Acquisition of object-robbing behavior in macaques: After stealing inedible & more or less valuable objects from humans, they appear to use them as tokens, by returning them to humans in exchange for food

Acquisition of object-robbing and object/food-bartering behaviours: a culturally maintained token economy in free-ranging long-tailed macaques. Jean-Baptiste Leca, Noëlle Gunst, Matthew Gardiner and I. Nengah Wandia. Philosophical Transactions of the Royal Society B: Biological Sciences, Volume 376, Issue 1819, January 11 2021. https://doi.org/10.1098/rstb.2019.0677

Abstract: The token exchange paradigm shows that monkeys and great apes are able to use objects as symbolic tools to request specific food rewards. Such studies provide insights into the cognitive underpinnings of economic behaviour in non-human primates. However, the ecological validity of these laboratory-based experimental situations tends to be limited. Our field research aims to address the need for a more ecologically valid primate model of trading systems in humans. Around the Uluwatu Temple in Bali, Indonesia, a large free-ranging population of long-tailed macaques spontaneously and routinely engage in token-mediated bartering interactions with humans. These interactions occur in two phases: after stealing inedible and more or less valuable objects from humans, the macaques appear to use them as tokens, by returning them to humans in exchange for food. Our field observational and experimental data showed (i) age differences in robbing/bartering success, indicative of experiential learning, and (ii) clear behavioural associations between value-based token possession and quantity or quality of food rewards rejected and accepted by subadult and adult monkeys, suggestive of robbing/bartering payoff maximization and economic decision-making. This population-specific, prevalent, cross-generational, learned and socially influenced practice may be the first example of a culturally maintained token economy in free-ranging animals.

4. Discussion

This field observational and experimental study of token-robbing and token/reward-bartering interactions in the free-ranging population of Balinese long-tailed macaques produced three main findings: (i) these behaviours need to be learned throughout juvenescence (i.e. until up to 4 years in this species) to be successfully performed; (ii) older monkeys preferentially selected tokens that were more valued by humans; and (iii) these more skilful and selective individuals appeared to make economic decisions, as evidenced by clear behavioural associations between value-based token possession and quantity or quality of food rewards rejected and accepted.

(a) Experiential learning

As predicted, we found a significant increase in token-robbing success from juveniles to subadults to adults, whereas the main behaviour patterns required for the successful performance of token/reward-bartering interactions were already in place from around 4 years (i.e. in subadults). Likewise, the ability to engage in more negotiated successful token/reward-bartering sequences—during which the monkey only returned the token after being proposed more food rewards, or after rejecting more food rewards, or after accepting a type of food reward different from the one(s) previously rejected—was not fully acquired before the subadult stage.

These results lend some support to the ‘experiential learning' hypothesis, whereby token-robbing and token/reward-bartering interactions are multi-stepped and complex behavioural sequences requiring perceptual learning, sensorimotor coordination and cognitive skills (e.g. memory, associative learning) to be successfully performed; they are thus gradually acquired through extended individual practice during the juvenile period, in part via experiential trial-and-error learning. It is noteworthy to mention that the development of (sub)adult-level proficiency at robbing/bartering is not only dependent on skill learning (e.g. detection, sneaky approach, self-control), but may also be constrained by physical maturation. This is particularly true during the token-robbing phase that often involves monkey–human body contact and/or requires muscular strength when a monkey has to yank on a flip-flop still worn by an adult human. In these cases, the limited physical capabilities of juveniles, and the maturing bodies of subadults, may partly explain the significant increase in token-robbing success from juveniles to subadults to adults.

Primates are characterized by the longest juvenile period in relation to life span of all mammals [20]. According to the ‘needing-to-learn' hypothesis [20], prolonged juvenility is associated with behavioural patterns that necessitate acquiring a proportionally large amount of information and/or skills to reach adult competence before individuals become reproductively mature. These behaviours include extractive foraging techniques [21] and (socio-)sexual behaviour patterns [22]. Our study indicates that both phases of the robbing/bartering practice also required experiential learning to be fully mastered.

