown from elsewhere, but the A flat piece is again the main theme of the first Ballade (Volume 2). The Lyon Prélude was a favourite of Liszt's, containing, as it does, a little flourish which includes all twelve notes of the chromatic scale in tonal ambiguity – he wrote this fragment on several other occasions; the long-sought Prélude omnitonique which follows turned out not to be a missing work so much as another flourish through all the notes of the chromatic scale, in a chord sequence that allows the bass to proceed in a whole-tone scale – the passage is familiar from the tenth of the Études d'exécution transcendante (Volume 4) – and on this occasion there is no resolution. The Leipzig leaf of 1840 is a further theme from the Valse mélancolique; the Berlin Preludio adds a splendid cadence into C major to the Prélude omnitonique, whilst the Braunschweig Preludio takes the same material into F sharp major. The little Serenade is kin to Die Zelle in Nonnenwerth (see Volumes 25, 26, 36, 51), and the Andante religioso, of which there are at least two examples extant, is an extended quotation from the early Harmonies poétiques et religieuses (Volume 7).

By way of a coda we include the doubtful Mazurka in F minor. The copyist's manuscript does not credit Liszt as composer or (more plausibly) transcriber, but the manuscript is held in an otherwise completely verifiable Liszt collection held by the Gesellschaft der Musikfreunde in Vienna.

The second version of Gaudeamus igitur – Paraphrase was prepared by Liszt for publication and survives in a fair copy in a copyist's hand with corrections and title-page in Liszt's hand. Unlike the first version (Volume 40), the revision was never published (and it is not to be confused with the later Gaudeamus igitur – Humoreske, also in Volume 40). It remains an amusing, if entirely unimportant, pièce d'occasion.

The group of ‘Rapsodies hongroises’ which follows shows some remarkable alternatives to the standard texts: the version of Rapsodie hongroise II contains all of the later additions to the work (made on three separate occasions) except the larger of the two cadenzas which will be found in the version contained in Volume 57. These additions give a completely different atmosphere to the work; the added gypsy elaborations range from the weighty extension of the opening motif to the imitation of the cimbalom at the end of the first part, and the last three chords are replaced with a coda much more reminiscent of the style of the old Liszt. The alternative passages in the versions of Rapsodie hongroise X and Rapsodie hongroise XV recorded here are to do with the replacement of difficult glissandi with other material. But the effect is not simply that of facilitation; the ossia texts also offer real musical changes. Rapsodie hongroise XVI is given in its original version: Liszt had it published, with the title and his name in Hungarian (as he preferred it in his last years); he very soon reissued it with several additions and changes (see Volume 57). Liszt made two attempts at a conclusion to the delightful and all-too-brief Rapsodie hongroise XVIII which remain in manuscript. They are both recorded here, separated by the Mephisto-Polka. The Polka has two texts: one, as performed on this occasion, simple and direct in its language; the alternative (Volume 28) so replete with decorative ossia passages as to sound almost another work altogether.

The remainder of this disc is devoted to fragments of various kinds: the effect as a recital will no doubt be strange, but the music, however partially worked out, remains intrinsically compelling. The Cadenzas to Un sospiro were written for Luisa Cognetti and Eduard Dannreuther in the mid-1880s and may be compared with three others (Volume 38) all designed to be inserted at the same point of the study: before the recapitulation. Three Album-Leaves follow: the one the catalogues list as Poco adagio aus der Graner Messe is an excerpt from the Agnus Dei of the composer's Missa Solemnis, S9; the next presents a strangely misquoted version of the principal theme from the symphonic poem Orpheus; and the third, although marked in the manuscript as an excerpt from the symphonic poem Die Ideale, is actually laid out in a way which more closely resembles the trio section of the Künstlerfestzug, whence the theme in question derives (see above).

The fifteen tracks which follow do not account for every last known Liszt sketch for the piano. There are hundreds of such things relating to works which he later completed, and to record all of them would be otiose in the extreme. Instead we present fragments known from published catalogues, or pieces which remained incomplete elsewhere. The unfinished oratorio St Stanislaus consists mostly of more-or-less completed numbers (see Volume 14, for example), but there is a tantalizing fragment in the Library of Congress which is worth preserving in sound. There may once have been more of the stirring Winzerchor (Winnowers' Chorus from the Prometheus choruses) since the manuscript ends at the end of a very full page which continues from the Schnitterchor transcription.

It is very sad indeed that La Mandragore (from Delibes's Jean de Nivelle) remains incomplete. All that exists is a free fantasy in Liszt's last and sparsest style on the orchestral introduction to the opera, then on the introduction to the ballad itself, and then – silence.

