Adherence to Menzerath's Law is the exception (not the rule) in three duetting primate species

Abstract

Across diverse systems including language, music and genomes, there is a tendency for longer sequences to contain shorter constituents; this phenomenon is known as Menzerath's Law. Whether Menzerath's Law is a universal in biological systems, is the result of compression (wherein shortest possible strings represent the maximum amount of information) or emerges from an inevitable relationship between sequence and constituent length remains a topic of debate. In non-human primates, the vocalizations of geladas, male gibbons and chimpanzees exhibit patterns consistent with Menzerath's Law. Here, we use existing datasets of three duetting primate species (tarsiers, titi monkeys and gibbons) to examine the wide-scale applicability of Menzerath's Law. Primate duets provide a useful comparative model to test for the broad-scale applicability of Menzerath's Law, as they evolved independently under presumably similar selection pressures and are emitted under the same context(s) across taxa. Only four out of the eight call types we examined were consistent with Menzerath's Law. Two of these call types exhibited a negative relationship between the position of the note in the call and note duration, indicating that adherence to Menzerath's Law in these call types may be related to breathing constraints. Exceptions to Menzerath's Law occur when notes are relatively homogeneous, or when species-specific call structure leads to a deterministic decrease in note duration. We show that adherence to Menzerath's Law is the exception rather than the rule in duetting primates. It is possible that selection pressures for long-range signals that can travel effectively over large distances was stronger than that of compression in primate duets. Future studies investigating adherence to Menzerath's Law across the vocal repertoires of these species will help us better elucidate the pressures that shape both short- and long-distance acoustic signals.

Phylogeny and mechanisms of shared hierarchical patterns in birdsong

“…Indeed, that patterns resembling ML are stronger in voiced speech than in written text is consistent with biophysical contributions to ML. 3,39,51 Our observations of ML among 15 songbird species can provide the foundation for further cross-species comparisons on the extent and strength of patterns resembling ML (even across vocal non-learning species 4,7,9,19,20 ). While it is difficult to directly compare the strength of the relationships observed in our analyses to those in other species (because of differences in quantification and analytical approaches), a couple of published studies provide useful comparisons.…”

“…Given these observations, it will be interesting to reanalyze behavioral sequences that fail to conform to ML using these randomization techniques to identify additional organizational factors that reduce ML-like patterns. 20,40 Our unified approach to the analysis of ML across songbird species allowed us to directly compare the strength of ML across taxa and across types and levels of song organization. In particular, examining the degree to which observed slopes deviated from chance allowed for more direct comparisons Phylogenies depict the difference from chance for the observed slopes (in Figures 2, 3, and 4).…”

“…Here, we examined the degrees to which birdsong exhibits patterns resembling ML across 15 songbird species and to which observed patterns differ from null models based on random sequencing (Figure 1) and vary across organizational levels and phylogeny. In addition, because ML can be observed in the unlearned vocalizations of non-human animals, 4,9,19,20 we also evaluated how sensory experiences during development affected the emergence of and species differences in ML in two songbird species in order to reveal the contribution of learning and motor production biases to ML.…”

Untitled

“…Besides, similar analyses can be extended to any system that forms a hierarchy of units of different levels. Consequently, Menzerath's law has also been observed in genomes [17][18][19], protein domains [20], penguin vocalizations [21], chimpanzee gestural communication [22] and primate vocalizations [23][24][25][26][27], including strepsirrhine species [28]. It is related to compression principles [23,29] and explained in terms of oral communication as follows: Given that lungs' air capacity is limited, when longer words are vocalized then their constituents tend to be more compressed because the speaker must 'hurry' to finish articulating the word before running out of air [12,25].…”

“…However, some authors have discussed whether MAL or Menzerath's law could be trivially explained in particular cases [19,30] or sometimes, just quietly not be observed [26]. Nonetheless, the lack of a standardized methodology of studying these statistical laws introduces a large variation in the results depending on the segmentation method selected, the units of study, and some additional variables that we will address here [10,14,31].…”