PROBLEMS OF POETICS the spectator may not notice it, there is a great deal of editing, with films having an average shot length of three to five seconds. Unlike the calligraphic trend, this approach avoids exaggerated action, complex camera movements, and dynamic montage. Unlike the pictorialist trend, this style does not rely on opaque shots; each composition is highly legible. This style's closest kin is Hollywood cinema 1917-25, the one-bit-of-information-per-shot approach of Fairbanks, Lloyd, Lubitsch, and William De Mille. There is also a certain playfulness about the style, as unexpected cuts and sudden screen entrances disturb the limpid flow of information. Examples would be Shingun (Marching On, 1930), Naruse's Kimi to wakarete (Apart from You, 1933) and Yogoto no yume (Nightly Dreams, 1933), Gosho's L'Amour (1933) and Izu no odoriko (Dancing Girl of Izu, 1933), Shimazu's Our Neighbor Miss Yae (1934), and Shimizu's Arigato-san (1935). All these trends offer the director ways of expressively or decoratively elaborating the narrative point. Yet to isolate them as I have done should not occlude the fact that many films furnish mixed cases. Town of Love, for example, fundamentally adheres to the piecemeal style, but it has 'calligraphic' moments as well. It is evident that Ozu worked with the piecemeal style - partly because of its generic predispositions, partly because of Kido's policy, and partly because of the possibilities it holds for mixing playfulness and rigor. In Chapter 5, I shall discuss how Ozu recast the piecemeal approach into his own distinctive stylistic system. But we should note at this point that Ozu's films are far purer instances of the piecemeal approach than we find in other directors, and this points to a more general line of reflection. Crudely speaking, poetic theory affords two principal conceptions of the relation between the major artist and the norms of his or her period. There is the 'consummate expression' model, whereby the artist perfectly embodies the tradition. Burch writes that 'there is no break in continuity between [the work of Ozu and Mizoguchi] and that of a large body of directors working in similar genres.'14 This ignores, however, Ozu's major differences from Japanese norms. The rigor of his style resembles neither the narrative linearity nor the decorative eclecticism present in the work of ordinary filmmakers. Ozu's work is, for better or worse, eccentric. (His peers thought so too.) Are we then justified in introducing a second model, one of conflict between the artist and the reigning norms? Such a view often assumes that transgression is necessary for broader aesthetic change. Yet Ozu does not fit this role easily either. He was not at war with norms; he simply sought a greater concentration, consistency, and individuality of style than his milieu required. In effect, he sought to purify and deepen the piecemeal approach to decoupage. Our study will reveal him to be far more stylistically rigorous (even in his earliest films) than were his contemporaries; far more stringent in his choice of elements; far more bold in his patterned expansion of such canonized elements as interruptive cutaways and deep space; and far more systematic, concentrating upon tactics by which stylistic organization, springing from the need to transmit and decorate narrative, becomes a vehicle of self-conscious narration and at times the dominant element of the film's structure. 24