Evolutionary game dynamics of controlled and automatic decision-making

Danielle F. P. Toupo, Steven H. Strogatz, Jonathan D. Cohen, David G. RandarXiv 2015 · DOI · arXiv

Abstract

We integrate dual-process theories of human cognition with evolutionary game theory to study the evolution of automatic and controlled decision-making processes. We introduce a model where agents who make decisions using either automatic or controlled processing compete with each other for survival. Agents using automatic processing act quickly and so are more likely to acquire resources, but agents using controlled processing are better planners and so make more effective use of the resources they have. Using the replicator equation, we characterize the conditions under which automatic or controlled agents dominate, when coexistence is possible, and when bistability occurs. We then extend the replicator equation to consider feedback between the state of the population and the environment. Under conditions where having a greater proportion of controlled agents either enriches the environment or enhances the competitive advantage of automatic agents, we find that limit cycles can occur, leading to persistent oscillations in the population dynamics. Critically, however, these limit cycles only emerge when feedback occurs on a sufficiently long time scale. Our results shed light on the connection between evolution and human cognition, and demonstrate necessary conditions for the rise and fall of rationality.

Introduction

In psychology and neuroscience, "dual-process" theory suggests our decisions emerge from two systems: an automatic one (fast, rigid, "hardwired") and a controlled one (slow, flexible, capable of planning). While controlled processing is a hallmark of human intelligence, evolutionary game theory (EGT) has traditionally assumed a single mode of decision-making. This paper addresses a fundamental evolutionary puzzle: If controlled, rational processing is so beneficial for long-term survival, why hasn't it completely outcompeted automatic processing? Furthermore, why does human history seem characterized by cycles of societal advancement followed by collapse and a return to short-sightedness?

Methods

The researchers developed an analytically tractable mathematical model using the replicator equation to describe how the frequency of strategies changes in a population based on relative fitness.

Key components of the model include:

  • Intertemporal Choice: Agents forage for resources. Automatic agents have a competitive advantage ($\beta$) in speed—they grab resources first—but suffer from diminishing marginal returns due to immediate consumption. Controlled agents are slower but "pace" their consumption to maximize fitness.
  • Diminishing Returns ($a$): A parameter representing the utility loss from immediate versus spread-out consumption.
  • Environmental Feedback: The model makes environmental parameters—resource availability ($\rho$) and competitive advantage ($\beta$)—dynamic variables that respond to the population's makeup with a specific time lag ($\tau$).
  • Bifurcation Analysis: Nonlinear dynamics are used to map stability regions, identifying where the population reaches a steady state versus where it begins to oscillate.

Key Results and Insights

The authors successfully categorized the ($\beta$, $\rho$) parameter space into five distinct regimes of dominance, coexistence, and bistability.

Figure 1: Stability regions in the ($\beta$, $\rho$) parameter space.

A striking insight is the "Success Breeds Failure" loop: controlled agents are often the architects of their own demise. If controlled agents enrich the environment (e.g., through technological innovation), they increase resource density. This higher density increases the relative advantage of "fast" automatic agents, allowing them to invade and eventually dominate.

Furthermore, the model predicts the emergence of limit cycles—persistent oscillations between rationality and myopia. These cycles only emerge when there is a significant time delay between a change in the population and the environment's response.

Figure 4: Time series showing persistent oscillations (limit cycles) between the proportion of controlled agents ($x$), automatic agents ($1-x$), and the competitive advantage ($\beta$).

Discussion and Limitations

The study demonstrates that under specific conditions (high lag and moderate resource levels), a population will forever swing between being mostly "rational planners" and mostly "impulsive consumers." However, the authors acknowledge several simplifications: the model treats strategies as binary rather than a continuum, assumes basic foraging rules, and lacks spatial structure. Future research could incorporate these elements to provide a more cognitively realistic version of evolutionary game theory.

Summary updated

Abstract

We integrate dual-process theories of human cognition with evolutionary game theory to study the evolution of automatic and controlled decision-making processes. We introduce a model where agents who make decisions using either automatic or controlled processing compete with each other for survival. Agents using automatic processing act quickly and so are more likely to acquire resources, but agents using controlled processing are better planners and so make more effective use of the resources they have. Using the replicator equation, we characterize the conditions under which automatic or controlled agents dominate, when coexistence is possible, and when bistability occurs. We then extend the replicator equation to consider feedback between the state of the population and the environment. Under conditions where having a greater proportion of controlled agents either enriches the environment or enhances the competitive advantage of automatic agents, we find that limit cycles can occur, leading to persistent oscillations in the population dynamics. Critically, however, these limit cycles only emerge when feedback occurs on a sufficiently long time scale. Our results shed light on the connection between evolution and human cognition, and demonstrate necessary conditions for the rise and fall of rationality.
BibTeX
@misc{toupo2015evolutionary,
  author = {Danielle F. P. Toupo and Steven H. Strogatz and Jonathan D. Cohen and David G. Rand},
  title = {{Evolutionary game dynamics of controlled and automatic decision-making}},
  year = {2015},
  eprint = {1507.01561},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1507.01561},
}
Figure