Tolerant and despotic macaques show divergent temperament but similar theory of mind

Abstract

The social intelligence hypothesis proposes that the demands of social life shape the evolution of cognition, but different aspects of social interactions may be relevant. To test how competitive versus cooperative interactions shape social cognition, we assessed multiple metrics of social cognition in Barbary macaques (Macaca sylvanus, n = 40) and rhesus macaques (Macaca mulatta, n = 60). These closely related species have similar social organization, but diverge in social styles: Barbary macaques are more tolerant, whereas rhesus macaques are more despotic. Monkeys completed a battery of experimental tasks measuring gaze-following (co-orienting with others), knowledge attribution (representing others' underlying knowledge states), goal attribution (interpreting others' actions in terms of underlying intentional goals) and temperament (boldness in response to exploring novelty). While the rhesus macaques were more willing to approach a novel object than were Barbary macaques, both species showed similar success in each social task. However, individual Barbary macaques were more likely to show greater overall proficiency across all social measures combined than were individual rhesus monkeys. Overall, these results indicate that similar social cognitive capacities may evolve in distinct social contexts, and suggest socio-cognitive skills may be relevant for both competitive and cooperative interactions in primates.This article is part of the Theo Murphy meeting issue 'Selection shapes diverse animal minds'.

Keywords: comparative cognition; primates; social cognition; theory of mind.

Figure 1.

Methods for the cognitive battery. (a) Setup for the gaze-following task: a human demonstrator looked either up (test trials) or down (control trials). (b) Example monkey gaze-following responses: left panel shows monkey looking at the actor, and right panel shows upwards look for a rhesus macaque (top row) and a Barbary macaque (bottom row). (c) Setup for the knowledge attribution task: a human demonstrator watched a fruit move on an apparatus and reached either to where it went (expected outcome) or to an empty location (unexpected outcome). (d) Example monkey looking-time responses: left panel shows monkey looking to the apparatus, and right panel shows looking away from the apparatus for a rhesus macaque (top row) and a Barbary macaque (bottom row). (e) Setup for the goal attribution task: a human demonstrator first reached for one of two possible items, and then, when their locations were switched, either continued to reach for it in a new location (expected outcome) or performed the same reach on a new object (unexpected outcome). (f) Setup for the temperament task: a monkey could choose whether to approach a fake-baited food box and a novel object (a toy) placed on the ground. (g) Example monkey approach responses: a rhesus macaque approaching the novel object (top panel) and a Barbary macaque choosing not to approach (bottom panel).

Figure 2.

Social cognitive performance across species. (a) Proportion of monkeys that looked up in test versus baseline control trials in the gaze-following task. Successful monkeys should look up more often in test trials than in control trials. Error bars indicate s.e. (b) Looking times in expected versus unexpected trials in the knowledge attribution task. Monkeys should look longer at unexpected than expected test events. (c) Looking time change (habituation looking – test looking) for the expected and unexpected trials in the goal attribution task; relative change in the test trial compared with the paired habituation trial is depicted because monkeys differed in their initial interest in the different sets of objects used in each pair of trials. Monkeys should look longer at unexpected than expected test events, after accounting for this baseline difference. In boxplots, hinges indicate the lower and upper quartiles, the horizontal line represents the median, diamonds indicate the mean, and whiskers indicate the minimum and maximum range of data; outliers are plotted as individual points.

Figure 3.

Boldness and social cognitive performance across species. (a) The proportion of monkeys in each species that approached the food box and the novel object in the temperament task. Error bars indicate s.e. (b) Variation in social task difference scores indexing individual’s overall performance in that task, relative to individual performance in the temperament task. Positive difference scores indicate response patterns concordant with task comprehension (e.g. looking up more in test than control trials in the gaze task, or relatively longer looking at the unexpected compared with the expected test trials in the knowledge and gaze attribution tasks). In boxplots, the horizontal line represents the median, diamonds indicate the mean and whiskers indicate the minimum and maximum range of data; outliers are plotted as individual points.

Figure 4.

Relationships between social cognitive tasks and species comparison of overall performance. (a) Bivariate correlations between the difference scores in the social cognition tasks; tasks did not show significant shared variance. (b) Average summary scores for Dimension 1 extracted from the PCA; the species did not differ in this integrative measure of social performance. Hinges indicate the lower and upper quartile, the horizontal line represents the median, diamonds indicate the mean, and whiskers indicate the minimum and maximum range of data; outlier is plotted as an individual point.