Learning consists of selecting the rules of grammar which identify from the large set of possible expressions in the Language of Thought the ones that best fit the data.

The mental representations of humans are much more structured than those of our competitors.

At birth a human brain has two key ingredients:

the machinery that makes it possible to generate a bunch of abstract expressions in the Language of Thought, and

the ability to choose among these expressions according to probabilities inferred from data using Bayes updates.

Our adult judgements combine two levels of insights:

the innate knowledge of our species, that is the set of plausible hypothesis inferred from evolution, and

the mentals models fine-tunned by personal experiences.