Our analyses indicate that the vocal repertoire of Formosan macaques conforms to the pattern predicted by the law of brevity, with more frequently uttered calls being shorter in duration. This provides, to our knowledge, the first evidence of such a pattern in the vocal communication system of a non-human animal. While there are clear and fundamental differences between human language and the vocal communication of other primates (Pinker 1995
; Bickerton 2003
), our findings tentatively suggest important shared ground between these communication systems in terms of the basic rules governing signal length and frequency of use. Moreover, our results suggest that similar evolutionary forces may have acted to increase the efficiency of coding in both human and non-human primate vocal communication, or that common self-organization principles (Köhler 1989
) underlie these different communication systems.
Previous attempts to explain variation in the form of calls within vocal repertoires have often focused on the information content of these signals, for example, investigating how internal factors such as motivational state may affect the spectral characteristics of vocalizations (Morton 1977
; Hauser 1993
). Other studies have explored how call structure, including call duration, reflects adaptation to reduce signal distortion in specific environments (e.g. Ey et al. 2009
). Our findings indicate the importance of also considering coding efficiency as a force shaping the temporal characteristics of calls. Simultaneously investigating the information content of vocalizations, their degradation within the species' ‘acoustic habitat’ (Brown & Gomez 1992
) and their coding efficiency may prove to be a useful approach in understanding the selective forces shaping the size and structure of animal vocal repertoires.
It is important to note that the law of brevity explored here is different from Zipf's law (Zipf 1949
), which explores the relationship between total frequency of occurrence of signals and their rank order (i.e. first, second, third … most commonly used). While Zipf's law has been more widely applied in studies of animal communication (e.g. McCowan et al. 1999
; Hanser et al. 2004
, but see also McCowan et al. 2005
; Suzuki et al. 2005
; Ferrer-i-Cancho & McCowan 2009
), only the work of Ferrer-i-Cancho & Lusseau (2009)
has previously tested for patterns consistent with the law of brevity in non-human animal communication. Application of this approach to the vocal repertoires of other primate and non-primate species, as well as to other modalities of non-human animal communication, is now needed to investigate the generality of the relationship between signal length and frequency of use.
A valuable extension of this work would be to explore a second, and perhaps equally important, component of the efficiency of the coding system, namely the energetic cost of call production. In addition, it would be informative to explore how call function affects call duration over and above any relationship to frequency of use; alarm calls, for example, may be short not just owing to coding efficiency but also to reduce conspicuousness. Finally, and particularly in light of recent theories surrounding the role of gestural communication in language evolution (Corballis 2003
; Arbib et al. 2008
; Tomasello 2008
), it would be informative to test for patterns consistent with the law of brevity in primate gestural repertoires, to see if they too share this fundamental characteristic of human language.