Plato’s Warning Has Been Vindicated | THR Web Features | Web Features | The Hedgehog Review

THR Web Features / August 5, 2026

Plato’s Warning Has Been Vindicated

Why AI Has Brought Nothing More than the Conceit of Wisdom.

Michael Marinaccio

( THR illustration; King Thamus with the ibis-headed Thoth/Theuth, Egyptian god of writing.)

Michael Marinaccio

Michael Marinaccio is Executive Director of the Center for Responsible Technology in Washington, DC.

Categories

Essays

Related Topics

Technology

Self

History

Literature

Share

Open any parenting book from the past fifty years, and you will find a section about technology that starts something like this: “Socrates predicted the book would destroy society, but that never happened—so don’t worry about your kids watching television, using iPads, or having smartphones.” So why am I not reassured when I hear folks resort to that chestnut to promote “artificial intelligence”?

Whether they know it or not, the laissez-faire techno-optimists in those parenting manuals are taking issue with the words of King Thamus, the legendary Egyptian monarch invoked by Plato in the Phaedrus. According to the parable, Theuth, the father of writing, boasted that his invention would aid human memory. Having none of it, King Thamus retorted that those who read will lose “their memory and become forgetful” because they will rely on books instead of their brains. People will be “thought very knowledgeable when they are for the most part quite ignorant,” and because they are filled “with the conceit of wisdom instead of real wisdom, they will be a burden to society.” That prediction may have been premature, but I contend that the widespread embrace of AI thoroughly vindicates Plato’s warning. As more and more people buy into the promises of this new technology, we are witnessing the full consummation of the “conceit of wisdom” that was made possible with the advent of writing and reading.

Why did Plato’s argument take more than two millennia to be borne out? Literacy did change the conditions for the handing down and acquisition of knowledge, easing some of the burden that oral cultures placed on memory while also promoting the rapid acceleration and dissemination of more abstract systems of thought, such as philosophy and science. The printing press followed the manuscript and codex, and, for the most part, human societies flourished. However, the same media that made it possible to store knowledge without relying on memory formalized and codified the underlying grammatical rules of languages. It is these rules that would eventually enable the prediction machines we call AI to produce a simulacrum of human language. This was a “language” without any connection to the generative conditions of true human language: that is, the experiences of actual living beings interacting with other living beings.

Many think of grammar as a set of rules governing the correct usages within a language, but grammar is also how the mind comprehends the rules of language and gives meaning to speech. Aristotle tells us in On Interpretation that “spoken sounds are symbols of affections in the soul, and written marks symbols of spoken sounds.” In a simpler way, grammar tells us how words fit together to form coherent, meaningful thoughts, ideas, and sentences. It was Marshall McLuhan, the Canadian philosopher and godfather of media theory, who wrote in his 1943 doctoral thesis The Classical Trivium, “Grammar is the art of gathering and interpreting congruous instances, whether phenomenal or textual,” that form our understanding of reality—what the Greeks called the logos. Only within a framework of grammar can we understand and creatively interpret the world, our politics, and history.

The reason large language models (LLMs) pose such a threat to knowledge is that they pretend to be grammar in the deepest sense, when they are, in fact, mathematical manipulations of grammatical rules. Words are, by definition, human utterances that are exclusive to logos and human intelligence. “[I]f words belong exclusively to intelligence, then what is produced by a thing that lacks intelligence cannot in actual fact be a word,” explains the philosopher D.C. Schindler. “It can sound like a word, look like a word, and even function as if it were a word, but it cannot be a word.” Chatbots seem to create “words,” but these are merely probabilistic expressions of complex algorithms. To perceive and receive mathematically generated simulacra as truly expressive human language follows the sad downward trend in the history of grammar and threatens to destroy reality as we know it. As Friedrich Nietzsche put it in The Twilight of the Idols, “I fear we shall never be rid of God, so long as we still believe in grammar.”

But what do Plato’s...