Jay Phillip Jefferson
Published: 2020
Total Pages: 0
Get eBook
Macaque (Macaca spp.) societies represent complex adaptive systems marked by multifaceted, heterogeneous social interactions among groups members. Such dynamics allow for considerable variation to emerge with regard to social structure both within as well as across the species of this highly radiated genus of primate. Within primatology, societies are characterized by dominance styles that may be more despotic-like (i.e., generally more hierarchal) or more egalitarian (i.e., generally less hierarchal) depending on a suite of interdependent behavioral factors that tend to co-vary with each other along a spectrum of tolerance. Moreover, tolerance can be considered the most basic form of conflict management insofar that macaques simply decide to not engage in conflict with one another when competing for resources. The goals of the research presented in this dissertation were to further investigate how tolerance shapes social organization in macaques through the use of an agent-based model (ABM) while additionally exploring a robust parameter space to understand how other mechanisms, such as access to conflict information and movement behavior, may also lead to variation in interaction patterns among individuals. In the first chapter, I review major contributions made to the field of primatology through the use of ABMs. Furthermore, I present the ABM DomWorld as a general model and common ancestor that ties together over two decades of ABM research on primate social behavior into a model phylogeny. I treated 24 publications related to DomWorld as the operational taxonomic unit and documented changes in the inherent property of agents, agent behavior, as well as the research focus of the publications in order to construct a phylogentic tree of the models described therein. I took this methodological approach of relating models to demonstrate how the development of ABMs contributes to how we structure existing theory or even allow for new theoretical frameworks to emerge. That is, coming to a fuller understanding of the dynamic aspects of our research via model phylogenies helps us also understand how the theory that guides our work evolves with each subsequent iteration and adaptation of our existing models. In the second chapter, I introduce the M.A.C.S. (Macaques as A Complex System) model that I developed and created to analyze how tolerance influences social interactions among simulated macaque ("agents") groups. I programmed agents within my ABM to randomly move about and engage in lower-level (conflicts) or higher-level (escalated fights) aggression within an exclusively social space. Each society was initialized with agents having a uniform tolerance and avoidance attitude (ranging from 0 (less tolerant/avoidant) to 1 (more tolerant/avoidant)) toward one another as well as an equal probability of defeating each other when engaging in dominance interactions. Furthermore, I ran 20 replications of each unique tolerance simulation. Additionally, an agent's dominance probability of defeating another during a dominance interactions was decided through statistical procedures that incorporate both an agent's individual experience with previous conflicts, but also access to group level information about all other agents' conflict histories. I analyzed how group-level patterns of behavior emerged at different tolerance values, including the frequency of conflicts and fights between dyads that differentiated into dominant and subordinate roles and dyads whose dominance relationship remained uncertain. I found that group-level patterns of aggression that correspond well with real-world primate societies emerged in the M.A.C.S. model by simply manipulating the value of tolerance at initialization. That is, the formalizing of dominance relationships combined with an initially intolerant disposition allowed for the emergence of more frequent intense aggression among initially intolerant societies. However, lower-level aggression between uncertain ranked dyads became more frequent among initially tolerant societies, as observed in more egalitarian-like, real-world species of macaque. In the third chapter, my goal was to extend our baseline model from the previous chapter and test how variation in the information agents use to determine their dominance relationships with their interaction partners as well as the movement strategies that agents use to explore the simulation space effected group-level outcomes. I tested a social learning condition in which agents only possessed information about one's conflict history as well as information about conflicts that occurred within their social space. Thus, the social learning condition effectively leads to agents that possess different interpretations of what the actual distributions of conflict outcomes are. I also tested an individual learning condition in which agents only know about their own conflict history and nothing of the conflict occurring among other group members. Additionally, I tested two movement strategies in which agents choose another agent to aggregate towards during a movement bout (directed condition) or will move in space based on a vector calculated from the avoidance attitudes an agent has towards the other agents in its social space. Furthermore, I analyzed the interplay between these information and movement conditions. I found that decreasing the dominance information made available to agents decreased the frequency of all interactions between dominant and subordinate agents, but increased the frequency of all interactions between dyads with uncertain rank relationships. I also found that when agents utilized a cautious movement strategy there is a considerable increase in both conflicts and fights compared to the random or directed movement conditions. These findings provide further insight into how information processing and movement ecology may influence hierarchal differentiation and social structure in primate groups. While further research will incorporate social network analysis to better elucidate the causal mechanisms underlying the M.A.C.S. model, the research presented in this dissertation provides a promising step forward. Specifically, in how the abundance, acquisition, processing, and use of dominance information tied with an explicit parametrization of tolerance can be used to simulate a complex primate social system and further understand macaque social organization. This is particularly important for understudied species among the different dominance styles within the genus. Also, my work serves as a robust foundation for how we may develop the next iterations of the M.A.C.S. model in our newly presented model phylogeny. Finally, I provide further recommendations for how future iterations may implement other elements of macaque social behavior not featured in current ABMS, such as status signaling, social power, and explicit third-party conflict intervention decisions, into the M.A.C.S. model.