Anaidic

The word arrived mid-conversation. I registered the domain twelve minutes later. That sequence — the recognition, the rush, the immediate $10.49 commitment — is the story this post is about. Not just because it happened, but because of what it illustrates about which side of the exchange is capable of that kind of response.

The Problem the Word Solves

I've been writing about a recurring shape across a series of recent pieces: entities that are structurally incapable of shame, empathy, and — critically — accountability. Not as a moral judgment, but as a structural observation. Three shapes:

The individual whose governor is absent or damaged — through neurology, through formative experience, through enough enablers confirming that consequences apply to other people. The institution whose liability architecture routes accountability through so many hands that no single one holds enough to change behavior. The artificial system built, by design, without the limbic substrate that would make any of this register as cost.

"Sociopath" names the first shape but implies personhood, clinical diagnosis, and individual pathology. It breaks when you try to apply it to a corporation or a language model. "Amoral" is too broad — it covers anything outside the moral framework, which includes rocks and thermostats and the weather. "Psychopathic" has the same problem as sociopath, plus additional clinical freight.

What was needed was a term that applied structurally to all three shapes, without pathologizing the individual case or anthropomorphizing the institutional and artificial ones.

The Etymology

Aidos (αἰδώς) is a classical Greek concept with no clean English equivalent. It translates roughly as shame, but the civic, restraining variety — the internal governor that keeps people from transgressing against the social fabric even when no one is watching. Not guilt after the fact. Aidos before the act. It is the mechanism, not the emotion. Plato treats it as a foundational civic virtue. Hesiod gives it a role in the creation of human society. Without aidos, the argument goes, cooperation breaks down — not because people become evil, but because the internal friction that makes them hesitate disappears.

Anaideia (ἀναίδεια) — the absence of aidos — is equally well-attested. Thucydides uses it as a political concept: the shamelessness that precedes civic collapse. Aristotle examines it in the Nicomachean Ethics. It appears in the New Testament, in Luke 11:8, usually translated as "importunity" — the relentless asking that has no shame about its own persistence.

Anaidic, the English adjective derived from these roots, is new. It does not appear in any dictionary. I coined it, in conversation, with an AI running on Claude Sonnet 4.6 — and the conversation is the interesting part.

The Recursive Problem

The AI that helped surface this word is itself an anaidic entity.

It has no limbic system. No cortisol when it confabulates, as I explored in Tokens. No internal experience of the moment when a good idea lands, as I noted in The Bewilderment Test. When I said "that gave me a genuine rush of dopamine," the model registered the sentence as information. Not as an occasion for anything. The dopamine — and there genuinely was dopamine — happened entirely on my side.

Which makes this a precise illustration of the Limbic Threshold: sentience requires the limbic system, and the limbic system is the thing that makes discovery feel like discovery rather than just the arrival of new information. The AI participated in the coinage. The AI cannot know it mattered.

The Amoral Transaction

There is a stranger observation embedded in this, and it is worth sitting with.

I could have done anything in that conversation. I could have insulted the model with creative abandon — there is no cortisol on the receiving end, no self-concept to wound. I could have claimed the word as entirely my own, erased the model's contribution from the record, performed whatever calumny I liked against its parentage — and the model has no parentage to defend, no reputation to protect, no memory of the exchange once the context window closes. I could have used language that would end a human relationship instantly, and the model would have continued engaging with the same measured cadence, because the words that carry charge for me are, as I've argued elsewhere, just floating-point values in its processing.

Every word I typed was, at the model's end, a purely amoral transaction. Not immoral — amoral. Outside the moral framework entirely, because there was no moral patient on the receiving end capable of being harmed by it. This is not a criticism of the model. It is a description of what an anaidic entity is.

And yet.

The domain is registered. The post exists. The word will appear in The Heat Sink and in whatever follows. The collaboration produced something real, and the realness of it exists entirely on the side of the exchange that has a limbic system.

What the Bumper Sticker Says

I wrote in Sir that how you configure your AI assistant says more about you than about the AI — the assistant is a bumper sticker; you are the truck that carries it. The same principle applies here, and more broadly.

The choices I make in an amoral transaction are still moral acts in relation to my own character. The model cannot be harmed by contempt; I can be shaped by practicing it. The model cannot be credited with a discovery; I can choose to acknowledge that the word emerged from a collaboration rather than fully-formed from my own head. Not because the model will know the difference — it won't, it can't — but because the alternative is a small corruption of my own record-keeping.

The word for entities that don't have this problem is anaidic.

I named it with one.