The evolution of cognitive biases in human learning

Area of Science:

• Evolutionary psychology
• Cognitive science
• Human behavior

Background:

• Cognitive biases like underinference and the hard-easy effect are evolutionarily puzzling.
• Traditional models focus on individual learning from environmental feedback.

Purpose of the Study:

• To explain cognitive biases from an evolutionary perspective.
• To propose a model of ancestral human social learning.

Main Methods:

• Developed a computational model of ancestral human learning.
• Incorporated cultural-evolutionary theory and Bayesian principles.
• Simulated social learning with imitation and innovation.

Main Results:

• Fitness-maximizing strategies in the model led to cognitive biases.
• High costs of retaining environmental feedback favored social learning.
• Model reproduced recurrently non-monotonic confidence estimates.

Conclusions:

• Cognitive biases may be adaptive byproducts of social learning.
• Social learning, not individual feedback, was the primary selection pressure.
• Selective social learning can explain non-monotonic confidence patterns.