Robin Holloway

Corrective to a fault

Glenn Gould called it ‘the greatest song cycle ever written’, entitling his notes on the two versions of Paul Hindemith’s masterpiece ‘A Tale of Two Marienlebens’.

Already a subscriber? Log in

This article is for subscribers only

Subscribe today to get 3 months' delivery of the magazine, as well as online and app access, for only £3.

  • Weekly delivery of the magazine
  • Unlimited access to our website and app
  • Enjoy Spectator newsletters and podcasts
  • Explore our online archive, going back to 1828

Sometimes the rewrite can be more thoroughgoing — Beethoven’s ruthless shake-up of unsuccessful Leonora into imperfectly satisfying Fidelio, Brahms’s drastic recomposition of his early B-major piano trio, in all its looseknit romantic extravagance, into the taut manner of his late maturity. While Bruckner is oddest of all: terminal diffidence alongside radical boldness caused him to tinker endlessly with his symphonies old and new, by no means to their obvious advantage, and often to the clear reverse.

Unique to the Marienleben revision is a polemic from its composer invoking not just improved workmanship so much as a shift of stance that is also an elevation to higher moral ground. It’s piquant that the revised score with its breast-beating Preface came out in 1948, year of the notorious committee censuring Shostakovich and other notable Soviet composers for crimes of ‘formalism’. Hindemith is his own judge as well as wrongdoer. We music-lovers are the jury.

In 1948, the mature composer took his youthful songs, completed in 1923 when he was 28, and reworked them to comply with compositional principles that, commencing intuitive, had slowly developed into a system every bit as rational and constrictive as the contemporaneous 12-note method of Schönberg to which Hindemith was polemically opposed. Another early masterpiece, the opera Cardillac, was also subjected to treatment more ideological than artistic. The two works were to stand as show examples of juvenile excess and error curbed by mature experience into responsible discipline.

The 15 songs of Marienleben, setting Rilke’s beautiful poems on the life and death of the Virgin, reveal unwonted delicacy and suppleness within Hindemith’s more familiar manner, all no-nonsense energy and abrasion, not a little coarse though always exhilarating. There is abrasion, however: in the piano part, harmony and especially counterpoint can be harsh as well as tender, and the demands upon the soprano’s stamina are fearsome. Yet my recent exposure has converted initial admiring discomfort into something approaching love.

Circumstances made it curiously oblique. Possessing only a score of 1948 and an LP of 1923, to hear the one with eyes on the other is to be compelled straight into the composer’s self-imposed alterations. Then, purchasing a CD of 1948, I could see without discrepancy, and began to accept them as faits accomplis as their perpetrator wished (he even attempted to embargo performances of the early version). Then with a newly purchased score of 1923 used to follow the 1948 CD, one sees/hears the damage with newly fuelled regret. And, finally, the original version to sight and sound, confirming that youthful inspiration is found here in all its intensity and daring, before its replacement by middle-aged spread, bland for all its liveliness, sententious for all its sagacity.

Sometimes Hindemith leaves well alone. Two songs are virtually unaltered, and there’s much good material in others. Several are completely reconceived, notably the centre point Wedding at Cana, where a succinct original becomes a sprawling scherzo/fugue in which the poem is engulfed rather than set. That 1923’s ‘fecklessness’ over key centres is replaced by 1948’s carefully elaborated symbolic scheme fails to register with the hoped-for weight (but one does notice the introduction of quasi-Wagnerian recurrences, for instance a bitter dissonance evoking the Crucifixion). Most disconcerting of all are the two widely separated songs built upon a ground bass (2 and 13): the bass remains unchanged, above it the poly-phonic voices in soprano and piano right hand diverge totally, while making the same general sense. One wonders how the composer could understand the changes as in any way corrective to a fault. And two other songs seem to me so remarkedly superior in their first versions as to make one wonder about his artistic sanity. The original Birth of Christ (7) is an exquisite study in parallel and contrary-motion thirds, as if the devotional songs in Wolf’s Spanish Songbook were transported into the 20th century; its replacement, a humdrum pastorale that the later Hindemith could and did extrude by the yard. Earlier, the same machine-made momentum in the revised Annunciation to the Virgin (4) replaces a lovely original wherein motion virtually ceases for the cycle’s sweetest moment, a rapt recitative concerning the Unicorn, also a creature born without carnal generation.

Though 1948 yields a few compensations and very occasional improvements, the ‘Two Marienlebens’ all in all tell a tale of genius defeated by pedantry. But the case remains absorbing even if the jury has long since decided and disbanded. Das Marienleben might not be ‘the greatest song cycle ever written’ — what about Schubert, Berlioz, Schumann, Mahler, Schoeck, Britten? — but its place is perennial: performers and listeners able to withstand its formidable demands will be formidably rewarded.

Comments

Join the debate for just $5 for 3 months

Be part of the conversation with other Spectator readers by getting your first three months for $5.

Already a subscriber? Log in