Children are using a lot of the techniques for learning that scientists use:
they're building models and everyday theories about the world around them
they "play", essentially understanding causal linkages by experimenting
kids can use complicated patterns of statistics, to figure out causal structures in the world
kids are better learners than adults in that they can figure out "unusual" or "unlikely" solutions faster than adults who stick to rigid strategies of solving problems
kids are designed for optimal learning; they are in perpetual explore mode
chimps are mature by 7 years old
even in hunter-gatherer cultures, humans weren't mature until at least 15 years old
the protected period of childhood (neoteny) allows for an extended period of learning and exploration and in adulthood they turn to exploit modes, making them efficient "doers" and efficient specific-problem-solvers
AI systems don't currently learn anywhere near as well as kids
children vs adult models of learning and behaviour are evolution's way of optimizing solving for individual contextual problems
childhood is super explorative where you don't have to worry about safety or feeding yourself or anything other than to optimize for learning a whole bunch of contextually relevant information
then in adulthood, we shift into exploit mode where you can efficiently solve complex problems quickly and effectively
now think of cultural evolution where you combine the extended periods of neoteny with exploit modes of adulthood that put to efficient use the predictive inference exploit modes
cultural packages then get passed down through authority/top-down teaching and modeling practices, as well as tool-use, school indoctrination, politics, work imperitives, and other social modes of forced conformity
The vast majority of learning that happens in the first few months of life is fundamentally social (Trevarthen, Schore, Stern, Fogel, Beebe, Tronick)
Developmental psychology has long held that infants begin by understanding "simple" things like objects and then as they become more sophisticated, they start "getting" other humans. This makes little sense evolutionarily (and on many other levels). They START social, then move to objects.
infants are motivated beyond instincts to be fed and safe; they want to connect and synchronize with other minds/bodies
Communication developmentally starts with interactive mimicry that involves emotions communicated through body-language that form proto-narratives, then develop into language, and then grow into cultural knowledge.
infants come equipped from birth to infer subjective states in receptive social partners (especially mothers). This happens far earlier than is assumed in ToM research.
Now we need to link cultural evolution packages (Henrich, Boyd, Muthukrishna) and humans' innate receptivity to "receiving" those packages through intersubjective channels initiated from birth (Trevarthen, Gopnik, Smith, Fogel, Stern, etc)
These are some bread-crumbs that combine: ontogeny, phylogeny, and scientific thinking modes (still to elaborate)