(b) Value-based token selection

The first step in economic decision-making requires the cognitive ability to distinguish among different expected material values of a given symbolic currency (e.g. tokens, cash, virtual money). After showing that token selection was not significantly affected by token availability and the relative ease with which different types of tokens were stolen by the monkeys from human targets, our observational data revealed a marked age difference in how the monkeys responded to a human-based three-level hierarchy of valuable objects. When considering the token selection among all the prospective human targets (i.e. temple visitors with potential tokens available in a given area), juveniles did not show any preferential selection among low-valued, medium-valued and high-valued tokens, whereas subadults and adults preferentially selected high- and medium-valued tokens over low-valued ones. When two tokens of different values were available on a given human target, subadults and adults preferentially selected the higher-value token, whereas juvenile individuals did not show any significant difference. We found a similar age difference in the value-based token selection, after experimentally controlling for token accessibility and lateral bias.

These results support the ‘value-based token selection' hypothesis, positing age differences in the selection of higher-valued tokens by the monkeys during the token-robbing phase that are indicative of a developmental trajectory toward more strategic choices in more mature individuals. Subadult and adult monkeys (but not juveniles yet) have learned to map their token-robbing behaviours onto the hierarchical (and arbitrary) scale of values attributed by humans to different tokens: they preferentially selected tokens that were more likely to be exchanged for food (e.g. electronic devices, pairs of glasses) over other objects that were less valuable for humans and typically not worth bartering (e.g. empty camera bags, hairpins). Our findings are consistent with data obtained in other non-human primate species, showing that subadult and adult capuchin monkeys and chimpanzees correctly preferred a high-valued token over a low-valued token in an experimental bartering situation [1,11,23].

(c) Robbing/bartering payoff maximization

The second step in economic decision-making requires the cognitive ability (i.e. mental processes involving associative learning and memorization) to respond differently to differentially valued tokens by trying to maximize one's payoff. We found evidence for such behavioural associations between value-based token possession and quantity or quality of food rewards rejected and accepted by subadult and adult monkeys (i.e. the most skilful and selective individuals) during the token/reward-bartering phase. They consistently and actively obtained either more food rewards or a more preferred food reward in exchange for a higher-valued token. They were also more likely to end a successful bartering interaction by accepting a less preferred food reward in exchange for a lower-valued token.

Our findings support the ‘robbing/bartering payoff maximization' hypothesis in that subadult and adult monkeys strategically responded to differentially valued tokens in their possession by adjusting the amount or type of food rewards they gained from the barters. The result showing that subadults (unlike adults) failed to significantly reject more low-preferred food rewards before returning a higher-valued token may be explained in terms of poorer temporal cognition or higher impulsivity (compared to adults): subadult long-tailed macaques may either not have yet acquired the cognitive capacity to anticipate the subsequent proffering of more preferred food rewards in this specific situation, or not be patient/self-controlled enough to wait for possibly more preferred food rewards.

Overall, our field observational data are in line with laboratory-based studies showing that several non-human primate species can (i) understand the effectiveness of tokens as secondary reinforcements to make simple calculations about quantities of reward, (ii) determine an item's value on the basis of its perceived utility (e.g. exchanging only a low-preferred reward for a tool necessary to reach a more preferred reward) and (iii) recognize the appropriate conditions in which a successful exchange could occur (e.g. presence/absence of the experimenter, safe/risky experimenter) [1,2,23–27]. Other cognitive skills and temperamental traits exhibited to varying extents by non-human primates engaging in token-aided economic behaviours include preference transitivity, self-control, delay of gratification, action planning and calculated reciprocity, because they may facilitate or constrain an individual's ability to make optimal economic decisions [1,5,23,28]. Even though these characteristics were not explicitly examined in this study, some of them will be the subject of our future observational and experimental investigations.