The feeling is a forecast
The world tells you that a feeling is a reaction. Something happens out there, the emotion dutifully follows, and your job is to manage it once it has landed. For sixteen years of investment decisions, my own and my committees' and my clients', that was the industry's working model, and the model was a simple one: emotion is the intruder. It bursts in from somewhere primitive, hijacks the analysis, and the professional's whole task is to keep it outside the room.
The neuroscientist Lisa Feldman Barrett spent thirty years taking that model apart, not so much to rehabilitate emotion as to relocate it. The old story says that emotions are hardwired, that fear and anger and joy each have their own circuit, triggered by the world and written on the face the same way in every culture. In How Emotions Are Made she goes looking for the fingerprints, the distinct bodily signature that ought to belong to each emotion, and she cannot find them: no reliable fear-face, no dedicated anger circuit. What she finds instead is stranger, which is that emotions are not reactions at all. They are predictions.
Her first move is to point out that the raw material is not emotion at all. It is affect, a simple and more or less constant signal running along two dimensions, pleasant or unpleasant, agitated or calm, and that is very nearly all the body actually sends up, so that everything else is interpretation laid over the top. The dread you feel at the desk on a red morning and the dread you feel before a difficult conversation are the same signal wearing two different concepts, which means the naming is not decoration added afterwards. The naming is the emotion.
Her second move is to show that the prediction draws on an archive, because the brain forecasts from what it has already lived through, which is exactly why the experienced investor's discomfort is worth more than the novice's, not because the instincts are somehow purer but because the forecast has been trained on more outcomes, including a humbling number of my own mistakes. Gut feel, on this account, is not a mystery at all; it is a prediction engine reporting back from a longer run of data. It can still be wrong, of course, since a model trained on the past always can be. But it is not noise.
Her third move is to make granularity a skill rather than a temperament, because the people who can distinguish finely between their feelings, who can tell irritation from dread from plain depletion, regulate themselves better and decide better than the people who cannot. The person who only ever knows that she feels bad has a single lever to pull, whilst the person who can tell that today's unease is fatigue rather than signal has several. Precision of feeling turns out to be very like precision of language, because it is precision of language.
The objection more or less writes itself. If emotions are constructed, are they not therefore false, mere stories the brain tells itself and safely ignored? Barrett's answer is the opposite of that, because constructed was never the same thing as arbitrary; the forecast is built out of real bodily accounting and real personal history. Ignore it, and you have not become more rational. You have simply become a decider who has switched off the one instrument still reporting on the state of the operator.
The practical shift, in the end, is a small one. You stop asking whether to let emotion into the decision, because it was never standing outside the decision in the first place; it is the medium the decision travels through. The only questions worth asking are whether the forecast has been properly trained, whether you have named it precisely enough to use it, and whether you can still tell the signal from the residue of a bad night's sleep.
The feeling was never the interruption; the feeling is the forecast. The work is making it a good one.