
An incomplete but perhaps defensible mini-history of the birth of cognitive science might go like this:
Cognitive science emerged in the 1950s in reaction to the then-dominant scientific paradigm of behaviorism as championed by B.F. Skinner, often described as the most influential psychologist of the 20th century, and for a time arguably the most famous scientist in the world.
The basic idea of behaviorism is that the mind doesn’t matter. Because we cannot observe what is happening inside our brains, the story goes, it is essentially impervious to the tools of scientific inquiry. Instead, when it comes to understanding humans, behavior is all we have to guide us. And our behavior is entirely the product of just two factors: (1) our genetic endowment, the capabilities bequeathed to us via our DNA; and (2) the conditions of our surrounding environment, and the rewards and punishments it offers.
The only way to affect the first factor is through genetic engineering, but Skinner mostly steered clear of that minefield. Instead, he trained his attention on transforming human capabilities by manipulating the environment—he favored nurture over nature. His term for this environmental manipulation was “operant conditioning,” and at heart it rests on the proposition that our individual capabilities can be engineered to improve. Skinner was so convinced of this from an education standpoint that he spent several decades on a fruitless quest to build “Teaching Machines” that would administer a programmed curriculum to students using operant-conditioning techniques.1

It's hard to imagine today just how widely Skinner’s ideas permeated throughout universities in the 1950s. Scholars who expressed an interest in “mentalism” were derided by Skinner and his broad network of acolytes as being no different than those who once attributed the movement of the heavens to the whims of divine beings. But in the span of a few decades, both the ideas of behaviorism and Skinner himself would end up marginalized and largely discredited, scientifically and socially.
What happened?
Well, the cognitive revolution happened. Despite Skinner’s insistence that mental states were illusory, a motley crew of heretics—including Noam Chomsky, Herbert Simon, and George Miller, among many others—began to develop theories of how our minds work that were capable of being tested scientifically.2 Cognitive scientists thus began to demonstrate empirically that we could make robust inferences about mental processes by theorizing a cognitive architecture of information processing that could be studied, tested, and emulated. This not only led to a dramatic improvement in our understanding of human thought, it also spawned breakthroughs in simulating our mental processes through technology—computers and artificial intelligence are both born of the cognitive revolution. (We’ll circle back to this shortly.)
In response to this, Skinner dug in his heels. In fact, he doubled down on behaviorism by taking it to its ultimate sociopolitical conclusion: That all of human society should be structured to maximize productivity through operant conditioning. Mass social engineering, in other words. And to achieve this utopia humans must rid themselves of their quaint, anti-scientific notions about freedom or dignity—hence the title of his 1973 book, Beyond Freedom and Dignity.
This book ended Skinner’s career as a public intellectual. In one of the most savage book reviews you’ll ever read, Noam Chomsky systematically shredded both Skinner’s description of science as well as his (Skinner’s) blighted vision for humanity. As to the former, Chomsky noted that Skinner simply assumed that the role of science was to explain Skinner’s behaviorist theories, but as Chomsky dryly observed, “surely the task of a scientific analysis is to discover the facts and explain them.” And as to Skinner’s larger political aims, well, Chomsky did not hold back:
Skinner is saying nothing about freedom and dignity, though he uses the words "freedom" and "dignity" in several odd and idiosyncratic senses. His speculations are devoid of scientific content and do not even hint at general outlines of a possible science of human behavior. Furthermore, Skinner imposes certain arbitrary limitations on scientific research which virtually guarantee continued failure.
As to its social implications, Skinner's science of human behavior, being quite vacuous, is as congenial to the libertarian as to the fascist.
And that was pretty much that for behaviorism, at least as championed by BF Skinner.
“That’s a mildly interesting history Ben,” I’d like to imagine you’re thinking, “but why are you sharing it now?” Fair question. Let me briefly offer two scattershot reasons why understanding this history of cognitive science helps me navigate larger scientific and sociopolitical debates swirling around AI at the moment.
Let’s start with Chomksy. As he was preparing to slay Skinner’s behaviorist dragon (as my friend Audrey Watters quips), Chomsky was also advancing his theory of “Universal Grammar,” an idea that would come to dominate the field of linguistics for several decades. In brief, Chomsky argued that humans must genetically inherit general information about the structure of languages because we are capable of inferring grammatical rules even at a very young age, at a time when we lack the necessary external knowledge to make such inferences (the “poverty of the stimulus”).
