To call or not to call

We have investigated the vocal communication of chimpanzees and our findings challenge the traditional view of their communication system and highlight some new important similarities with human language. Studying the communication of non-human primates (primates) can help us understand how and when elements of human language might have evolved. Whilst previous research had shown us that monkeys can extract complex information from calls, we know far less about great ape vocal communication and the mechanisms that underlie call production. The traditional view is that primate vocalisations are individualistic expressions of emotion, that are not produced voluntarily or intentionally to communicate with others. We aimed to test this traditional view of vocal production in our closest living relative, the chimpanzee.

We conducted experiments with wild chimpanzees to test the extent to which food-associated calls are (i) automatic, involuntary responses to food or (ii) signals directed at specific individuals in a flexible manner. Once a lone individual had fed silently for 5 minutes we broadcast an individually distinctive vocalisation from a group member to simulate the arrival of that individual into the area. We found that chimpanzees did not produce calls in response to the food alone, nor to the arrival of any individual; instead they only called when the playback simulated the arrival of a high ranking friend. This experiment showed that chimpanzees selectively direct calls to specific individuals.

We also examined whether wild chimpanzees produce alarm calls intentionally to warn others of danger, or rather as individual expressions of fear. Chimpanzees were presented with a model python when alone and in a group. We evaluated the vocal behaviour of the chimpanzees against established criteria for intentional signalling and found good evidence that at least two call types are produced intentionally. These calls were (i) socially directed and given when friends arrived, (ii) associated with visual monitoring the audience and gaze alternations, and (iii) goal directed, as calling only stopped when recipients were safe from the predator. Thus, the vocal production of our closest living relatives can be intentional and chimpanzees produce alarm calls to warn others of danger.

Although our results from the wild show chimpanzees can use calls intentionally, our captive study suggests that they may have limited capacity to alter the acoustic structure of their calls. We recorded the structure of food-associated calls produced by two groups of chimpanzees before and after they were integrated into one single group. We found that prior to integration the two groups produced calls with different structures for certain food types. One year after integration and exposure to the other group's calls for the same food, we found no evidence of call structure change. While this could indicate a lack of ability to converge on a new group label for a food, it could also reflect lack of motivation to do so: an analysis of social data indicated that one year after integration, the chimpanzees were still preferentially interacting with members of their original group. To distinguish between these explanations we have just completed collecting another set of vocal data, two years after integration, where there is more evidence of interaction between all individuals.

In sum, our project has shown that chimpanzees share key hallmarks of human language production in that they can direct calls at specific recipients in a voluntary and intentional manner, but they may have limited ability to dynamically change the structure of their vocalisations. These results are the first to systematically demonstrate similarities between human language and ape vocal communication in terms of intentional call production. Our findings support theories suggesting language originated directly from the vocal communication system of our ancestors.