His intention was nothing if not polemical. Most writing about music, he claimed, was tautological in that it tended to describe features which attentive listeners could hear for themselves, whereas what it ought to concern itself with was bringing to the surface the more hidden unities beneath the contrasting surfaces of this or that masterpiece. But since, in his submission, the laws of musical continuity and conceptual reasoning were essentially different, verbal commentary upon music would always remain problematic. Hence his new method of wordless 'Functional Analysis', which proposed to scrutinise music entirely through music.
In retrospect, it might seem surprising that something of the sort had never been tried before. Keller went on to devise another dozen FAs, as he called them, over the next 20 years. If the idea has, so far, failed to catch on more widely, the reasons doubtless range from the paucity of musicians with anything like Keller's variety of skills and depth of insight to the suspicion that an FA of, say, a Mahler symphony, could go on for hours. Keller himself seems to have tacitly acknowledged the drawbacks, since he also continued to pour forth inimitable articles on music up to his death in 1985.
All the same, there can be few scribblers of programme notes, let alone academic analysts, who have not at some time or other doubted the point of what they were doing. Keller himself used to argue that no technical commentary or background information could explain the workings of a masterpiece to a listener who had not instinctively understood it in the first place. His theory of musical understanding was, admittedly, rather specific. What true composers did, he argued, was to summon up the common experience of whichever genre they were writing in - whether symphony or popular song - as a background of expectations, which they then proceeded to modify or contradict point by point in their foreground invention; the tension between foreground and background constituting an exact measure of the new content they were communicating.
It has been objected that such an explanation would hardly account for non-European traditions; that it ignores textural, social, ideological and historical issues which may also impinge on musical meaning, and that it implies a context in which composers and audiences share the same musical backgrounds - something the 20th century has no longer been able to take for granted, with its endless attempts to invent new musical languages (to say nothing of its endless composers' programme notes attempting to explain them). Yet for confirmation of his theory, Keller needed only to point to the Austro- German tradition in which he had been raised. For a substantial proportion of its repertoire was understood and accepted long before anyone ever heard of a programme note.
Those who attended Haydn and Mozart concerts in the late 18th century or Beethoven premieres in the early 19th would have had little more than handbills listing works and movement titles to guide their hearing - perhaps with the addition of a few descriptive sentences once more programmatic pieces, such as Weber's Konzertstuck and Berlioz's Symphonie fantastique, came into vogue. But it was not till the 1820s that something like programme notes in the more familiar form began to emerge, and then probably for social rather than musical reasons: to fulfil the cultural aspirations of the same upwardly mobile middle classes that were anxiously purchasing grammars and books on etiquette.
Long before the end of the 19th century, annotators such as Sir George Grove had arrived at the kind of genteel, pseudo-academic programme note that so provoked the derision of George Bernard Shaw. In our own century, such activities have, of course, spread to the craft of radio announcing and the record sleeve note, and though the genre has thrown up the rare genius - most obviously, Tovey - it is sometimes hard not to feel that the old programme notes so frequently recycled for standard concerts by the London orchestras are designed less to inform the audience than to wrap revenue-raising adverts around.
Nothing daunted, Cambridge University Press has been putting out a new series of music handbooks over the last three years, each of them constituting in effect a book-length programme note on a major work, or group of works, in the concert repertoire. The general editor, Julian Rushton, and CUP's commissioning editor for music books, Penny Souster, seem to have plumped for representing most of the standard figures, but not necessarily by their most obvious works. Twenty handbooks have appeared to date, and despite a brief that has restricted each writer to some 35,000 words and a dozen music examples, the volumes so far have proved nicely varied in approach.
Some writers, such as Malcolm Boyd on Bach's Brandenburg Concertos, or W Dean Sutcliffe on Haydn's String Quartets Op 50, adopt - but adopt very well - a straightforward treatment: the genres are described, the source materials scrutinised, performing problems discussed and each work clearly analysed. Others are more slanted. Elaine Sisman brings 18th- century concepts of rhetoric to bear on Mozart's Jupiter Symphony; Susan Youens is concerned to rescue Wilhelm Muller, the poet of Schubert's Die Schone Mullerin, from charges of mediocrity; Stephen Walsh conducts a mordant inquiry into the dramaturgical puzzles of Stravinsky's Oedipus Rex; and Michael Russ is necessarily preoccupied with the graphic source material of Pictures at an Exhibition - though he also finds illuminating things to say about Mussorgsky's compositional originality.
More questionable, perhaps, is Nicholas Cook's account of Beethoven's Ninth Symphony largely as a story of the ideological distortions subsequent generations have thrust upon it - an exercise in the currently fashionable mode of reception history that will infuriate more traditionalist musicians. Yet James Hepokoski's study of Sibelius's Fifth Symphony is little short of revelatory. For a start, he is the first to have gained access to the complete sketches, and it is as touching to see the familiar tunes emerging from Sibelius's initial doodles as it is awesome to learn the full story of the work's tortuous conception. But within a mere 102 pages, the author also manages not only to outline a novel theory of Sibelius's compositional procedure - which he calls rotation form - but to sketch an entire new overview of the late Romantic-Modern period and Sibelius's place within it.
Hans Keller had a freely confessed deaf spot for Sibelius - the Violin Concerto aside. Yet he would surely have conceded the value of a book that should enable even listeners who have loved the Fifth Symphony to death to approach it again with fresh ears.
Cambridge Music Handbooks: pounds 19.95 each, hardback; pounds 6.95, paperback