Tireless Lisztian William Wright was kind enough to supply a fragment (of an operatic fantasy?) held in Glasgow. So far, it has proved impossible to find the rest of the piece, or to identify the work upon which it is based. The succeeding Operatic aria – and sketched variation consists of a transcribed melody full of elaborate musical curlicues, and dislocated bits of an incomplete variation. It appears to date from Liszt's youth. The same manuscript contains just the theme of Meyerbeer's Valse infernale (from Robert le diable) in rather a different form from that found in the famous Réminiscences (Volume 30).

Also to William Wright goes the credit for unearthing the Allegro maestoso which may well be a sketch for the opening of an Étude in F sharp major. Such a study would have been the next piece in the projected continuation of the Étude en douze exercices (originally announced as 48 rather than 12 pieces, but only the first set ever appeared). Once again we have Mária Eckhardt to thank for the publication of the Rákóczi-Marsch in the incomplete simplified text which follows in the same manuscript upon the text of the so-called first version (see Volume 27).

The Harmonie nach Rossini's Carità is marked to follow ‘Palestrina’, and so belongs to the first planned cycle of Harmonies poétiques et religieuses (see Volumes 7, 47) and would have succeeded the Miserere d'après Palestrina. It is interesting that La charité was considered for this volume; so too was the Stabat Mater, S172b (Volume 7) – a Liszt original but with an introduction taken from Rossini's Cujus animam (Stabat Mater). Of course Liszt later transcribed this last piece and issued it with the final version of La charité (Volume 24), and so there was no further connection between Rossini and the Harmonies.

The single page of Marie-Poème is all that survives of a projected set: Marie-Poème en 6 chants, inspired by the poetry of Auguste Brizeux, planned in 1835. Nothing is known of the origin of the Andante sensibilissimo, which must have been some kind of private musical joke since it is marked 'avec miaulement' ('with miaowing'); the manuscript is in the Library of Congress (scribbled at the foot of a page of sketches). The oft-cited Melodie in Dorische Tonart consists simply of the two themes from Totentanz (Volumes 9, 53a, 53b) which come from a version of the burial service; the chants also appear in open score on a leaf at the Goethe-Schiller Archive marked 'Prose des morts' (no musical connection with the piece to which we have given that title in the first cycle of Harmonies poétiques).

The Dante fragment appears to be the earliest sketch of the opening of the Paralipomènes à la Divina Commedia, later the Après une lecture du Dante (see Volumes 51 and 43 respectively). Polnisch may well be a sketch for an earlier projected piece which was replaced by the work of the same title in the suite Weihnachtsbaum (Volumes 8, 51); and the Korrekturblatt turned up attached to a Liszt letter and is now at the Istituto Liszt in Bologna. It certainly applies to the earliest manuscript of La lugubre gondola (a manuscript which has only recently surfaced and which is not yet available for recording) which dates from December 1882, and is intended as an alternative passage for nine bars towards the end of the work.

Liszt arranged for all twelve of his Weimar-period symphonic poems to be published in versions for solo piano, each one under the name of one of his young protégés as putative arranger. As we have already seen, Liszt frequently made many anonymous contributions and improvements to such arrangements. This series of recordings has included all of them that incontrovertibly bear his rearranging hand. In the case of Mazeppa we probably have a case of a rejected effort by Theodor Forchhammer. Forchhammer's versions of Tasso, Heroïde funèbre and Hamlet all appeared in the Breitkopf edition of the twelve pieces, but the published version of Mazeppa is that of Ludwig Stark, who also arranged Ce qu'on entend sur la montagne, Prometheus, Festklänge (see above), and Hunnenschlacht. (The other pieces were arranged by Karl Klausner – Les préludes (Volume 38), Arthur Hahn – Die Ideale – and Friedrich Spiro – Orpheus (see above) and Hungaria (see Volume 55). Of course, Liszt was himself the sole arranger of his later symphonic poem Von der Wiege bis zum Grabe (Volume 25). The manuscript of the unpublished Forchhammer version is to be found in the Liszt Research Centre in Budapest, and again thanks are due to Mária Eckhardt for arranging for the present writer to study it in situ and to bring away a copy of it. The manuscript is full of Liszt's corrections, although from time to time these are written in shorthand – in parallel passages, for example – and at one point at the very end he has deleted Forchhammer's text for four bars without indicating its replacement, but that could easily be supplied with reference to the orchestral score.

The manuscript of the Valse surfaced at a sale of autographs in Paris in early 1995 where the present writer made its acquaintance. It has since been published in America in an excellent edition by Rena Charnin Mueller. (Curiously the handwriting of the title in the manuscript shows 'Walse' – a spelling which Liszt did not otherwise employ.)

The Ländler, Dumka, and Air cosaque were all published in facsimile in 1960, and the manuscripts originally came from the library of the Countess Lyudmila Zamoyska (known to Lisztians as the original composer of Puszta-Wehmuth – see Volume 12). Whether or how much the countess contributed to these little pieces cannot be easily determined, but their charm is incontrovertible. (The Liszt Society published the pieces in its 1997 Journal, computer-set.)

