Cecilia Heyes, Cognitive Gadgets. Harvard University Press, 2018. Read in September 2026.
The book’s central thesis is simple to state. Theories of cultural evolution already assume, more or less across the board, that knowledge and skills are learned. In Heyes’s metaphor these are the “grist”: how to make a fire, how to prepare food, how to use a tool, the content of a language (its vocabulary), fishing… There is a fairly broad consensus that all of these are obviously learned rather than innate. Things get more complicated with the cognitive mechanisms that let us use them in the first place; there, theories of origin multiply. These “mills” include imitation (including the sub-skill of picking whom it is worth imitating), language as a capacity, thinking about one’s own thinking, a sense for normative behaviour (what is and isn’t done, and who gets to decide) and “mindreading” – theory of mind, the ability to work out what someone else thinks and knows.
According to Heyes, all of these are tools that each person assembles during childhood out of cultural material, not inborn instincts. Genes supply only a starter kit, and a surprisingly modest one: a taste for being around people and tolerating them, attention that turns of its own accord towards faces and voices, and general-purpose learning somewhat more powerful than a chimpanzee’s. The rest is culture.
With this, Heyes enters a dispute I frankly have no ambition to map. On one side, evolutionary psychology with the mind as a Swiss army knife full of Pleistocene tools. On the other, cultural evolutionists (Joseph Henrich and The Secret of Our Success, say), who rejected modularity and showed that the engine of human success is cumulative culture. But even they, Heyes argues, take the capacity to absorb culture itself as given: for them imitation, teaching and mindreading are genetic adaptations that make culture possible. Culture supplies the grist, the mills are innate. Heyes claims that culture also supplies the mills that do the processing – and those mills then need not be universally human. This relatively small shift has fundamental consequences for how we think about… thinking.
How does she argue?
The strongest chapters are the ones on imitation and mindreading. Imitation looks like something newborns can do out of the box, and there are experiments to back it up: Meltzoff and Moore with their “stick your tongue out at a newborn and it will stick its tongue out more often”. Heyes points out that a large 2016 replication did not confirm this, and that sticking out the tongue is a fairly common newborn reaction to anything interesting whatsoever.
Theory of mind she compares to literacy. Nobody assumes reading is innate even though almost everyone can do it: it needs instruction, a culture that cultivates it, and a brain that adapts to it. The same, she argues, holds for the ability to attribute beliefs and intentions to others. The evidence: children in different cultures acquire the components of theory of mind in different orders, and Samoan children, for instance, pass the false-belief test years later than a module would allow; deaf children of hearing parents who get access to language late also get theory of mind late; and the “implicit” understanding of false belief in toddlers that was supposed to prove innateness falls apart in replications. Most of us evidently acquire the ability to read minds much earlier than the ability to read letters, but we still have to learn it.
The chapter on language is built the same way: Heyes attacks the various versions of “universal grammar” and innate language. I won’t lie, this held no interest for me, because this corner of linguistics sits for me somewhere between “suspicious”, “I don’t understand it” and “I don’t believe it”, so I was easily satisfied with the refutation. Actually, not a refutation: Heyes does not directly and unambiguously refute universal grammar; she rather refuses to grant language the exemption it would get as the only one of the gadgets, and that is enough for her.
And here is the book’s problem, of which the author is aware. The overwhelming majority of the evidence is negative. The procedure is always the same: take a nativist claim, show that it doesn’t replicate or has a learning explanation, and file it under “gadget”. But from the fact that we have no evidence for an innate module it does not follow that we have evidence for cultural construction. Heyes offers two options where in theory there are infinitely many, and various finer-grained spectra. Between a ready-made instinct and a tool built from culture lies ordinary nativism – systems that come pre-prepared but are completed by learning that is steered in advance. Most of the data Heyes cites is explained just as well by this middle model. And perhaps we know too little about the brain; perhaps it is something else entirely. Associative learning and executive functions are supposed to produce the ability to think about what others think: but exactly how, the book promises rather than shows. Heyes is honest about this: she says she is offering a hypothesis and a research programme, not a result. Which is fair: it is a first major push in some direction with a strong model, not a definitive treatment of a theory.
Why do I want to believe it?
Why, then, do I still trust this model more than the alternatives?
It is parsimonious in the right place: it minimises the number of things that had to arise through natural selection, and in a window that is, by evolutionary standards, very short. It doesn’t just produce post-hoc rationalisations full of exceptions, it makes predictions: it counts on cultural variability, on the possibility of losing these tools, on their rapid transformation – and at the same time it sits firmly with the historical record. A human fifty thousand years ago had practically the same genes but a visibly different mind and a different ability to work with it.
It is also worth admitting that her theory suits me emotionally. If these tools are a matter of culture, they are deeply ingrained but tractable: they can be trained, neglected and lost. The environment that shapes them matters: family, school, institutions, screens. It also fits what we fear in the debate about children and screens – although the evidence there is thin so far, and this is exactly the spot where I want to be careful about what I want to believe. And it carries real hope in that it rather convincingly refutes the basic maxims of the type “the human mind was fully formed in prehistory and ever since our conflicts have just been tribal warfare writ large, nothing changes”. We bear considerably more responsibility for ourselves and our environment.
It has practical consequences for how to live, and it confirms a certain view of education and upbringing: learning is not the mystical unlocking of what is already in the child, but the construction of tools that would not exist without deliberate work. And that holds even for things we take for granted, like the ability to understand that the other person doesn’t know what I know: a child who can’t manage this in third grade need not be written off as lost forever; it can be worked on.
And then there is the angle on artificial intelligence, which Heyes does not yet develop in the book but which follows from it directly: if human cognition is largely a technology assembled from cultural material, then there is no reason to be surprised that a machine fed cultural material can do things we took to be exclusively human. That doesn’t mean machines suddenly become people: it means that we, as humans, are not exclusive in our capacity for cultural learning. The question is not whether it is “real thinking” (and what kind of category that even is), but which tools and packages it lacks – and perhaps will lack forever, perhaps not.
To me this looks like a stronger model than most of the competition. But Heyes does not deliver proof that it is so. She delivers a model that can be worked with well, provided it is worked with carefully: as a hypothesis, not as a result. Little that is innate, a ton of cultural tools, and consequently more weight on will and on what we choose to do.