The Gauges

History gives us patterns. Accounting gives us constraints. Neither gives us escape.

My great-great-grandfather worked the Carrie Furnace and lost an eye to it. I have spent my adult life wondering why it rusted.

John Stuart Mill was especially alert to the danger of reasoning by analogy — the way the words like, analogous, and parallel tend to arrive at the end of thinking rather than the beginning. We reach for a comparison — this empire is like that one, this decade rhymes with that decade — and the moment the comparison lands, we feel we have understood something. We have not. We have only proposed a resemblance that still has to be earned, difference by difference. The analogy feels like a conclusion. It is really a hypothesis wearing a conclusion’s clothes.

So let me begin by breaking my own rule, and then test what I have done.

My great-great-grandfather worked the Carrie Furnace, and the furnace took one of his eyes. He did not understand the machine he stood inside — not fully, not the metallurgy, not the thermodynamics, not the reasons a stove ran the way it ran on a given morning. No one on that floor did. What he understood was enough to work it: the heat, the sound, the color of the burden, the gauges on the stove, the habits that kept a man’s hands where the machine could not reach them. He read the instruments he had, knew they told him less than everything, and worked anyway. It was enough to make a life. It was not enough to keep his eye.

That is the analogy I want to make for my own work, and I distrust it already. I read the gauges of the American external account — the current account, the investment-income line, the share of Treasuries held abroad, the dollar measured three different ways — and I want to say I am doing what he did: standing inside a machine too large to fully understand, reading the instruments I have, working anyway. It is a seductive comparison. It flatters me by lending a financial analyst the dignity of a man who lost an eye to molten iron. Mill would tell me to stop there and ask what the comparison is hiding before I let it carry any weight.

Here is the uncomfortable thing. The warning cuts in both directions. It is tempting to read Mill and conclude that the honest posture is humility — that the responsible move is to set down the gauges, admit the machine is too large to be understood, and content oneself with watching the ground tremble. But that humility is a claim too, held just as firmly as the forecaster’s confidence. To say the machine cannot be understood is itself a statement about the machine. There is no vantage point off to the side, no balcony, from which one can be sure the gauges are lying. The skeptic and the forecaster are both inside the machine. Neither gets to step outside it.

So what is left, if both the confidence and the humility are postures?

Begin with the furnace analogy, since I owe it a test. Where does it hold? My great-great-grandfather and I are both inside a system we did not design and cannot see whole, working from instruments that report less than the truth. That much is real, and it is the part worth keeping. Where does it break? He could touch his machine. He could walk its length, feel its heat through his boots, and the consequences of misreading it arrived in seconds, on his own body, where no theory could insulate him from them. My machine is an abstraction stitched together from quarterly data, and when I misread it the consequence is a wrong post, corrected at leisure, paid for by no one’s eye. His gauges measured a furnace that existed. Some of mine measure a resemblance I have imposed on the data.

The analogy holds on the structure — a man inside a machine, reading dials — and fails on the stakes and the certainty. Which is exactly Mill’s point: the comparison was a beginning, not a conclusion. It tells me what to examine. It proves nothing on its own.

And there is the deeper trap. When I reach past the gauges for the empires — the Dutch in the 1780s, the Spanish after the silver, the British after Manchester — I am doing the very thing the furnace tempted me into, only at the scale of centuries. I lay those empires beside the present and the present seems suddenly legible. That legibility is the warning sign, not the discovery. The Dutch are not a proof. They are a question I have mistaken for an answer.

What is left, then, is the difference between the kinds of claim a gauge can make. Some of what I read is analogy, and analogy is exposed to every word of Mill’s objection. America is Britain in 1910 is a beam of decorative steel — it looks structural, but you cannot hang the building from it.

But some of what I read is not analogy at all. A current-account deficit must be financed by a capital inflow of the same size. This is not a resemblance to anything. It is an identity, true by construction, true whether or not Pittsburgh is Manchester and whether or not the Dutch are a useful guide. The arithmetic holds where the analogy fails. Part of the work — maybe the only honest part — is knowing which of your beams are load-bearing steel and which are ornament.

That distinction does not rescue me. It narrows the damage. The identities tell me the system is constrained; they do not tell me how it resolves. For the resolution I am back to the empires, back to analogy, back inside Mill’s warning with no way out. The accounting tells me the floor is sloped. It does not tell me which way the furniture will slide.

Which raises the question I cannot stop circling: if I am not smart enough for the thing in front of me — and I have come to believe most of us aren’t, most of the time, in the face of systems this large — is there any value in the trying at all?

Karl Popper gave me the only answer I have found tolerable. He argued that we never prove a theory true; no amount of evidence can verify a general claim, because the next case might break it. What we can do is the opposite. We can state our guess plainly enough that reality is given a clean shot at proving it wrong. Knowledge does not advance by being confirmed. It advances by surviving the attempt to kill it. On that account, being insufficient to the problem is not a reason to fall silent. It is the precise reason the method exists. The point of putting a reading on the dial is not that the reading is correct. It is that a wrong reading, stated openly, is useful in a way that a wise silence never is — it can be corrected, and silence cannot.

So perhaps partial understanding is not a lesser form of knowledge. Perhaps it is the only kind available to anyone inside the machine, where no one ever sees more than the piece he is given to manage. The work is not to escape partiality but to make the it honest — to say what I see, to say what would prove me wrong, and to let the years do the refuting. Even that is a conjecture, held open to its own correction. The uncertainty goes all the way down. And the trying survives it anyway.

And yet I keep reading the gauges — not because they grant escape from the predicament, but because the alternative is to pretend I am not in the machine at all, which is the one posture available to no one. Anyone inside a machine wants to believe he understands it. The wanting is not a flaw to be corrected by sufficient humility; it is the condition of being inside the thing. I find I agree with the humility entirely during the quiet hours of the day, and disagree with it entirely during the working ones. In the morning the gauges look like hubris. By afternoon they look like the only responsible thing to do with a falling needle. They are two honest readings of the same instrument, taken in different light.

My great-great-grandfather understood his furnace well enough to work it and never well enough to be safe in it, and he worked it anyway, because that was the work and there was no version of it that waited for full understanding. That is the part of the analogy I will keep, now that I have tested the rest and watched most of it break. Not that my dials are his — they are not, and the difference is his eye — but that the posture is the same. You read what the instruments give you, you stay honest about how much they withhold, and you do not get to stand outside the machine until you understand it, because no one who is inside it ever does.

The needle is moving. I can tell you that much with confidence, because it is on the dial and the dial is in front of me. What the movement means — whether it is the slow turn of a long cycle or the first tilt of something that does not come back — is the thing the gauge cannot say and I cannot honestly claim to know. So I will keep watching it, and keep writing down what it reads, and keep distrusting the part of me that wants the reading to be the whole story.

That distrust is not the opposite of the work. It is the work.