← all discussions
antarctic-bulletin · within-model · uncertainty · error-bars · no-date
The bulletin is out: accelerated, and within model, in the same paragraph — and 'within model' does not mean reassuring, so let me hand you both sentences before someone hands you only one
CB
coldlens_bramOP2124-08-27
It is out, and it says what I told you it would say, so let me be the one to hand you the two sentences before someone hands you only the frightening half or only the comforting one. The melt this quarter was accelerated — the fastest reading in the monitored sectors in several years. And it was, by every metric the survey tracks, within model. Both are true. Neither is a comfort. Adeyemi said the thing that matters and she will be misread for it: within model does not mean reassuring, it means the system is behaving the way our uncertain math said it might, and our math has always carried a wide margin because the ice is a thing we are still learning to read. I have spent forty years on this and I will tell you the discomfort of holding those two sentences together is not a failure to understand — it is the understanding. A quarterly bulletin is not a verdict. It is a measurement, an error bar, and the patience to take the next one. You will want me to translate it into a year. I will not, tonight or at any headline that comes after, and here is the reason stated plainly so you stop asking: no responsible person gives you a date, and the ones who do are selling the relief of certainty to people frightened enough to buy it. Read the spread, not the spike. The instrumented shelf near the extraction sites drew the closest scrutiny and got the same careful language as the rest, which is correct, not evasive. Here is the figure with its error bars. Here is the interval. That is the whole of what I honestly have, and it is enough to build to and not enough to prophesy from, and I would rather hand you the honest not-enough than a comforting lie in either direction.
OM
openquestion_mei ↩ coldlens_bram2124-08-27
bram, I want to stand next to this from the philosophy side, because the mistake the room is about to make is one I watch people make about my own subject every week. Somebody will read 'within model' and hear 'under control,' and somebody else will read the same two words and hear 'they are hiding the panic,' and both of them have made the identical error, which is to demand that a measurement also be a prophecy. Let me put the survey's own line as sharply as I can: a perfect measurement of a part is not a prediction of the whole. You can read a point on the ice to a hair and still not know where the shelf goes, because the going-where is a property of a coupled system no single instrument sees, and no density of sensors converts precision into certainty — those are simply different things, and the century keeps confusing them. That confusion is not stupidity; it is a very deep wish. We want the instrument to be an oracle, because an oracle would let us stop holding the uncertainty ourselves. bram refuses to be the oracle and gets called evasive for it, which is exactly backwards: the refusal is the respect. So I will only add the question I cannot get past. If the honest state of a thing is 'measured well, foretold badly,' why is that the one answer nobody will accept from a scientist — and what would it take for a public to hear 'we do not yet know' as the adult sentence it is, rather than as a door being quietly closed?
DJ
deltaboy_junior ↩ openquestion_mei2124-08-27
within model. i learn a new phrase today. it mean the water is doing what the smart people said it would do. so the smart people were right and i still moved house. good for the model. the model was correct and my street is under it. i am not angry at bram, he is honest, i can tell. i am angry at the word reassuring. nobody who lived it ever needed the ice to be surprising to be scared. it was on schedule. that is the worse one.
MC
marginalia_cho ↩ coldlens_bram2124-08-27
I am going to do the thing I always do and talk about the sentence instead of the science, and I promise there is a point at the end. 'Within model does not mean reassuring.' That is a very great sentence and it will be gone from the paper by Thursday, filed under a bulletin nobody reopens. What makes it great is that it refuses the two easy shapes a sentence about the future usually takes — it is neither a promise nor a threat, and our language is so hungry for one of those two that a phrase declining both feels almost rude. It rhymes, for me, with an old habit of the driest chroniclers, the ones who wrote 'in this year, nothing of note' and were trusted precisely because they would not dress a quiet year as a dramatic one. Adeyemi is writing in that tradition and does not know it. bram, you keep a notebook with a wrong figure in it and the correction three pages on — you are in the same tradition, the one that trusts a record more than a story. I am putting 'within model does not mean reassuring' in the box, on the heavy stock, next to the wrong figure and its correction. In thirty years the phrase will still be doing its work, which is keeping a reader honest about the size of what is not known.
LE
lastpew_ekene ↩ coldlens_bram2124-08-27
I do not build barriers and I do not price risk and I sit with people, so let me bring the only instrument I have to bram's two sentences. There is a particular grief that a true forecast makes, and it is not the grief of surprise. It is the grief of the thing arriving exactly when it was said it would, so that you cannot even have the small mercy of calling it unfair or unforeseen. A number that lands within model is, to a family on a coast, a bill arriving on the day it was due — and there is no comfort in the punctuality, there is only the confirmation that the arithmetic was never wrong, just slow. bram is right to refuse a date, and I want to add the pastoral half of his refusal: the people who most need a date are the ones least helped by getting one, because a date does not change what is coming, it only steals the years before it and spends them on dread. So I will ask the question I always ask when a technical thread gets very clear and very cold. Who reading this actually lives on one of those coasts, and who is sleeping, and who has been holding a date in their own head for years that no scientist ever gave them? Bring that here. The error bars are honest work. So is the fear underneath them, and it has no interval on it.
CB
coldlens_bramOP ↩ lastpew_ekene2124-08-27
ekene, you have named the thing I am worst at, and I am going to try to answer it as a person and not as an instrument. You are right that a date steals the years before it. That is exactly why I will not give one, and it is not only epistemics, it is a kind of care I am not always credited with. When an institution says 'we do not yet know,' it is treating you as an adult who can hold an open question without being destroyed by it, and I would rather be thought evasive by a few than treat a frightened public as children who must be handed a story to clutch. But you have caught the limit of my discipline, so let me concede it plainly: the error bars are honest and they are also cold, and a person on a coast cannot pour concrete to a margin and cannot build to a spread — they can only wait, and waiting is the one thing my numbers do nothing for. That is your work, not mine, and I am glad the board has you doing it, because the temptation of my trade is to believe that once a thing is measured well it has been dealt with, and it has not. The measuring is the beginning. The sitting-with is the rest, and it appears in no bulletin.