Robin Holloway, composer and critic, teaches in the Cambridge Music Faculty. His song-cycle, The Lovers' Well, is a setting for baritone and piano of poems from Geoffrey Hill's 'The Pentecost Castle'. It was first performed at Emmanuel College, Cambridge in April 1984 and broadcast on Radio 3 in January 1985. The cycle will be heard again at the Cambridge Poetry Festival.

So far as I know the only previous settings of Geoffrey Hill are of a group of poems from his first book in a cycle of the same name, For the Unfallen, by Gordon Crosse, and the cantata Ad Incensum Lucernae (with words from Tenebrae) by James Brown. I had wanted to try since first reading Mercian Hymns: the idea - of polyphonic strata of criss-crossed time and place presided over by a timeless placeless figure coarse benevolence - remains highly suggestive; the actual words of these prose-poems, though eminently projectable, suggest rather the possibility of a musical equivalent than a setting to be sung.

Almost all the rest of this small distinguished output seems to me music-resistant. These poems are at the opposite pole from a Campion or Herrick (or for that matter the lyric work of a Tennyson or Wallace Stevens) who positively invite the composer into 'numbers'. (The comparison here is Donne, most of whose 'songs' are far too subtle for setting.) This is ironic since Hill's inspiration so often comes from music - witness the 'chorale preludes' upon Celan or the seven sonnets after the 'seven passionate pavanes' of Dowland's Lachrimae. It is not just that he is complex in thought and can be dense to the point of tortuosity in expression, for another kind of word-setting can flourish upon this, achieving music against the grain of the words to the surprising enhancement of both. It is that his work is so finished, wrought and worked-over to a point of self-fulfilment which gives music no foothold or entry - in fact nothing to do. A setting of such poems as the architecture sequence from Tenebrae or a section from the Péguy-cycle that responded on the same level to their fulness of imagery and meaning is scarcely conceivable; moreover a musical response that did not spoil by dilution would spoil by overkill, as if the Wolf of the Spanish songbook had lavished such baroque ardour upon Gongora rather than upon elegantly-turned conceits and common-places. Hill is given to putting forward as an ideal for verse Milton's famous 'simple, sensuous and passionate'. So far as his own is concerned, this is a hopeful trinity of adjectives even for the willing reader; and for the composer, in its operative first word that unlocks the others, completely untrue!

The exception is 'The Pentecost Castle' sequence, where sensuousness and passion glow unimpeded through tiny unpunctuated stanzas of lucid structure and direct utterance. Here in a word is simplicity, though the overall effect is anything but simple. Complexity has been melted and the labour of a 'difficult' author done by him rather than for him. These at-first-sight transparent lyrics have been distilled over and over to achieve an intensity filled to the brim with content that does not sacrifice pellucidity. Here, unique in Geoffrey Hill's oeuvre, a musician may take and drink.

I was at once attracted to this sequence when Tenebrae came out, feeling strongly its possibility for a song-cycle. When later I heard the poet read it aloud, the possibility was renewed because his simple, sensuous and passionate performance realized a powerful drama only latent in the lyrics as seen on the page. But now comes a difficulty quite different from that of setting a complex text; one that concerns temperamental incompatibilities that may well be unbridgeable in spite of strong attraction. Most of 'The Pentecost Castle' indeed 'walks into music', but a certain atavistic rationality cannot swallow the 'churchiness' of some of the imagery. My position lies somewhere between the shrill trumpet-tones of post-Nietzschean anti-Christianity and the plain middlebrow honesty of Philip Larkin when he says 'I am not going to fall on my face every time someone uses words such as Orpheus or Faust or Judas'. My reaction to the 'churchy' element in 'The Pentecost Castle' is both artistic irritation at what Larkin calls 'dead spots', where Christian imagery is used to invoke an effect that has not been earned, and the anger of the affronted Nietzschean. This is not just to jib and jibe at Jesse-trees and Jesus; I find these references accompany a distinct loss of the pellucid essentialized quality that shines over the sequence as a whole. They seem to me second-hand and 'literary', and can moreover produce individual stanzas and isolated lines of startling badness.

This is not really a question of faith and disbelief. If these images spoke there could be no objection, just a straight-forward difference of outlook. But I feel here that something of great beauty and integrity is spoilt by the intrusion of elements that are mawkish, self-indulgent and moribund. 'Christ the deceiver' with his wounds 'so cunning and so true' evokes only the shudder of the 'dead spot', where the proper names that remain unglossed in the first poem - 'Medina's pride, Olmedo's flower' - catch with inexplicable power (and total simplicity) a note of sensuousness and passion that reverberates on throughout the rest of the sequence. But the religious imagery is too thick with unharmonized association to do any such thing. That its use may be in part ironic does not help the reader for whom irony as a device belongs decidedly among the discarded stage-props of the modern movement (how does one square irony with 'simple, sensuous, passionate'?): the words cannot be any other than what is given on the page; if the attitude that underlies them has not got into them, the irony is as unspeaking as the imagery.

So if this bigot was to set 'The Pentecost Castle' the 'dead spots' had to be omitted as a matter of conscience, mainly artistic. Which is at bottom practical - one cannot fake a response in areas like this - it would come out as a dead spot in the music, exposed as such by proximity to passages written in full accord with the poetry. And though I have only put the negative side, the positive incentive to set what I loved in Hill's sequence was still strong enough to outweigh it.

Once I felt free to cut as taste directed (regretting only the ravishing no. 7) the music followed fairly quickly. So nine only of fifteen tiny poems are set without break, and the cycle ends with a strong sense of closing the circle as the final music brings to its full development the half-formed shapes and shadows from which it had first emerged. In place of the main block of omitted poems (7-10 inclusive) comes an interlude for piano alone (added, it must be confessed, well after the composition of the vocal settings) which can be offered as compensation as well as apology. This interlude sets out from the cycle's point of repose ('there love/rests and is saved') into a chalorous paragraph that looks first to the work's outermost ends, then forward to the 'proud citadel' of love fulfilled (no. 13), then back in a lyrical development of the vernal love-death tenderness of no. 2, before turning finally down down down like the wounded heron to the point where fidelity to the poet is resumed, and the voice re-enters with its urgent questions and reproaches ('why do you hold back/dearest heart').

The song-cycle's new title The Lover's Well (taken from the sequence's final poem) epitomizes the different emphasis. The fusion in 'The Pentecost Castle' of mystical and erotic has yielded to a simpler concentration on the erotic alone. But its musical presentation certainly remains dark and spiritualized, and the many half-unconscious echoes of Wagner's Parsifal (the ultimate instance of this particular fusion) are in resonance with what has been consciously dropped from the texts (ironic, perhaps!). I can only justify this cavalier way with a poet's intentions as well as his words, by a final result where words and music have fused. The omissions possibly even intensify that feeling of latent drama so strongly communicated in the poet's live performance which provided the musician's directest inspiration. It is as if the manuscript of a story already elliptical had come down to the twentieth century incomplete from a remote past. During the breaks in continuity made first by the poet then by the composer, the latent story moves inevitably along its subterranean course; when the stream re-surfaces nothing is further explained than it had been before but the goal comes ever nearer - a final condition of clear-eyed acquiescent truthfulness unsullied by doctrinal difference.