One Reading, Two Unknowns

Contents

1. The feeling has no label
2. A reading is a line, not a point
3. Why the world gets convicted
4. The second line of position
5. The written denominator
6. Where this stops being true
7. Exercises

01 The feeling has no label

The Reality Equation is three letters and a bar, R = A / E, and its second line, S = ln R, is the one you feel — the real part of it, which is the only part a body reads. I have used the same example for years because it refuses to be impressive. You expected eight dollars and were handed five. R = 5/8 = 0.625, and S = ln 0.625 = −0.470. A small, sour, ordinary afternoon.

Now change the story and keep the number. You were handed eight dollars — exactly what you were handed last month and the month before. The world did not move. But somewhere over the summer, without a meeting and without your consent, the denominator had grown to twelve dollars and eighty cents. R = 8/12.8 = 0.625. S = −0.470. The same sour afternoon, to the third decimal, and the world did nothing.

The Reality Equation does not know which of these happened to you. That much is obvious; an equation knows nothing. The claim of this essay is stronger and, I think, unwelcome: neither do you. The reading arrives as one number. Nothing in it is labelled with which term moved. I said that in the essay on language, where it explained one long argument. It deserves to be the whole lesson, because most of what people believe about the causes of their own feelings is a guess made on the far side of this fact.

Let me say once what kind of claim this is. The Reality Equation is a proposal about experience — that what you undergo is a comparison, that the comparison is a ratio, and that the ratio is read on a log scale. Nothing here proves the proposal. But given the proposal, what follows is not a further proposal. It is arithmetic, and arithmetic does not soften for anyone.

02 A reading is a line, not a point

Take the log of the ratio and it becomes a difference: S = ln A − ln E. Write a for the log of the Actual and e for the log of the size of the Expectation, and the whole of felt surprise is one line long:

A single equation in two unknowns does not have a solution. It has a family of solutions, one for every value you care to assign to the term you did not measure. Solve for the world and you get a = e + S: a straight line of slope one in the plane whose axes are forecast and world, sitting S above the diagonal — below it, when the news was bad. Every point on that line is a different pair — a different world met by a different forecast — and every point produces exactly the reading you are holding. Undo the logs and the line becomes a ray from the origin, A = R · E, and the ray is the same object: the set of all the lives in which you would have felt precisely this.

Figure 01 · Computed

Six afternoons that feel identical

Expectation, E	Actual, A	R = A / E	S = ln R	What happened
8	5	0.625	−0.470	the world fell short
12.8	8	0.625	−0.470	the world held; the denominator had grown
16	10	0.625	−0.470	both doubled
1.6	1	0.625	−0.470	a small room, the same proportion
80	50	0.625	−0.470	a large one
4	2.5	0.625	−0.470	the world halved; so had the forecast

Navigators have a name for this. A single bearing to a lighthouse does not tell you where you are; it tells you that you are somewhere on a line — a line of position, as the navigators’ manuals put it, Bowditch’s among them — and a competent navigator will not put a dot on the chart until a second line, from an independent source, crosses the first. One bearing is not a fix. One reading of surprise is not a diagnosis. It is a line of position through the space of everything that could have produced it.

The sign is no help either, which surprises people. A negative reading says the ratio fell below one. It does not say whether the numerator sank or the denominator rose, because those are the same event in a ratio. Disappointment is a line, and it runs through both stories.

Figure 02 · Computed

The same reading, drawn in both planes

• The level set R = 0.625 — every pair that reads −0.470
• E = 8, A = 5 — the eight-dollar afternoon
• E = 12.8, A = 8 — the world held, the denominator grew
• E = 16, A = 10 — both doubled

One narrowing before I go on. This is the magnitude face of the equation, where A and E are amounts and the ratio can fall below one. On the probability face — where the Actual is the thing that happened, so A = 1, and the reading is Shannon’s −ln p — there is only one unknown, and the reading identifies it. The face that has disappointment is the face with two unknowns. That is not a coincidence; it is the same fact from two sides.

03 Why the world gets convicted

When a feeling arrives, you go looking for its cause. You are, whether you would put it this way or not, solving the inverse problem: given the reading, find the pair. And you can only search the terms you can inspect.

Look at the two. The numerator is a record. It happened; it has a time, a place, a shape; it can be named, described, pointed to, argued about with other people who were in the room. It is finished and still, which is exactly what makes it searchable. The denominator is none of those things. It was brought unasked, before the Actual arrived; it was built from repetition and family weather and every prior arrival, most of it before you had language to describe it; and it does not report. Ask a person what they expected of the evening and watch them fail to say. The gate knows. The gate does not file.

So the search for the cause of a feeling is confined to one term — not by carelessness, and not by any bias a person could be talked out of, but by construction. One of the two unknowns has a file and the other does not. When the search comes back with a culprit, the culprit is always the one that could be searched.

The numerator is not guilty. It is searchable. The denominator is not acquitted. It was never looked for.

