My colleagues and I have begun to build a model of the development of young infants’ physical reasoning. The model is based on the assumption that infants are born not with substantive beliefs about objects (e.g., intuitive notions of impenetrability, continuity, or force), as researchers such as Spelke and Leslie have proposed, but with highly constrained mechanisms that guide the development of infants’ reasoning about objects. The model is derived from findings concerning infants’ intuitions about different physical phenomena (e.g., support, collision, and unveiling phenomena). Comparison of these findings points to two developmental patterns that recur across ages and phenomena. We assume that these patterns reflect, at least indirectly, the nature and properties of infants’ learning mechanisms. In this review, I describe the patterns and summarize some of the evidence supporting them.

Our perceptual systems are products of evolution and have been shaped, in part, by natural selection. It is widely assumed that natural selection favors veridical perceptions – that is, perceptions that accurately describe aspects of the objective world relevant to fitness. This assumption has been tested using the mathematics of evolutionary game theory. It is false. Monte Carlo simulations reveal that veridical perceptions are never more fit, and generically are less fit, than nonveridical perceptions of equal complexity that are tuned to fitness. Veridical perceptions go extinct, and their extinction rate increases as complexity increases. These results motivate a new theory of perceptual systems – as species-specific interfaces shaped by natural selection to hide objective reality and guide adaptive behavior. For Homo sapiens, space-time is the desktop of the interface and physical objects are icons on the desktop. The shapes and colors of physical objects no more resemble objective reality than the shapes and colors of desktop icons resemble files in a computer.

Key words: evolution, natural selection, evolutionary game theory, perception, bayesian inference, veridical perception, illusion.

What is the nature of human learning, and what insights can be gained from understanding early learning in infants and young children? This is an important question for understanding the human mind, the origins of knowledge, scientific reasoning, and how to best structure our educational environment. In this article, we argue for a new approach to cognitive development: rational constructivism. This view characterizes the child as a rational constructive learner, and it sees early learning as rational, statistical, and inferential. Empirical evidence for this approach has been accumulating rapidly, and a set of domain-general statistical and inferential mechanisms have been uncovered to explain why infants and young children learn so fast and so well.

Key words: rational constructivism, learning mechanisms, cognitive development