Native language determines parsing of nonlinguistic rhythmic stimuli.

Kiyomi Kusumoto and Elliott Moreton

Given a rhythmic sound stimulus, the hearer automatically parses it
into recurrent units.  Researchers since Woodrow [Arch. Psych. 14,
1-66 (1909)] have consistently found that stimulus elements differing
in duration are parsed iambically, while those differing in intensity
are parsed trochaically.  Hayes [Metrical Stress Theory (1995)]
observes that no language has quantity-insensitive iambic feet, and
suggests that the nonlinguistic parsing tendencies may be at the root
of this language universal.  However, the relevant nonlinguistic
parsing experiments were done with native English speakers.  Using
comparable nonspeech rhythms, we compared parsing behavior in native
speakers of English and in two dialects of Japanese.  Our findings
suggest that language determines parsing preference, not the other way
around.

[Abstract in:  Journal of the Acoustical Society of America 102 (5,
Part 2):3204 (1997).]