the event is not the exposure

So I should really be writing down my ideas in some systematic form. But I never know where to start. So I’ll start here, which is to say, at a random place. We’ll see where this goes.

Today’s idea is the distinction between event and exposure. Event is whatever happens “out there in the world”, and exposure is your payoffs in relation to that event. So for example the event might be “tsunami” and your exposure would be “loss of property”.

This seems obvious enough. Why do we need a mantra? Well, it’s because of this other mantra “the map is not the territory”. That one is basically an epistemic rationalist’s slogan. The map, which is to say your model of reality, is not reality itself. Again, as a propositional statement it should be obvious. The problem is in what it emphasizes. The assumed epistemology there is that we have this model of reality, and we have our intention of how we want reality to be, and based on our model of how reality works, we do certain things in order to bring about our vision of reality.

That is wrong. Wrong as a description of how human beings actually behave. Wrong also as a prescription for how human beings should behave. We go about our daily lives mostly depending on heuristics and rules of thumb, not theories. To a mind brainwashed by the dominant enlightenment ideology of our times this can seem like a bizarre assertion. I don’t pretend I can convince the skeptic here. It is something you have to be on the lookout for and something you must train yourself to notice in your daily life. The narratives are certainly there, only they merely adorn behavior and do not govern it.

Taleb would call those who believe theories drive human behavior “fools”. But I know better than to be so insufficiently cynical. Hanson would say theory-talk is simply far mode story-telling. Homo hypocritus is never stupid enough to walk the talk, but hypocritically and sanely chooses to rely on his trustworthy heuristics when it comes to the mundane practice of day to day living.

So since we do not after all rely on maps to get around in the territory, the mantra of “the map is not the territory” is at best irrelevant and quite often actually harmful. Homo hypocritus’s hypocrisy performs best at the level of the individual when he is in charge of his own stories and behavior. He knows precisely when not to take his stories too seriously if he cares at all about the life he actually lives, as opposed to the fairy-tale life his stories describe. But it all goes painfully astray in modern democracies. (Oh shit, here comes another idea I thought was meant to be in another post; you see the problem here; I will try not to get carried away). With democracy, it is veritable orgy of far mode idealization with virtually no near mode accountability. You can tell stories and even when they get disastrously implemented in policy you will not be at all personally punished by the consequences. But like I said in the parentheses a detailed analysis of this phenomenon deserves another blog post.

Theoretically (ha! caveat emptor), it is not model error per se which is pernicious but a function of that model error, viz. your payoffs. Sure, you say, the map is not the territory, we are well aware of the presence of error. But mathematically speaking the model error is but one cost function among many. A person guided by the map-territory mantra will try to minimize the cost function which is the model error, when he should in fact be minimizing the cost function which is true penalty, i.e. that which does violence to his values, of getting the model wrong.

All very technical, but perhaps I tossed that in there to perhaps convince a rationalist to be willing to consider that other mantra, the event-exposure one.

So should we just build models in observation of the true subjective cost (relative to our values) of getting things wrong? As opposed to some objective measure? Son, if you think that is what human beings have been doing ever since anatomically modern homo sapiens, then I must say you are a supreme specimen of the fool. The whole point of paying attention to your payoffs is that it is what you in fact care about. You don’t care about what the model predicts, you care about whatever shit reality is prepared to dish out to you. That is to say, you should take a non-predictive, as opposed to a predictive stance (oh god, is this yet another blog post?).

Give up on predicting reality. Rather, pay attention to insulating yourself from unacceptable harm, while at the same time allowing yourself the option of seizing on any serendipitous opportunities for great payoffs which come your way. This is what Taleb calls being antifragile. Notice that this does not require a model. For example, you don’t try and predict the weather to plan ahead for what you should do for the weekend. You only make sure that the worse possible weather isn’t going to ruin your weekend, and that you have options for the basic contingencies of rain or shine. And if it snows, great, you can bust your sled and have some fun. The event is not the exposure. You should have fun regardless of the weather.

Advertisement