CLASSICAL MUSIC / Where words fail: Bayan Northcott reflects on the ambiguities behind a new anthology of (nearly) all the non-operatic texts set by Benjamin Britten: Benjamin Britten's Poets - edited by Boris Ford: Carcanet, pounds 25

The surprise is that nobody had thought of it before. Seventeen years after Britten's death, the literature around his life and work has already swelled to vast proportions. The libretti of his operas, for instance, were long since gathered into a single volume with a penetrating preface from Hans Keller. Yet only now has the huge array of non-operatic texts Britten set been collected as an anthology, entitled Benjamin Britten's Poets, edited by Boris Ford and scheduled for publication by the Poetry Press, Carcanet, on 9 June.

Of course, those who possess a selection of Britten recordings, or some of the more standard verse anthologies from which he himself drew, will have many of the texts already. The point of assembling them in one volume is rather to see what emerges about his patterns of choice, about his overall development and sensibility. In this respect, two of Professor Ford's decisions are helpful. The texts are ordered not under writer or historical period but in the sequence that Britten set them. And where short extracts were taken from longer poems, these have been reprinted in extenso - though Ford's indications of exactly what was, and was not, set are occasionally awry.

His brief annotations could also sometimes have been fuller. He fails to mention, for instance, that the touching little Shepherd's Carol of 1944 was one of only two brief sections that Britten set from the text of an entire Christmas oratorio, entitled For the Time Being, which W H Auden had written for him - a default that marked the beginning of the end of their creative relationship. In fact, the other section, entitled Chorale (1944), is missing altogether from Ford's anthology, together with two Thomas Hardy songs originally intended for the cycle Winter Words (1953) and a couple of unpublished Edith Sitwell settings from 1956 - all of which have more recently been disinterred and recorded.

Yet, quibbles aside, the array of over 360 separate texts that Britten composed between the little Robert Burns ditty 'O that I had ne'er been married' from his ninth year, to his final set of folksong arrangements of 1976, is striking enough. Since Ford has included all Britten's folksong volumes, it is not surprising that the largest single category, some 80 texts, turns out to be anonymous. More surprising are the no less than 74 sacred settings. Maybe Britten was a natural Anglican; certainly, cantatas and anthems suitable for amateur performance and expressive of communal values flowed from his pen, but one hardly thinks of him as a religious composer in the same sense as, say, Stravinsky. Even the War Requiem conveys an essentially humanist compassion; the unworldly numinosity of A Symphony of Psalms was outside his range.

Impressive in a different way are the 50- odd songs and choruses to non-English texts - and not only liturgical Latin. Already at 14, and fresh from his discovery of Debussy and Ravel, Britten was setting Verlaine and Baudelaire in the original with a sensitive precocity; the Seven Sonnets of Michelangelo (1940) achieved a marvellous Mediterraneanising of style, while the Six Holderlin Fragments (1958) and The Poet's Echo (1965), on poems of Pushkin, made daring forays into the soundscapes of the German lied and the Russian romance. Then there are the works built around single writers: Auden himself and the poets Auden pressed upon Britten during that intense seven years of collaboration from their first meeting in 1935 - Donne, Christopher Smart, Rimbaud, Wilfred Owen, T S Eliot - or the authors Britten subsequently discovered for himself, such as the Scots recluse, William Soutar, so hauntingly set in Who Are These Children? (1969), one of his later and most underrated cycles. In a pregnant essay, 'Composer and Poet', written for the Faber Britten Companion (and reprinted by Ford in a regrettably shortened version), the poet Peter Porter confessed that there were writings such as Donne's Holy Sonnets that he understood only imperfectly until he had heard Britten's settings, so uncanny was the latter's ability to complement a text, to realise in music what even the greatest poetry left unsaid.

Or, especially the greatest poetry. For it is Ford's central proposition, and the justification for his anthology, that the poetry Britten chose to set in such vast quantities was more varied in range and consistent in quality than that of any other great vocal composer, Monteverdi, Schubert et al not excepted. Porter seems to probe deeper when he argues that, before they became professionalised specialities, music and verse were perceived as springing from a single poetic source, and that Britten had somehow worked back to such a unified sensibility.

But this would seem to suggest that kind of frictionless lyricism of a Renaissance poet-composer such as Thomas Campion, and while a few Britten works may approach such an ideal - the seraphic Christopher Smart cantata Rejoice in the lamb (1943), for instance - much of Britten's vocal music seems fraught with tensions and ambiguities. And never more so than in that sequence of anthology works centring on the Serenade for tenor, horn and strings (1943), the Spring Symphony (1949) and the Nocturne (1958), where Britten was able to compile from multiple sources the texts best-suited to his formal and expressive purposes and which, aside from the operas, constitute for many listeners his most richly characteristic achievement.

What precisely was the nature of those ambiguities? Since the dissociation of sensibility Porter hints at between words and music in the 17th century, the relation between the two media has been traditionally viewed as a matter of dominance, with solutions ranging from Italian bel canto, in which the function of words may be to supply no more than a mood and a stream of pleasing vowel sounds for frankly musical display, to the kind of lieder composed by Hugo Wolf, in which the music is merely supposed to support and articulate the nuances of a dominant text. Actually, Wolf often seems as concerned with the harmonic subtleties of his accompaniments as with the verbal responsiveness of his vocal lines - as adept as Schubert before him, and Britten after, at filling the perceived gap between words and music with implications.

Indeed, Keller argued that Britten's lifelong bias towards vocal music 'was intimately bound up with his ability to create tensions between musical and verbal rhythm', and that 'these tensions were of a magnitude that made it possible for him to make extended structures with comparative ease'. But he also noted the paradox that, while consistently inspired by words, Britten seemed almost pathologically averse to explaining his music in words, and speculated that Britten may at some level have felt guilty precisely about his ease of vocal composition compared with the difficulties of the 'pure' instrumental music he thought he ought to be writing.

No doubt it would be simplistic to suggest a causal link between any such guilt and Britten's affinity for ambiguous, even nightmarish, subject matter; for the strange resonances that can seem to accumulate not just during, but between, the separate numbers of a work like the Serenade. A composer seeking a lost unity of sensibility by methods that made it impossible to recapture and which he felt he should not be indulging anyway? But, whatever the explanation, the fascination for interpreters and listeners alike of the singularly complex world of feeling behind Britten's deceptively simple lines and textures seems likely to be lasting.

'Benjamin Britten's Poets', edited by Boris Ford, Carcanet, 323pp, pounds 25