The Marche funèbre was the young Liszt's immediate response to the death of his father. It was composed at Boulogne on 30 August 1827, two days after Adam Liszt's death, and shows a tremendous possession of heroic spirit (as George Nugent so rightly described it in his article published with the facsimile of the manuscript in the 1993 Liszt-Saeculum), much influenced by Beethoven.

Magyar tempo is familiar from various sources: the Ungarischer Nationalmelodien No 1, the Magyar Dalok No 5 and the Rapsodie hongroise VI (see Volumes 27, 29, 51, 57) although this manuscript from the Bodleian Library differs in many minor details, especially of rhythm.

Pásztor Lakodalmas ('The Pastor's Wedding') is a work of Count Leo Festetics, very much in the style of the pieces in Ungarischer Romanzero, with elaborations by Liszt. Apparently Liszt made some further changes to the piece post-publication, but the whereabouts of these is unknown, as is the extent of Liszt's involvement in the present version.

The next group of four pieces entails Liszt's charming idea of supplying an introduction and coda to a finished composition by one of his friends. (The resulting form is an echo of such pieces as Weber's Invitation to the Dance.) It is curious that, although Liszt's contribution is small, it is nonetheless able to determine in some decisive way the listener's response to the work thus framed. The Einleitung und Coda zu Raffs Walzer in Des-dur has been published long since in the Liszt-Pedagögium. Unfortunately that publication identifies the Raff waltz as being in A flat, and so the work could not be assembled until the correct piece was identified: Tanz-Kaprize Opus 54/1. Similarly, the Einleitung und Coda zu Rubinsteins Étude in C-dur – an unpublished manuscript in the Library of Congress – was for a long time listed as an introduction to Rubinstein's Staccato-Étude, Opus 23/1. It has also been identified correctly as belonging to the once-famous Étude on wrong notes (a trick with mild dissonances, really) but then postulated to be the introduction to an unfinished arrangement. As with the other pieces of this kind, the paper tells the tale: the introductory bars are on one sheet, the concluding bars (or in this case, just one bar of flourish) on another. The Einleitung und Coda zu Smetanas Polka also appeared in the Liszt-Pedagögium, but had the great fortune to be republished with the Smetana piece in an encore album edited by the much-missed virtuoso Raymond Lewenthal. The Einleitung und Schlußtakte zu Tausigs 3. Valse-Caprice suffered an even stranger fate before it was finally published in its correct context by the Liszt Society in its 1989 Journal: it appeared in an American publication as a single piano piece without the Strauss/Tausig Waltz to separate its two leaves – and thence was mis-catalogued in Grove VI, under a title derived from its tempo indication ('Ruhig'), and was given a misleading catalogue number. Liszt's pages clearly contain the thematic material of Johann Strauss's Wahlstimmen ('Voting Papers'), Opus 250. The great scholar and editor Robert Threlfall was the first person correctly to identify this manuscript held in the Library of Congress.

The original conception of the Third Mephisto-Waltz has never been published, but the comparison between this version and the work's final form (in Volume 1) is an engrossing study: the first version is almost too concise, whilst the later version is right at the edge of prolixity. Both contain a reflective musing passage – risking the loss of the waltz character in an excursion into 4/4 time – and both deliberately exploit an uneasy fluctuation between D sharp minor (the unusual key of the piece) and F sharp major (its more comfortable relation).

It would seem that the unpublished Petite Valse was written some time between the composition of the third and the fourth Valses oubliées, since it is described as a pendant to the three waltzes. The manuscript (in the Goethe-Schiller Archive) is in a terrible state: at some stage it has been torn asunder, both from top to bottom and from side to side – a crime if the deed was not perpetrated by Liszt in a moment of self-doubt, merely a pity otherwise, because the fragment is of that other-worldly nostalgic beauty unique to the gentler works of Liszt's old age. The piece consists of 101 bars of music, and a final change of key signature indicating a return to earlier material. The present writer has completed the piece by adding twenty-five bars, twenty of which are entirely Liszt's, and the last five bars only contain the vanishing spectre of the previous phrase, in imitation of a passage in the Troisième Valse oubliée. This is simply too haunting a piece to discard unheard.

To end with a revivification of the spirits, and at the same time to hark back to the very first disc in this series, we offer the original text of the flamboyant and witty Grande Valse di bravura (subtitled Le bal de Berne in one of the early editions). This is of the musical and pianistic stamp which first made people pay attention to Liszt, and which, as so often, contains both the public and private humours of his musical expression, where immanent and daring originality is garbed in a glittering carapace. Leslie Howard © 1999
All Tracks & Credits