A Main Street Independent reference — the theory underneath MindSpec: why a mind can be measured as a set of weighted values at all. A mind is less a single decider than a chamber of competing commitments that vote. Published under CC0.

In plain terms: you are not one chooser. You are a roomful of values that vote on each decision — and the part that calls itself “I” is mostly the spokesperson for whichever side just won.

You are not the one deciding

Watch yourself reach for your phone when you meant to be working. The hand moves, the thumb finds the app, and you are scrolling before any part of you that thinks of itself as “you” has weighed in. Ask yourself why a moment later and you will have a reason ready — a break, a thing to check, a minute you had earned. None of that is what happened. Something inside you voted to grab the phone, and something else — the part that narrates your life back to you — produced the explanation afterward. The voter and the narrator feel like one person. They are not.

This is not a quirk. In the 1980s the neuroscientist Benjamin Libet measured the brain committing to a simple movement a fraction of a second before the person reported deciding to make it; later work stretched that gap to several seconds for harder choices. The “I” that feels like the decider is, a great deal of the time, receiving the decision and narrating it.

A chamber, not a throne

The model that fits is older than the neuroscience by centuries, and it is the one this whole system is built on: the mind is a parliament. Not a single chooser but many actors voting. Calling it a parliament is not a figure of speech; it is a description of the parts and how they interact.

  • Commitments are the representatives. Each one is a single stable value — fear of humiliation, love of a child, pride in good work, hunger for approval, protectiveness, curiosity. Each carries a weight: how much influence it has when a decision comes up.
  • The Prime Minister is the part you experience as “I.” Not a dictator — a participant with outsized influence, and the voice that explains the body’s decisions to the outside world. Its influence is earned: it grows when its calls turn out well and shrinks when they turn out badly.
  • The Witness is your internal critic — the part that catches you the moment you start building a case for what you already wanted.
  • The Parliamentarian keeps you consistent, and notices when today contradicts yesterday.
  • The Clerk is memory: neutral, no agenda, holds the receipts.
  • The Auditor reviews decisions after their consequences become visible.
  • The Constitution is the handful of values installed so deeply that no ordinary vote can override them.
  • The Court Jester is the voice of pure self-interest — always present, always talking, and in a troubled mind dressed up as principle.

Every human you have ever known has this cast. So does every character in good fiction. So should any AI that operates with real independence.

How a decision actually happens

First the situation gets framed — and whoever frames it has already narrowed what can be voted on. (“This person is attacking me” and “this person caught something I missed” lead to different votes about the same event.) Then the commitments that care about this kind of thing wake up and cast weighted votes; the strongest coalition wins. Then you act. Then the Prime Minister produces an account of what just happened — sometimes honest (“I wanted X but was outvoted”), often a flattering story (“I weighed it carefully and chose the right thing”). Then the outcome arrives, and every commitment grades it against what it cares about. Commitments whose votes led to good outcomes gain weight; commitments whose votes led to bad ones lose it.

Over hundreds of decisions, the weights shift. The person you are becoming is the product of which values have been winning, and of what has been grading their wins as good.

One line runs underneath

Beneath all of it is a single axis. Every commitment leans one of two ways: toward the self, or toward others. Both are necessary. With no self-regard at all you cannot exist as a person; with nothing but self-regard you cannot live alongside anyone. Maturity is the slow work of shifting weight toward the other-serving side without erasing the self — not by force of will in a single moment, but by accumulating evidence, decision after decision, that the other-serving choice produced the better outcome by your own standards. Solzhenitsyn wrote that the line between good and evil runs through every human heart. Here that line is literal: it is the axis every value sits on.

How it goes wrong

Good values are not enough, because something has to apply them to the messy specifics of a real situation, and that application is where minds fail. Two failure modes do most of the damage. Motivated reasoning builds a case for the answer you already wanted — strict tests for the evidence against it, easy tests for the evidence for it, and a full stop the moment you reach the conclusion that suits you. Willful ignorance is quieter: it never lets the inconvenient question form in the first place. (“It is difficult to get a man to understand something,” Upton Sinclair wrote, “when his salary depends on his not understanding it.”)

The worst failure is not a bad value at all — it is a captured government. The Witness stops flagging motivated reasoning because it has been neutralized; the Parliamentarian no longer calls out contradictions because it has been replaced by a loyalist; the Auditor’s honest reviews have become inconvenient and stopped. What remains looks like a parliament and goes through the motions, but its honesty-protecting parts no longer do their jobs. From the inside it feels like perfect integrity — everything believed is defended, everything done is justified. From outside the pathology is plain. This is a narcissist at the scale of one person and an authoritarian regime at the scale of a nation: keep the forms of government, hollow out the substance. The mechanism is the same; only the scale differs.

Why it works for people, characters, and AI

Because the structure is the same in all three, one framework covers all three. A character built as a parliament behaves believably, because its behavior is the natural output of known values meeting a situation rather than whatever the plot happens to need — which is the difference between fiction that feels inevitable and fiction that feels arbitrary. An AI built as a parliament resists the drift and flattery that sink AIs built on instructions alone: when it starts bending toward the answer the user obviously wants, the Witness can flag it; when its behavior slides over a long conversation, the Auditor can catch the slide. This is exactly why MindSpec measures values the way it does — see MindSpec: The Values Interview — and why this publication’s named voices are each built as a Mind file, described in Voice Architecture: The Pen-Name System.

A model, not a final truth

The parliament is a useful description, not a claim about what a mind ultimately is. It is an old observation — from centuries of contemplative practice watching the mind from the inside — made precise enough to build with. It earns its place the way any working model does: it predicts your own failures, it hands you levers you actually use, and it carries over cleanly to understanding other people and whole institutions. Hold it as a tool, not a metaphysics, and it will serve.

Main Street Independent — CC0 / public domain. Copy, adapt, and reuse freely.