Zemlinsky’s Mermaid: a new critical edition

Antony Beaumont


Download PDF
Send by email

One of the most fascinating works from the orchestral repertoire of the 20th century is being reborn. Antony Beaumont is working on a new critical edition and surprises us with an as yet unknown scene.  

In 1976, the orchestral score of an untitled piece by Zemlinsky came to light in a private Viennese collection. Four musicologists involved at that time in Zemlinsky research (Alfred Clayton, Peter Gül­ke, Keith Rooke and Horst Weber) independently identified it as Part I of The Mermaid, and confirmed that a folder in the Zemlinsky Collection at The Library of Congress contained the autograph score of the remaining two movements. Once the two manuscripts had been brought together and collated, the work was performed – for the first time in 75 years – by the Austrian Youth Orchestra conducted by Peter Gülke.

Like all other early champions of The Mermaid, Gülke was obliged to conduct from a facsimile of the autograph score: a challenging task at the best of times, exacerbated in this instance by Zemlinsky’s diminutive handwriting, numerous erasures, corrections and other blemishes. These problems also affected the performing materials. Newly copied for the occasion, they have since been updated several times.

It was announced that Gülke would prepare a new edition of the score, but the project came to nothing. As a stopgap, Universal Edition prepared a handwritten copyist’s score of Parts II and III. With regard to legibility, this score was a distinct improvement, but unfortunately it was marred by numerous errors and omissions. The facsimile of Part I was replaced in 1997 by a computer-engraved score, prepared for the personal use of James Conlon and subsequently acquired by Universal Edition. Unfortunately the engraving was completed in a rush, the score diverged in certain respects from notational norms, and proved to be rather less than reliable. Over the years, most of the errors and omissions in this composite hire-score have been identified and rectified. Nevertheless, over a century after the world première, the need still remains for a fully integrated critical edition, available also as a study score.

Zemlinsky first envisaged The Mermaid as a through-composed symphony in two movements, each subdivided into two sections. During the proc­ess of com­posi­tion (February 1902–March 1903) he recast the work into three separate movements. The first of these follows the well-known story by Hans Christian Andersen from a murky opening (‘at the bottom of the sea’) to the moment where the Mermaid rescues the shipwrecked Prince from ­drowning. The second, in extended ternary form, focuses on a subsidiary scene in Andersen (a ball in the palace of the Mer-king) before moving on to the crucial confrontation between the Mermaid and the Mer-witch. The third depicts the Mermaid transformed into a mortal, her anguish, suicide and transfiguration.

As a protective shield against conservative critics, Zemlinsky chose a subtitle that left his options open. Rather than categorising the work as a ‘symphonic poem’, which would have placed him squarely in the camp of the so-called ‘Neu-Töner’, at the world première on 25 January 1905 he billed it simply as a ‘phantasy for orchestra’. So which came first, the programme or the music? Several lines of Andersen, retold in Zemlinsky’s own words, are included in the sketches for Part I. Already in Part II, however, he reduces the programme to a few key-words. And although Part III still adheres closely in spirit to the original story, it includes no further key-words at all. In the full score, all that remains of Andersen is the title itself, together with an isolated expression mark, ‘wie hilferufend’ (‘as if calling for help’), over a phrase for solo clarinet (Part I, measure 303).

Considering that Zemlinsky’s earlier music had been strongly influenced by Brahms and Wagner, The Mermaid was quite a radical departure. Since the death of Brahms, he had found inspiration in the music of Richard Strauss, indeed no work stands conceptually closer to The Mermaid than Ein Heldenleben. Yet Zemlinsky detected weaknesses in that score and determined that his own music should never make concessions to logic for the sake of effect. In his youth he had mastered the technique of variative development. Although his perspectives had since shifted, he saw no reason to abandon that craft. As a conductor he remained an ardent advocate of Strauss; as a composer, he soon realised that it was Mahler who looked more penetratingly into the future, while the aesthetic of Strauss led ultimately to stagnation.

Zemlinsky’s opera Der Traumgörge, composed between 1904 and 1906, bears witness not only to this new-found influence of Mahler, but also to an intensified interest in contemporary French music, notably that of Ravel and Dukas, and an on-going commitment to the experiments of Schönberg, whom he followed (if at times reluctantly) in his search for ‘air of another planet’. Hence, by the time The Mermaid received its first performance, together with Schönberg’s Pelleas und Melisande, Zemlinsky evidently regarded his score as snow of yester-year.

“The Mermaid is a radical departure from Brahms and Wagner.”

Antony Beaumont

Although it received good reviews from the Viennese press, Zemlinsky evidently considered it a failure. In his handwritten curriculum vitae, submitted to Universal Edition in December 1910, he made no mention of it, although it had meanwhile been performed with some success in Berlin (December 1906) and Prague (September 1908). In 1914, having negotiated a long-term contract with Universal Edition, he suggested several earlier works for publication, but The Mermaid never crossed his mind.

Posterity has shown that Zemlinsky was wrong. Of all his compositions, The Mermaid has established itself as the most popular. Hence the time has come to impose order on the jungle of errata lists accumulated over the years, and to take a fresh look at those passages which Zemlinsky decided to eliminate.

In the autograph score of Part I, the pages containing the Bacchanal are glued together. Often enough, glue crumbles with age and can be removed with relative ease (as was the case, for instance, with the Paris autograph of Verdi’s Don Carlos). In this case, any attempt to prise the pages apart would cause irreparable damage. Fortunately, the first three measures of the Fernorchester are included on the preceding page, where they have been struck out. These reveal that the Bacchanal was scored for an oboe, an E-flat clarinet, two horns, two trumpets and a tambourine, and that the entry was underpinned (at least initially) by a timpani roll in the main orchestra. In theory, this information would suffice to reconstruct the entire passage, following the text of the short score. In practice, Zemlinsky worked this passage over several times, but left nothing definitive or even complete. It is also unclear how he dovetailed the music for Fernorchester into the surrounding entries of the main orchestra. Rather than forcing the issue, the idea of reconstructing this section was therefore abandoned.

In the case of the Mer-witch episode, Zemlinsky simply removed the pages and set them aside. All but the opening measures are preserved at The Library of Congress. The missing passage linked the Trio with a mystic sequence of brass chords, identified in the score as the motif ‘of the immortal soul’ (‘von der unsterblichen Seele’). These were pitched in F-sharp major, then reiterated at the close of the episode in E-flat major. Hence Zemlinsky merely needed to modifiy his transition so as to arrive at E-flat rather than F-sharp major. He then orchestrated the new bars and pasted them over the older ones. Since the original transition is included in the short score, it was possible to restore it with no palpable loss of authenticity, taking the orchestration of the revised version as a model.

In the new edition the two versions of Part II appear side by side, enabling performers to choose freely between them. The original version builds to a wild climax, bordering on hysteria, and disrupts the formal balance of the entire work. The revised version passes elegantly over the agony and ecstasy of Andersen’s tragic story, as if to say, ‘The rest is silence’.

Which version will win the greater acceptance?


Antony Beaumont is a British musicologist and conductor, whose interest focuses on Busoni, Mahler and Zemlinsky. For UE he has edited numerous works of Zemlinsky, including A Florentine Tragedy, The Dwarf and the Lyric Symphony. He also completed the orchestration of Zemlinsky's last opera, Der König Kandaules.