Hans Neuhoff, Timo Fischinger and I developed an experiment to explore whether the performance timing patterns found to characterize the jembe piece Manjanin (Polak 2010) would also pertain to perception (paper in review). We found that Malian expert listeners (drummers and dancers) can well perceive the differences between all of the rhythm patterns we used as stimuli, save for short-medium-long (SML) and short-long-long (SLL), which mostly were perceived to be identical. These same two timing patterns (SML and SLL), which are typical of the performance practice of Manjanin in Mali (as shown by Polak 2010 and Polak, London, and Jacoby 2016), were clearly preferred over all others. Not surprisingly, as stimuli deviated further and further from the characteristic SML and SLL patterns, stimuli were less preferred (Neuhoff, Polak, and Fischinger, in press).
This finding is relevant in two perspectives. On the one hand, the nuanced SML pattern prevalent in the performance timings (Polak 2010) pertains only to
rhythm production. Only two different subdivision classes, short and long, can be discriminated in the context of the given rhythmic patterns and relatively fast tempos.
On the other hand, expert listeners discern SML and SLL timings from all other timing patterns including isochrony, and they clearly consider the
the SML/SLL metric pattern to be the better in the context of Manjanin phrases. Evidently, the NI timing pattern not
only is firmly inscribed into the temporal structure of Manjanin as performed, but also is perceptually relevant; it shapes the aesthetic ideal of that piece.