Every time I see someone describe their beliefs as their "priors," I have to fight the urge to punch a hole in the wall.
-
-
Replying to @MetaHumean
Is that because Hume frequently assumed first-person and second-person knowledge being based on two different logical functions (to facilitate the flow of his interlocutors), when "you" and "I" are in fact the same form and one logical function?
1 reply 0 retweets 1 like -
Replying to @RattelyrDragon
No, it's because no one ever says "prior" when they mean "low priors," they ALWAYS mean "beliefs." Also, Kahneman and Tversky proved that humans aren't probabilistic reasoners, so priors don't exist. But beliefs do. And you mean "beliefs." So say "beliefs!"
2 replies 0 retweets 4 likes -
Replying to @MetaHumean @RattelyrDragon
there are actually no statisticians in existence
2 replies 0 retweets 4 likes -
Replying to @eigenrobot @RattelyrDragon
Of course people can build and use statistical models. But our brains don't encode subjective probabilities. Probability models exist; subjective priors don't.
3 replies 0 retweets 2 likes -
Replying to @MetaHumean @RattelyrDragon
hrmr im not sure the difference between native wetware encodings and the models we kludge out on top of it is really important brains probably dont directly encode words or numbers greater than 20 or w/e either or am I missing your point?
1 reply 0 retweets 2 likes -
Replying to @eigenrobot @RattelyrDragon
Kind of missing the point. A prior isn't supposed to be a belief about probability. It's supposed to be a mental state with a probabilistic functional profile. Ramsay showed that, given a coherent set of preferences, we can model all human action as maximizing expected utility...
1 reply 0 retweets 2 likes -
This is the basis of the homo economicus model of human cognition. It assumes that all our desires are utilities, and all our beliefs are probabilities. But if this model is accurate, it makes concrete predictions about how humans will reason. K&T tested those predictions...
1 reply 0 retweets 2 likes -
And found them to be wrong. So much for subjective probabilities and expected utility maximization. Some (many) still hold that Ramsay described an ideal of rationality, rather than providing a descriptively accurate model of human reasoning...
1 reply 0 retweets 1 like -
But the ideal model is not very compelling if it assumes just a completely inaccurate model of human cognition. And people who talk about "my priors" aren't speaking about an ideal, they're describing their own mental states. Inaccurately. But Bayesianism is trendy, so
1 reply 0 retweets 1 like
yeah I mean I come out of microeconomics we know all of our models are wrong in the sense that core utility function / preference axioms dont hold in reality, but they still serve as a pretty useful approximation in many cases map is not the territory, maps are still helpful
-
-
Replying to @eigenrobot @RattelyrDragon
Sure, we learned a lot from Ramsay, he was a genius. But behavioral econ is taking over, for good reason. Probabilistic maps are close to reality in some respects, quite far removed in others. Let's get some better maps.
0 replies 0 retweets 1 likeThanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.