Four Letters

For years I have called hierarchy, significance, fairness, and symmetry the four cardinal ideas, and in all that time I never said what the word cardinal meant. I treated the four as a list I had curated — the ideas that mattered most, chosen from many. Privately I suspected something stronger: that they were not a selection at all but an alphabet, the letters everything else is spelled in. I was never brave enough to publish that.

Bravery turns out to be unnecessary. There is a derivation.

Start with a fact about mathematics that has nothing to do with my framework, so that whatever else happens, this part cannot be argued with.

Rotation is not a primitive. The exponential that turns a bearing through an angle is a sum of terms, and not one of the terms rotates. Each is a pure power — a stretch, a scaling, a lengthening along a fixed direction. Write the series out and you are looking at a list of things that do not turn.

And yet the sum turns.

Rotation is what interference among non-rotating quantities looks like.

I first saw this years ago and it would not leave me alone: the Taylor series of eₓ as a thing in motion, a stack of orthogonal pieces that generates a turn none of its parts contains. It felt like watching something come from nowhere. By my own doctrine it was not coming from anywhere — I was turning into it — but it took the rest of the framework to see what it was actually showing me.

Read the series as a walk. The n-th term is a step: its length is θⁿ/n!, and its direction is iⁿ — the imaginary unit raised to the n-th power.

Now the arithmetic that carries the whole essay. The powers of i cycle: 1, then i, then −1, then −i, then back to 1, forever. Four values. On the plane those are four directions: right, up, left, down.

So every step of every turn — any angle, any bearing, anywhere on the circle — points along one of exactly four directions. The strides shrink factorially, the corners come faster and faster, and the walk closes on its destination. But it never once steps diagonally.

Nothing else is ever used.

So far, mathematics. Now the identification, and I will mark it as such: everything from here depends on reading i as something more than notation.

In my framework the denominator of experience has always had two components: prediction on the real axis — habit, pattern, the subconscious model of what comes next — and ideation on the imaginary axis, orientation toward what is not yet actual. For two years I treated that as a statement about geography: ideation is located perpendicular to prediction.

The stronger reading is that i is an operator. Ideation is not a place. It is a move — the quarter-turn off the prediction axis — and it can be applied repeatedly. Apply it to prediction and you get pure ideation. Apply it again and you get prediction inverted. The powers of i are the states of iterated ideation, and there are four of them because ideation has period four.

• i⁰ = 1 Hierarchy · 0 No ideation. Pure prediction — the condition of distinction itself, what expectation is before any turn is taken.
• i¹ = i Significance · π/2 Ideation once. Orthogonal to prediction — what could be otherwise. The bid for consciousness rather than the forecast.
• i² = −1 Fairness · π Ideation twice. The inversion of prediction — and inverting distinction is precisely what answering an imbalance means.
• i³ = −i Symmetry · 3π/2 Ideation three times. The inversion of significance — closure, the settling of what bidding opened.

This did something I had needed for a long time. The placement of hierarchy at 0 and fairness at π was, until recently, a ruling — a choice I made and defended. It is now a consequence. Hierarchy sits at zero because prediction is what un-ideated expectation is. Fairness sits opposite because ideating twice inverts, and i² = −1.

I did not choose the placement. The operator did.

Put the two halves together and the claim I was never brave enough to make becomes unavoidable.

Every bearing on the circle — every named condition, every idea you will ever be in relationship with — is reachable as a walk whose steps point only along hierarchy, significance, fairness, and symmetry, in that cyclic order, with factorially decaying stride.

Every idea is a word spelled in four letters.

The factorial decay matters. The early letters carry almost all of any word: the first stride and the second dwarf everything after them, which is to say that hierarchy and significance do most of the spelling of every idea, with fairness and symmetry arriving as ever-finer corrections. That is not a moral ranking. It is the shape of the walk.

And notice what the claim retires. The four cardinals are not the four most important ideas. They are not a teaching convenience, not a curated shortlist, not four among many. They are the complete set of directions out of which every other idea is composed — a basis, in the exact sense a mathematician means it.

Nothing was chosen.

A result you reach once might be an artifact of the road you took. This one arrives twice, by operations that have nothing to do with each other.

• By reflectionMirror the circle across the prediction axis. Exactly two bearings are their own reflection: 0 and π. And exactly one reflected pair lands antipodal rather than adjacent: π/2 and 3π/2. Four positions, singled out by symmetry alone.
• By the seriesThe walk that assembles any rotation uses exactly four step directions: the powers of i. Four positions, singled out by construction alone.

Same four points — the fourth roots of unity. I did not arrange the agreement, and when I noticed it I trusted the alphabet claim for the first time. Two derivations that share no machinery converging on the same small set is what being on to something feels like from the inside.

I hold this framework to a rule: claim exactly as much as the argument carries, and say the rest out loud. Here is the honest ledger for this piece.

Proved: the four-direction structure of rotation. That is a theorem about the exponential and it holds whether or not anything else I believe is true.

Claimed: that those four directions are hierarchy, significance, fairness, and symmetry. This rests on reading i as the ideation operator — to ideate is to turn a quarter off prediction, and doing it four times returns you unchanged. That is a claim about ideation, not about arithmetic, and the whole identification stands or falls with it.

Open: less than when I began the essay. The residue I carried in — whether the psychological act of ideation is rightly identified with the geometric quarter-turn — closed while these pieces were in press: to keep a thing whole while moving is to move perpendicular to it, and ideation is exactly the move that reorients everything while spending nothing actual. Wholeness-preservation and orthogonality are one condition. What still remains: that exactly one new dimension is taken, which is parsimony rather than proof, and that the sense of the turn is convention. The full argument, and where the second dimension comes from, is the companion piece published today.

You cross thousands of bearings between waking and sleeping. A handful have names, and the named ones are your ideas. Every one of them — the noble ones, the obsessive ones, the one that will not resolve until you make its mark — is a word in the same alphabet, spelled right, up, left, down, in strides that fade like factorials.

Four letters. No fifth. Nothing was chosen.