In the 1970s and 80s, Gopnik and her colleagues [...] focused not on what children said, but on what they did in creative and problem-solving situations. “It turned out that even the youngest babies already knew more and learned more than we ever would have thought,” Gopnik went on.

This was one of the things that really surprised me when becoming a parent. If you take the time to just observe very small children doing their thing, they are surprisingly methodical, consistent, and hypothesis-driven. It looks like chaos at first, but then I've sort of come to realise that it's really about performing actions that are likely to trigger a response with high information content.