For reasons I won’t go into here, by the time ChatGPT came into our lives Chomsky’s theory was already on the scientific ropes, but AI may have delivered the final knockout blow. That’s because, as linguistic researchers such as Kyle Mahowald and others have described, large-language models demonstrate “formal linguistic competence” on par with humans, meaning, LLMs consistently produce text that conforms to the abstract, semantic rules of human languages. This is a big deal! And it’s upended broad swaths of the field of linguistics, as this wonderful if wonky oral history explores in detail. Among other things, researchers are now establishing “a direct mapping between internal LLM representations and neural activity patterns within the language network [in the human brain]. This mapping can be successfully used to predict brain responses to novel sentences and words in previously unseen contexts” (again quoting Mahowald).
In other words, when it comes to how humans develop their linguistic abilities, LLMs offer great scientific promise for helping us to understand ourselves. This, by the way, is why I have low tolerance for claims that AI is “fake.” For all my criticisms of the how the technology is being deployed commercially—and I have many!—to deny any possibility it might help us to better understand ourselves strikes me as fundamentally anti-scientific.
In response to this, Chomsky—much like Skinner before him—has dug in his heels. He continues to vehemently deny this technology has anything useful to offer us scientifically, it is merely “superficial and dubious.” But surely the task of a scientific analysis is to discover the facts and explain them—and not cling to a theory in the face of mounting evidence against it?
The other reason I keep recuring to the history of cognitive science versus behaviorism is that it reminds me that ideas rarely disappear, they simply transform. This should be the subject of a separate essay, but I see a direct through line from Skinner’s behaviorism to the modern claim of scientists such as Robert Sapolsky who contend that all of human behavior—indeed, all of the universe—is “determined” by forces outside of our control. Sapolsky believes this counsels us toward a progressive, tolerant society that would not mistakenly hold people accountable for what they have no power over, but in providing a readymade scientific justification for any social arrangement as the product of naturalistic forces set in motion in antiquity, well, his ideas are as congenial to the technofascist as they are to the socialist.3
That’s too much for one penultimate paragraph, I’m sorry. Behaviorism makes me cranky. Remember kids, read your history. Let’s close out with some Donovan:
Histories of ages past
Unenlightened shadows cast
Down through all eternity
The crying of humanity

Audrey Watters has written a terrific history of Skinner’s failed attempts at this, and his career more broadly, titled Teaching Machines: The History of Personalized Learning. Much of my understanding of Skinner draws from her work.
We cannot and should not ignore that cognitive science as a field was steered almost entirely by white men, as I’ve explored previously.
Mercifully, Kevin Mitchell has undertaken the slog of batting down Sapolsky’s claims one by one—start here if you’re curious. We can also see echoes of determinstic behaviorism today from Charles Murray (on the right) and Kathryn Paige Harden (on the left) when they argue that much of who we are, and our capacity to change and learn, has been written into our genetic code. That’s another essay I’ll need to write at some point.
I stand corrected on the Transformer invention date.
Rhetorical Q, though: Were transformers more of an evolution (a difference in degree) or was it an entirely new technology (a difference in kind)? Seems to be the former... (https://www.linkedin.com/pulse/story-ai-evolution-before-ml-era-transformers-gpt-3-beyond-ghosh/). The foundations for it were laid by Hinton when both of us were still kids, right?
~~
The cultural transference of language thing is interesting. Although, at first blush, I am not seeing anything in it that precludes there being a UG. More of a belt-and-suspenders thing. Alone, it doesn't seem to explain the robustness of language growth in any given person. Indeed, at least one cultural-based refutation didn't stand up to scrutiny back in the day (https://cognitionandculture.net/wp-content/uploads/nevinsEtAl_07_Piraha-Exce.pdf)
Final thought on Noam: I noticed earlier that you posed Noam as someone who sort of unscientifically dismissed his critics, while simultaneously dinging him for evolving his theories when presented with evidence. Can't have it both ways. ;-)
But, this was a great discussion (at least for me). Thank you!
_Mark
I don't disagree with you here (although the date of the citation you mention doesn't predate "the recent advances in AI"...LLMs and transformers go back farther, it's the training based on all the stolen data that is the recent "advance", I think).
And, all of the evolution you mention both in Chomsky's thought and in the critics—those all still exist in Chomsky's sandbox, yes? I mean, NO ONE thinks the infant brain is a tabula rasa, anymore, right?
As an attorney, I am just always sensitive to undermining my own point by using opposing frames. "Functional Competence" (FC) is how the AI/"AGI"-hypesters want to talk about this, because it elevates prep for the job market to the #1 goal of education.