The Prestige of the Sentence
De Tocqueville Predicted AI Slop
Do you love or hate AI writing, sometimes called “slop”?
Let me suggest that our vibe war is not simply a left-right issue, but a class war over what Marx might call the means of “linguistic production.” When I hear people talk about hallucinations or AI bias, I hear them displacing their deeper angst about their own power and status position in a world that endangers their competitive advantage.
To understand our moment, look back to Alexis de Tocqueville’s diagnostic anxiety about the language of democracy and the insurgent neo-traditionalism of the High Modernists.
Tocqueville observed that language is the ultimate barometer of social structure. In aristocratic societies, it functions like a temple: high entry costs, rigid architecture, inherited property. Precision is piety. To speak correctly is to submit to a hierarchy that precedes the speaker, what T.S. Eliot would later call the “Mind of Europe,” a curated, difficult continuity where every syllable has its place.
Democracy acts as a universal solvent. It demands that language be liquid utility rather than guarded estate. As ranks mingle, words become mobile, capacious, and vague. Tocqueville described democratic language as “boxes with false bottoms.”
The current panic among the professional writer class is the discovery that their temple, the ability to synthesize complex ideas into fluid, persuasive prose, has been converted into a public park. The false bottom described by Toqueville has become an explicit feature of democratic language post GPT-3.
In the early 20th century, Eliot and Ezra Pound sensed the democratic slither and launched a rearguard action. Their poetic project was part and parcel of their political project to renew aristocracy. They were the ultimate linguistic aristocrats. Pound’s injunction to “Make It New” was not an invitation to democratic novelty; it was a demand to strip away the lazy abstractions of the crowd and return to the hard image. He wanted language as precise as a ledger. Eliot, meanwhile, built fortresses of allusion—The Waste Land as citadel—to ensure that poetry remained difficult. For the Modernists, the value of a word lay in its scarcity and its resistance. If anyone could say it, it wasn’t worth saying.
The anti-AI movement is the direct descendant of this High Modernist elitism, which is somewhat ironic given the preponderance of AI-skepticism on the academic left. When critics call LLM output “soulless,” they are often using soul as code for labor-intensive. They are mourning a world where good writing reliably signaled cognitive status. The complaint is not aesthetic. It is economic.
If the Modernists were the last defenders of the temple, Ludwig Wittgenstein is the patron saint of the LLM.
In his Philosophical Investigations, Wittgenstein destroyed the idea that words possess essential or noble meanings. Meaning, he argued, is use. If the crowd uses a word a certain way, that is what it means. There is no higher court of appeal.
LLMs are the physical manifestation of Wittgenstein’s language games. They take the slither Pound despised and render it mathematically certain. The LLM proves that you do not need a soul to play the language game.
This is the third wave of democracy. The first leveled political power. The second leveled access to information. The third levels articulation itself.
The pro-LLM and anti-LLM camps map onto a class divide over the nature of language.
The anti-AI position treats language as identity. The pro-AI position treats language as utility. Identity is scarce.
Utility is a democratic vector. Identity is an aristocratic one. LLMs break the monopoly of the articulate class. Utility is common.
There is a middle way, though. To seek utility when utility counts and identity when identity counts. Not to be a precious as Eliot and Pound and yet to still seek to find a unique voice and style and not simply outsource all communication to the statistical average. Democracy and aristocracy both make good points; both have limits. Tocqueville himself understood the dangers of pure democracy and found solace in checks and balances against the tyranny of the majority, or, we might say, the tyranny of the average.
In other words, let’s celebrate democratization without giving up on the ideals of excellence, nobility, and striving around which our ancestors rallied.



Nice! Your take “There is a middle way, though. To seek utility when utility counts and identity when identity counts.” resonates with your excellent piece on Adam I and Adam II (from the two creation stories in Genesis), where Adam I can be replaced by AI, but Adam II requires the addition of the particularist, relational I, or AI+I.
This expands on Wittgenstein’s “If the crowd uses a word a certain way, that is what it means. There is no higher court of appeal.” Such plain meaning is ‘peshat’, the first step in the four-part Jewish interpretation approach of PaRDeS (Peshat, Remez, Drash, Sod) for deep text understanding.
Of note, the Greek word using these consonants appears in the Septuagint’s translation of Genesis 2’s creation story of Adam II, placed in the ‘Garden of Eden’ in verse 2:8. There is no garden in the first creation story. The Greek word used for garden in 2:8 is παράδεισον – meaning garden / paradise, hence our idea that the Garden of Eden represented paradise. The Hebrew word for "paradise," derived from Persian, is Pardes (פַּרְדֵּס), the same as PaRDeS.
This further supports we need AI+I, and that AI will need us to “seek utility when utility counts and identity when identity counts” to enable us to seek the paradise that the creation of Adam II promised.
I think you are right-- there is an anti-AI movement that is economic and not aesthetic.
And the pro-AI group often is /democratic/ in their attitude.
But I think that it would be wrong to think that there are different variations against AI. The economic argument often utilizes an aesthetic argument for its cause. And in my opinion, the economic argument is wrong, but the aesthetic argument stands. I can viscerally feel the pain when I try to read something that I know is AI-generated and presented as natural writing. (As an aside, the aesthetic argument often dissipates when it's acknowledged as AI. This is like calling a song a pop-song. A classical music fan is not offended by an unchanging four chord progression for 3 minutes, because nobody ever presented it as anything but that.)
And similarly, while the pro-AI group often has democratic motivations, I'm pretty confident AI will not make our societies more egalitarian. As opposed to technologies like the library, internet tech and AI as well, are not merely irrelevant for those that do not know how to wield them. They actively destroy them and their capacity to rise in quality.
I don't think this contradicts anything you said. But these are thoughts that was able to articulate thanks to your essay.