Lessons from the marmosets

By Paul Bouissac

New study of infant marmosets, Callithrix jacchus, yields insights into the development of vocal learning that are relevant across species, including Homo. Rather than being the “hopeful monster” of Chomskyan lore, this research suggests that human language may share the same developmental program as other primates as well as songbirds (Science vol. 349, issue 6249, 14 August 2015, pp. 688-9 and 734-38).

Investigations of the ways in which animals communicate by signals in various sensorial modalities usually take for granted a fundamental difference with the nature of human languages. There have been many descriptions of repertories of postures or patterned movements as well as distinct acoustic productions which can be associated with typical situations and elicit specific responses in a great variety of species. The easiest data to collect are alarm calls which are within the range of our own discriminating auditory capabilities. But over the last decades advanced acoustic recording made it possible to refine the scientific descriptions of these signalling repertories and to show they can function referentially contrary to earlier assumptions.  For instance, Cheney and Seyfarth (1990) identified distinct alarm calls referring to hawks, leopards, and snakes, in Vervet monkeys , Cercopithecus aethiops, each triggering an appropriate escape behavior. In search of evolutionary antecedents to human language, some were inclined to claim that such signals were arbitrary. Of course, their distinct acoustic formants could simply be determined by the position of the head with respect to predators that necessarily appear on different levels of the monkeys’ horizon. But this could not question their eventual functional codification. Rendall et al. published in 1999 a important article on “The meaning and function of grunt variants in baboons” and later presented evidence of similarity between the formant profiles of human vowels and vowel-like baboon grunts (Rendall et al. 2005, Journal of Acoustical Society of America 117:944-955).

However, these synchronic descriptions of vocal repertories within well-defined groups did not raise the issues of their universality across the species considered, nor did they address the problem of their acquisition by individual animals during their developmental stage. With respect to the former issue, it has been noted that when two groups of captive chimpanzees are merged, the “immigrants” learn and adopt the particular “food grunts” which are used by the residents and differ from their own. The latter problem – “language” development — is more difficult to investigate in species which mature relatively slowly. However, marmoset infants have the marked advantage of maturing in two month, thus providing the possibility of observing their vocal productions from birth to the time when they are fully socialized in their group.

As it has been observed in human language development, marmoset infants utter a rich babbling that includes a wide array of phonetic productions which get progressively pruned toward the restricted repertories of meaningful signals used by the adult marmosets of the group within which they are born. These characteristics offered ideal conditions for Takahashi et al. (Princeton Neuroscience Institute) who report the results of their research in “The developmental dynamic of marmoset monkey vocal production” in the section Language Development of Science. As usual for articles considered particularly important, Science features in the same issue a shorter account of the main article, aimed at less specialized readers who can nevertheless appreciate the impact of the results in the broader context of language evolution.  The findings point indeed “to an ancient substrate for vocal learning that an evolving large hominid brain could take advantage of, thus continuing the evolutionary process that has enabled communication in other animals” (689). In their lucid summary of the Takahashi et al. article, Margoliash and Tchernichovski include a comparative visual rendering of the clusters of vocal sounds in marmosets, birds, and humans in early development, late development, and adulthood. It shows on the first level a wide distribution of acoustic data. The second level reveals several distinct clusters resulting from the progressive selection of formants. On the adulthood level, much smaller sets correspond to the contrastive functional acoustic signals used in communication among the species concerned. By playfully entitling their comments “Marmoset kids actually listen”, they point to the vocal learning process that is observed across the three species which are known to develop system of acoustic communication.  

In Takahashi et al.’s words: “Our findings demonstrate that infant marmoset calls undergo dramatic changes during the first two months of life, transforming from cries into mature adult-like phee calls. The timing of this transition is partly attributable to maturation but is also influence by contingent parental vocal feedback. This is consistent with preverbal vocal development in humans, whereby (i) natural categories of sounds change as respiratory, laryngeal, and facial components mature, and (ii) in parallel, vocal feedback sensitizes infants to certain features of those sounds, and the sounds are modified accordingly. […] The complex and socially dependent vocal development we observed in marmoset monkeys may be a necessary condition of the vocal learning observed in humans”.

These observations open an interesting vista on the origin of language in humans. A crucial article by Ackermann, Hage, and Ziegler (“Brain mechanisms of acoustic communication in humans and nonhuman primates: An evolutionary perspective”, Behavioral and Brain Sciences, 2014 December: 37 (6) 529-546) provides an enlightening theoretical and empirical framework for understanding the significance of the above observations with respect to the origin of oral language in humans. Other recent observations bear upon the temporal organization of distinct sounds in patterns akin to syntax (e.g., Gustison and al. 2016; Thompson 2016; Zielinski 2015; quoted in Science News 25/04/2016: https://www.sciencenews.org/article/gelada-monkeys-know-their-linguistic-math . While there is evidence that speech and gestures developed in synergy in Homo, the common ancestor of primates must have relied exclusively on acoustic communication as a vital social resource. Indeed, a tree-dwelling organism needs all its limbs for the primary functions of locomotion and stabilization in a mostly opaque environment of trunks, branches, leaves, and other vegetal growths in which distal visual communication would hardly be adaptive. The communicative functions of gestures can only have emerged when hominins became bipedal in relatively open space. Visual signaling could then expand the reach of the messages and possibly secure them from exposure to eavesdropping competitors and predators. The visual medium carries information farther than sound waves of the kind humans usually produce.

Note 

For over a decade, research has been conducted on marmosets’ communication at the University of Zurich by a team under the leadership of Dr. Judith Burkart who writes: “We are currently studying dialect differences in different marmoset colonies, and find that they actually are distinct. We then moved animals from one of these colonies to our one, to see if they adjust their call structures — which they do! Finally, we paired marmosets from Zurich with animals from this other colony, and measured both the development of their pair bond and how much they accommodate their vocal output. It seems that the pairs who like each other better also accommodate more to each other. As a consequence, we became very interested in vocal accommodation in human and non-human primates and have written a review article, currently under review in Biological Reviews”. (personal communication, October 18, 2016)