Detaching rhythm and meter from harmony rather in the way that Meyer prioritizes melody (but, like Meyer, preserving a loose sense of hierarchy), Wallace Berry proposes that rhythm merges into meter at larger spans and thus multiple accent streams can be read hierarchically, with a layering of accents analogous to the layering of structural levels in a Schenkerian analysis – the difference being that the accents at the upper end of the hierarchy acquire their position by cumulation, rather than syntactical differentation.
Such expressive accent groups have the advantage that they can model the dramatic or emotional curve(s) of a composition's unfolding far better than analyses that rely on metric regularities or the hierarchies of harmony. As a matter of method, however, Berry demands too much of the analyst – every level requires considerable, intuitive gathering and sorting of accents, and the resulting analysis graph can never reflect the complexity of those decisions. In the case of D779n13, reasons for the choice of the "primary" accent are not likely to be obvious, because one element of the decision is a denial of expectations: our structural highpoint or highest-level accent, m. 31 (see the graphic below), should have been as loud as the forte of the C#-major phrase, and it should have received a strong hypermetric accent (which we can confer on it but only with the help of explanations such as those engendered by Lerdahl and Jackendoff's various preference rules). The measure is marked partly by these notable absences, and in both cases, the preceding bars of A7 chords rob m. 31 of those features. Thus, even at the end there remains residual doubt about whether the primary accent belongs to the C#-major chord of m. 23, the A7 chord of m.29, or the B-minor 6/3 chord with its suspension in m. 31. The deciding factor, I think, is register, as depicted in the lower part of the example, where a steady progress of the initial F# across the piece obtains, and the moment of arrival coincides with an inversion of the initial soprano-alto interval. After this moment, the reprise and final cadential ascent sound "anticlimactic" – in Berry's terms "reactive" and recessive. (Berry's priorities resemble Robert Fink's, a fact which has motivated my notation using the angled arrow/beam.)
Berry's conception of musical hearing is the endpoint in a line that began with Schenker and moves through Meyer. All believe in hierarchy: Schenker's is the strictest, with its single generating structure determining priorities throughout the levels; Meyer loosens this to permit multiple simultaneous patterns but he clearly believes in relative significance based especially on a shorter-scale/larger-scale distinction; Berry believes in multiple, autonomous streams which establish hierarchies by statistics, the accumulation of coincident accents (one might say that this concept of meter is at the top of the hierarchy). Its combination of multiple streams of activity and dramatic accent makes Berry's method the most cinematic of any mode we have considered so far – indeed, his notion of levels of metric accent as applied to music is indistinguishable from Michel Chion's "audiovisual phrasing" as applied to a film soundtrack.
References.
Berry, Wallace. "Metric and Rhythmic Articulation in Music." Music Theory Spectrum 7 (1985): 7-33.
Berry, Wallace. Structural Functions of Music. Englewood Cliffs, NJ: Prentice Hall, 1975; reprint ed. New York: Dover Books, 1987.
Fink, Robert. "Going Flat: Post-Hierarchical Music Theory and the Musical Surface." In Nicholas Cook and Mark Everist, eds. Rethinking Music, 102-137. 2d ed. New York: Oxford University Press, 2001.
Chion, Michel. Claudia Gorbman, tr. Audio-Vision: Sound in Film. New York: Columbia University Press, 1994.
