The uproar has scarcely ceased since. How could Stravinsky refute a view that had seemed obvious at least since Monteverdi? How could he deny the deep religious feeling of Bach, the sonorous ideology of Beethoven, the psychological insight of Wagner, or the evocative correspondences of Debussy - whatever he thought of their actual music? Come to that, how could he forget the magic of his own Firebird, the pathos of his Petrushka, the primitive excitement of his Rite of Spring? In fact, putting those achievements into perspective, if not actually denying them, was part of the subtext of Stravinsky's Autobiography. Still only 30 when the riotous reception of the Rite consummated his world fame, he had to face correspondingly early the choice of whether to go on exploiting a successful formula or whether to risk everything on new departures.
In the event, it took some years for Stravinsky's choice to become apparent, partly because of the disruptions of the First World War and partly because the premiere of the culminating masterpiece of his Russian period, Les Noces, was held up until 1923. But already in The Soldier's Tale (1918) and other, shorter pieces composed in his Swiss exile during the later war years, he was constructing little forms from Western vernacular sources - waltzes, tangos, ragtimes - rather than from Russian folklore. Then, with the Communist takeover in the October Revolution, he was literally cut off from his homeland and language (not to mention his private income). When Diaghilev flourished a bunch of Baroque pieces purporting to be by Pergolesi at him in 1919 and asked if he would like to arrange them as a ballet, Stravinsky seems to have leapt at the chance to remake himself as a composer. Forty years later, he was to recall: 'Pulcinella was my rediscovery of the past . . . but it was a look in the mirror too.'
And so arose the dreaded notion of Neo-Classicism, which ever since has been the standard explanation for Stravinsky's output over the following three decades up to The Rake's Progress of 1951 - after which, he is supposed to have belatedly salvaged his radical credentials by going over to a post-Schoenbergian serialism. Yet Stravinsky disliked the label, and even on the most superficial level it hardly begins to account for the real variety of his middle years. If anything, his typical figures, rhythms and textures during the period owed more to the Baroque than to Classicism. While the Piano Sonata (1924) or The Rake's Progress may invoke Beethoven or Mozart, the Piano Concerto (1923), Violin Concerto (1931) and the chamber concerto Dumbarton Oaks (1938) are all redolent of Bach. Yet Romantic elements are also to be found: Verdi behind the ostensibly Baroque facade of Oedipus Rex (1927), Delibes in the bland diaphony of Apollo (1928), Tchaikovsky all over the place after The Fairy's Kiss (1928). And later, Renaissance, even medieval mannerisms were to inflect the shadowy elegance of Orpheus (1947) and the serenely hieratic Mass (1948).
Yet possibly the most intriguing fact about Stravinsky's middle-period music is that it has been understood quite differently by three successive generations. The musical intelligentsia between the wars seemed amused enough at first by what it heard as his almost annual changes of style, but tended to regard these as the smart moves of a deracine creator who had left his soul in the Russian period. By contrast post-war listeners, prompted by such avant-garde ideologues as Boulez, revalued his early works not for their Russianness but their radicalism - writing off the so-called Neo-Classical decades as a reactionary interregnum. Yet today it is difficult to hear the stylistic references of the middle-period output as less Stravinskian than those of before or after, so that the question arises whether all the allusions, the Russian and serial no less than the Neo-Classical, were successively required and consciously invoked to realise afresh an essentially unchanging set of basic compositional procedures.
Which brings us back to the famous ruling about expression. Considered dispassionately, it is not so very startling. What Stravinsky was suggesting is that a sweeping phrase is not intrinsically romantic nor a minor key intrinsically sad, but that such devices have acquired their apparent expressivity by constant association, over the Western centuries, with certain texts, rituals or dramatic situations. Theorists who take an opposing, Expressionist, view might of course ask why those devices had been associated with those particular texts, rituals or dramatic situations in the first place. But Stravinsky was not denying that such associations should be used - indeed, from the austerely sacred Symphony of Psalms (1930) to the frivolously profane Jeu de Cartes (1936), his middle period is crammed with traditional expressive affects: but affects invoked, placed, slanted in peculiarly Stravinskian ways as if, far from comprising the essence of the music, they were elements as objectively deployed as his many stylistic allusions simply to articulate, to colour the surface of something more constant.
What was that something? In the Autobiography, Stravinsky proposes, almost theologically, that 'Music is the sole domain in which man realises the present' - that is, the only means of establishing stability amid the experiential flux of past and future. Since the language of late-Romantic music, and in particular Wagner's notion of 'music as the art of transition', depended upon the manipulation of that experiential flux, and since the young Stravinsky showed singularly little talent for handling it, his cleaving to an ideal of present-centred objectivity might be considered as a mere aesthetic compensation. And if music history itself is conceived as an evolutionary flux of past and future, then his position might also be construed as a pretty conservative one.
In fact, as his earliest music before the first great ballets already suggests, Stravinsky was in many ways a natural conservative. Even the apparent iconoclasm of The Rite of Spring had more to do with restoring the distant past to the present, than with any deliberate attempt to invent the Modernist future. What his mature aesthetic enabled him to escape was precisely the apparent ongoing imperative of historicism, allowing him instead to renew his art through the continual reshuffling and cross-cutting of history, from the archaic (medieval canons) to the futuristic (serialism), for his present purposes.
Because Stravinsky's sustaining of a constant immediacy depends so much upon his denial of flux, there will always be aficionados of the expressive surge who find his apparently static, cut-and-paste constructions frustrating. Yet the distinguished line-up of outside presenters Radio 3 has been able to call on to introduce next week's mini-festival of middle-period Stravinsky - including Richard Alston, Joanna MacGregor and John Cox - only emphasises how widely an output once considered merely artificial, modish or reactionary has come to appeal, delight and indeed to move.
Composer of the Week, Radio 3, 9am from Monday. Also on Radio 3: Violin Concerto, 11.30am Tuesday; Persephone, 11.30pm Tuesday, repeated 12 noon Wednesday