Sporadic articles for such editors as HL Mencken he had begun writing back in his Harvard days just after the First World War - an apprenticeship that culminated in his first book, The State of Music, published in 1939: a mordant survey of the artistic pretensions and economic vicissitudes of modern musicians. But only in 1940, after 15 expatriate years in Paris, where he had initially gone to study together with his lifelong friend Aaron Copland, did he accept a professional post on the New York Herald Tribune: beginning, as he intended to go on, with an attack on the New York Philharmonic as irrelevant to the city's intellectual life plus a side swipe at Sibelius's then vastly popular Second Symphony as "provisional beyond all description".
Over the next 14 years, an impressive array of prejudices were to be liberally indulged: against such performers as Toscanini and Heifetz, against Germanic music in general, and against Benjamin Britten, whose American operetta Paul Bunyan he seems to have regarded as a personal challenge. But he was also far more responsive to the latest from Stravinsky, Schoenberg, Bartok, Webern or, for that matter, to jazz and gospel music, than most of his New York colleagues, and better informed in the other arts, particularly literature and painting; moreover, readers could never be sure his insouciant prose was not about to throw up some original insight or long-pondered proposition.
How he managed it all is something of a mystery, since he was well known for groaning "No good can come of this" during the opening bars of concerts or operas and sleeping soundly through to the end. Yet the immaculate notices he regularly turned in continued to resurface in anthologies long after he had retired from the Tribune in 1954. Meanwhile, in 1966 he brought forth his engaging autobiography, Virgil Thomson by Virgil Thomson; in 1970, a canny if erratic survey of American Music Since 1910; and, shortly before he died in 1989 at 93, a bouncy Selected Letters. Yet, once he had made a somewhat belated start in his early twenties, Thomson undoubtedly thought of himself primarily as a composer, steadily accumulating a colossal output over the next 70 years. And, having realised by the later 1920s that his French contemporaries such as Milhaud and Poulenc were not going to accept him as an honorary member of Les Six, he proceeded to renounce their naughty sophistications so completely that Copland was later to call him "a kind of inverse iconoclast: his music was so plain and direct, so baby-simple, so to speak, that it was startling".
Of course, simplicity can mean a lot of different things, as the history of 20th-century music so abundantly shows. For a start, an urge to simplify, to substitute for the diffuse, vast, ramified post-Wagnerian language of such late Romantics as Mahler, Richard Strauss and early Schoenberg, something starker and more condensed, was a strong impulse in Modernism itself: whether in the textural purging of a Webern or the rejection of conventional harmony in order to concentrate on the rhythm of a Varese.
Indeed, the injunction to clarify was duly to pass to such relatively traditionalist social consciences as Copland, Shostakovich and Britten. Granted, there were also naturally complex minds such as Schoenberg and Boulez who immediately proceeded to elaborate the new clarities; but these, in turn, have tended to reanimate the simplifiers, so that the classical Minimalism of early Riley, Reich and Glass, for instance, is probably better understood as a Modernist reductio (ad absurdum?) than an anti- Modernist movement.
On the other hand, there has been a recurrent enterprise to bring simplicity in to so-called art music by appropriating and "treating" material from ancient, exotic or vernacular sources: the neo-primitivism of Stravinsky's Rite of Spring; the folkloristic inflections of Bartok; the music hall, jazz and cafe chantant cartwheels of Les Six - or the more acidulated Weimar Republic vernacular of Weill. And, rather removed from either of these tendencies, there have been a few solipsists producing pieces so singular and complete within their limitations - Satie's Gymnopedies, for instance, ever revolving in their own serene space, or certain of Morton Feldman's tranced traceries - that their simplicity amounts to a mystery.
Yet, although Thomson revered Satie, particularly the hieratic word-setting of Socrate; although he invoked a comparable range of vernacular sources to Les Six, out of popular music and his Midwest boyhood; and although he deliberately unlearnt a sophisticated technique - he was, after all, a pupil of Nadia Boulanger - his approach to simplicity was quite distinct. His central tenet was that modern music had become too self-conscious and that true composers ought to be able to compose off the top of their heads without too much effort, or even attention - and be prepared to stand by the results. This he called the Discipline of Spontaneity. One of his gambits was to sit friends down and, as quickly as possible, to compose their portraits - a whimsical notion he pursued so implacably that eventually there were more than 100 such pieces. Yet Picasso seemed to assent when he told Thomson, "Ah, yes. I understand. If you are in the room while I am working, whatever I do is automatically your portrait."
It was, indeed, the subject of one of Picasso's most famous portraits who really got Thomson going. Intoning Gertrude Stein's neo-nonsense at the piano, he found he could easily conjure up complementary accompaniments. The result was the opera Four Saints in Three Acts (1928), which actually contains four acts, more than 30 saints and no plot at all - amounting, in that sense, to the first purpose-composed "producer's opera" ever.
On the page, the music looks like a blandly inconsequential stream of waltzes, patter songs, hymns and tangos - its orchestration wheezy with accordion vampings. Yet, as sung and danced by an all-black cast, with cellophane designs from Florine Stettheimer and a scenario of saintly processionals and epiphanies ("Pigeons on the grass alas") staged by the young Frederick Ashton, the 1934 premiere exerted such mesmeric charm that it subsequently ran for six months on Broadway. Whether Thomson's other Stein opera, The Mother of Us All (1947), pulled off a comparable trick, whether his later Lord Byron (1968) really is the crowning opus he intended, is difficult to know because neither have ever received really first-class professional productions.
And that is the crux. To the extent that his enduring quest seems to have been for the poetry of ordinariness - the musical equivalent, perhaps, of the Red Wheelbarrow school of William Carlos Williams - Thomson's scores look fatally easy to bring off and have too often been poorly performed and recorded. Not that he lacked successes: his score for the film Louisiana Story (1949) received a Pulitzer Prize, even though his treatment of its Cajun melodies was straightforward to the point of anonymity. Nor was his music always so sweetly homespun: by means of multiple doublings, his "orchestral picture" Wheatfield at Noon (1948) piles up a dense atonality with an almost childlike empiricism. What continues to bemuse, perhaps to disturb, is that a sophisticated musician should have intentionally devoted his long life to defining an idiom that often sounds virtually indistinguishable from the jottings of a not very talented amateur. Yet his criticism remains one of the sprightliest records of the musical mid- century. He did achieve something special in Four Saints. And one can think of not a few portentous or doctrinaire composers active today who might benefit from a brief submission to the Discipline of Spontaneity