I wrote in March that the numerator gets the blame because it is visible. That was the observation. This is the mechanism: the blame is the output of an inverse problem solved with half its unknowns off-limits, and it would come out the same way for any Actualizer whose expectation leaves no record, which is every Actualizer that has ever lived until roughly now.

And the habit runs in the other direction just as easily, which is the part worth being fair about. There are people whose first move, on every sour reading, is I expected too much. That is not insight. It is the same line, resolved by a different fixed rule. Choosing a point on the line without a second reading is not a measurement in either direction; it is a temperament. The blamer and the self-blamer are standing on the same line of position, each with a dot they drew from habit, each certain of a fix they never took.

04 The second line of position

A fix needs a second line, and it needs to be independent — two bearings to the same lighthouse are one line drawn twice. The arithmetic of the equation offers exactly three ways to get one, and each has a name you already use.

Hold the Actual and change the Actualizer. Two people receive the same Actual. Their readings are S₁ = ln A − ln E₁ and S₂ = ln A − ln E₂. Subtract:

The world drops out entirely. What is left is the ratio of two denominators, and that is what a disagreement is: two readings of one record, differing by exactly the log of how differently the two of you were prepared to receive it. Take the two afternoons from the top: the person who expected eight reads −0.470; a person beside them who expected twelve-eighty, handed the same five, reads ln(5/12.8) = −0.940. The gap is 0.470, which is ln 1.6, which is the ratio of their expectations and nothing else. Agreement, by the same arithmetic, reads zero — and a zero tells you only that two denominators matched, not what either one was. The essay on agreement argued that the disagreement is the more informative event. Here is the line of arithmetic that makes it so.

It also delivers a corollary I did not see when I wrote about the common denominator. If millions of gates are being filled from the same few models, the denominators converge, and as they converge the difference above goes to zero for every Actual there is. The instrument reads flat. Nobody reading it is wrong; the two of you really did feel the same. But the only instrument that ever measured a denominator was another denominator, and it works only while they differ. Convergence does not just correlate surprise. It blinds the one gauge that could have shown you the term you cannot see.

Hold the denominator and change the Actual. The mirror case: one person, two arrivals close enough together that the denominator has not yet retrained. Then S₁ − S₂ = ln(A₁/A₂), the denominator cancels, and what is left is a pure measurement of the world. That is what a comparison is. Handed five and then ten against a standing eight, the two readings differ by ln 2 = 0.693 — exactly one bit — and the eight never enters. It works only while the denominator holds still, and the noise-floor essay says it never quite does; every arrival retrains it a little, below the line where you would notice. So this line of position is good for minutes and hours and untrustworthy across months.

Watch for a reading that will not move. This is the one that requires no second person and no fast succession, only a record and some honesty about it. Suppose the world genuinely changes — new job, new city, new partner, three different Actuals by any public measure — and the reading comes back the same each time.

Figure 03 · Computed

Three rooms, one reading: the repeat test

Room	Actual, A	Reading, S	Implied E = A / R	What the denominator did
first	5	−0.470	8	sat 1.6× above the world
second	20	−0.470	32	rose with the world; still 1.6× above it
third	50	−0.470	80	rose again; still 1.6× above it

For S to hold still while A moves, E must have moved in lockstep — the denominator retraining onto each new world and overshooting it by the same proportion, room after room. That is a fact about the denominator’s retraining rule. It is not a fact about any of the rooms. The inflation essay described the ratchet from the inside; this is what it looks like from the outside, and it is the one diagnosis a person can make alone. If the reading repeats while the world changes, the constant is you.

I want to be exact about what that sentence hands you, because it is not a lever. It does not let you reach into the denominator and turn it down; nothing does, and the sentence would be worthless if it pretended to. It is a diagnostic — telemetry about which term has been doing the work — and a diagnostic is only ever worth what it changes about the next action.

05 The written denominator

There is one Actualizer for which the inverse problem has been solvable for a long time, and it is worth seeing why, because it shows what the remedy is made of.

A market has a Reality; the equation does not care that the entity is not a person. And a market’s denominator, alone among denominators, leaves a record. Before a company prints its quarter, a few dozen analysts publish what they expect it to print; the average of those artifacts is the consensus estimate, and it is written down, timestamped and public, before the Actual arrives. So when the number lands and the stock lurches, the market can do what no individual can: it can look up a written estimate of its own denominator, made before the fact. That is why did the business fall short, or was the bar too high is an argument that can be had at all — and why it is had, in earnest, every quarter. The conservation essay walked through what companies do with that written bar. The point here is narrower. The argument exists because the second line of position was on file.

A person’s denominator has never left an artifact like that. Until, as the legible-gate essay argued, roughly now. A stack that carries your memory writes down what you asked, what you kept, what you expected the answer to be — and what is written can be read. The calibration essay asked you to forecast before you delegate, to reclaim a little private surprise from the layer that smooths it away. Here is the second, and I think larger, reason for the same practice. A forecast written before the Actual arrives is the only artifact that makes the next reading identifiable. It is the second bearing. Without it every sour afternoon is a line; with it, some of them become points.

