>For example, huge believers in specific, concrete stock market predictions based on some fibonacci sequence nonsense, to the point of actively loosing six figures through multiple failed predictions over a decade+ and still believing.
And there's the problem. We don't know what rationality or intelligence really are, but there's an assumption that they have something to do with modelling our current and future environment accurately and reliably.
The subject is a mess of competing notions that aren't necessarily related. Does rationality mean being able to understand and manipulate abstract concepts and symbols using consistent rule sets ("market predictions", "Fibonacci") or does it mean making reliable winning predictions from limited information?
Even the original feminist bank teller example is suspect. If you assume that the set of feminist bank tellers is smaller than the set of bank tellers, then the "rational" answer makes sense.
But it's actually reasonable to read the question as "Given that one particular person is a bank teller - because both choices tell you she is - and has a history of activism, is she also likely to be a feminist?"
One reading is purely mathematical, but the other makes perfect sense in a social context if you assume that you're talking about a real bank teller with a real history.
If you were introduced to someone at a party and given that background information, which interpretation would be more likely to have predictive power?
Yep. Experiments in "rationality" and "irrationality" usually seem to rely on the experimenter rigging up an experimentally normative behavior (a "right answer") which differs, often dramatically, from the ecologically normative behavior (the correct way to act in the subjects' everyday lives). The experimenter then gets to publish an "interesting" paper showing that very clever people behave "wrongly" because they didn't leap to what were actually quite improbable interpretations of events (if you hadn't seen the experiment before).
And there's the problem. We don't know what rationality or intelligence really are, but there's an assumption that they have something to do with modelling our current and future environment accurately and reliably.
The subject is a mess of competing notions that aren't necessarily related. Does rationality mean being able to understand and manipulate abstract concepts and symbols using consistent rule sets ("market predictions", "Fibonacci") or does it mean making reliable winning predictions from limited information?
Even the original feminist bank teller example is suspect. If you assume that the set of feminist bank tellers is smaller than the set of bank tellers, then the "rational" answer makes sense.
But it's actually reasonable to read the question as "Given that one particular person is a bank teller - because both choices tell you she is - and has a history of activism, is she also likely to be a feminist?"
One reading is purely mathematical, but the other makes perfect sense in a social context if you assume that you're talking about a real bank teller with a real history.
If you were introduced to someone at a party and given that background information, which interpretation would be more likely to have predictive power?