You cannot touch the denominator. You can leave a record of it before the world arrives — and a record is the one thing an inverse problem can use.

Notice what this is and is not. It is not a hand on E. Writing your expectation down does not change your expectation; the gate has already brought what it brought, and the impossibility rule stands. It is an action, and actions leave artifacts, and this particular artifact happens to be the missing equation. Everything the framework permits is present in that sentence and nothing it forbids.

Two limits, stated plainly. First, a written forecast is an estimate of the prediction, made by the conscious mind about a term it inherits rather than sets; it is a good second bearing, not the denominator itself. And it does not recover the pull of the ideas that were aboard, which also contribute to the size of E and which, as the delegate essay found, leave almost no trace in the record of what you did. So even a fully written denominator fixes the point on one axis and leaves the other open. Second, none of this touches the imaginary part of surprise. The angle is not felt, so it is not a reading, so there is nothing to invert; the unidentifiability I have described belongs to the real part, and the angle’s invisibility is a different and deeper fact that the imaginary-part essay already owns.

What the diagnosis changes, then, is not the terms. It is the history you make next. A denominator-caused reading met with world-directed action — leave the job, fire the vendor, end the arrangement — leaves artifacts about the world, and the denominator rides along unretrained into the next room, where the reading will repeat; Figure 03 is what that life looks like on paper. A world-caused reading met with reflexive self-blame retrains nothing either, because I expected too much is a sentence and not an artifact. The one action that makes the next reading identifiable, whichever term moved, is the dull one: put the forecast on file before the world arrives, so that the next time the ratio drops, there are two lines on the chart and not one.

06 Where this stops being true

What is load-bearing

The algebra. S = ln A − ln E is one equation in two unknowns; its solution set is the line a = e + S in log coordinates and the ray A = R·E in the ratio plane. On a shared Actual, S₁ − S₂ = ln(E₂/E₁); on a held denominator, S₁ − S₂ = ln(A₁/A₂). A constant S across varying A forces E proportional to A. Every number in all three figures is computed. That a single reading of a ratio cannot be inverted for its factors is not a claim about psychology; it is the definition of a ratio.

What is convention

Calling the world the unknown you solve for rather than the forecast — the line is the same either way. The unit of the log. Reading the numerator as “searchable” and the denominator as “unwritten” is a fact about availability in the Actualizer, not a fact about the equation, and an Actualizer whose expectation has left an estimate on file — a market, a person with a stack — is the case where the convention flips. The Reality Equation itself remains a proposal about experience, not a theorem.

Where the shorthand breaks

“Same Actual” is an assumption, not a given. The eyepiece shows a sample, and two people in one room are looking through two apertures; the disagreement instrument measures a denominator ratio only to the extent the Actuals were in fact shared, and some of what reads as a denominator difference is a sampling difference. The comparison instrument assumes a denominator that holds still, which the noise floor says it never does; it degrades with the interval. And the repeat test needs the record to certify that the Actuals really differed — a claim about history that can be checked, but must be, since the same rule that inflates the denominator is happy to tell you the rooms were all alike.

Where I would push back on myself

The whole essay rests on the readout being one number. If experience carries a second channel — if a reading against a bloated denominator feels hollow, as I myself wrote in March of a success that landed that way, where the same reading against a world that fell short feels sharp — then there are two readings, not one, and the inverse problem is solvable from the inside after all. The predictive-coding literature, in its later form, quietly carries such a channel: prediction errors arrive weighted by a precision the organism also estimates and updates. That would be a real failure of this essay’s central claim, and it is testable: give people the two afternoons from Figure 01 as lived experiences and see whether they can tell them apart better than chance. If they can, the feeling has a label after all, and I would rather know.

07 Exercises

1. Draw the lineTake a reading of S = +0.693 — one bit of good news. Write down four pairs (E, A) that produce it, including one in which the world did nothing unusual. Plot them in the log plane and confirm they are collinear with slope one. Say what the intercept is.
2. The disagreement gaugeTwo colleagues receive the same review. One reads it at S = −0.223, the other at S = −1.204. Compute the ratio of their expectations without knowing what the review said. Then explain in one sentence why you could.
3. The repeat test, honestlyPick three episodes from your own record that you would describe with the same word — underwhelming, say. Write down, from the record and not from memory of the feeling, whether the Actuals differed. If they did and the reading did not, solve for the denominator in each and state its rule.
4. Write the barBefore the next result you are waiting on — a number, a reply, a verdict — write down what you expect it to be, with a date. When it arrives, compute S from your note and from your feeling separately. Report which one you would have trusted without the note, and which term the note convicts.
5. Argue the other sideMake the case, at full strength, that the disappointment of an inflated expectation and the disappointment of a shortfall are phenomenologically distinct — that people can and do tell them apart. Then say exactly what the equation would need to carry, beyond one real number, for that to be true, and what else in the Reality Equation changes once it carries it.