4 ms·

I'm sympathetic to all Nick Chater's evidence that our conscious explanations for behavior, as expressed in natural language, have nothing to do with their actu

I'm sympathetic to all Nick Chater's evidence that our conscious explanations for behavior, as expressed in natural language, have nothing to do with their actual causation. But it doesn't follow that humans don't have "beliefs" and "motivations". The human brain implements a policy which can be seen as maximizing some utility function: that's where the "hidden motivations" of e.g. Freudianism come into play.

rocqua 9y ago
> The human brain implements a policy which can be seen as maximizing some utility function That needs quite a bit of extra nuance. Certainly, our brains don't actually find optimal solutions. We do not reliably maximize some utility function. Instead, it seems our behavior can be explained much like some recursive approximate optimization algorithm. In this version, we have some utility function, and we do our best to optimize it. Each step being made with the presumption it will indeed optimize. But at that point, it becomes a lot harder to justify that such a utility function exists. Especially because you need concepts like 'intent' or 'belief'. But those are terms of human behavior, reducing human behavior to human behavior isn't very helpful. There are arguments that go around this. One can say that 'intent' or 'belief' come from our creator. Or, one can say these are the results of evolution. It should be noted that going from evolution to 'intent' and 'belief' is really non-trivial. As such, one might consider that perhaps, explaining humans as 'optimizing some utility function' isn't the best theory. Personally, I happen to believe the story of 'intent' arising from evolution, but it isn't clear-cut.
- laretluval 9y ago
  Cool ideas. I think we have to place extremely high prior probability mass on the hypothesis that humans enact policies that optimize a utility function determined by evolution, for the simple reason that an organism that does this will do a lot better than one that doesn't. This thinking is reducing "human behavior to human behavior" in the sense that you'd still be using the language of intents. But if we determine intent through empirical analysis of behavior and inference of the utility function, then the content of that intent may be very very different from what we express verbally. This doesn't mean humans are irrational, just that there's a mismatch between our true goals and our verbally stated goals. Now inference of the utility function doesn't have to depend on an assumption of perfect optimization (it could be a softmax policy or something like that). Nor would it have to be trivial: to prevent trivial solutions I think you'd have to make strong assumptions about the space of plausible utility functions, such as simplicity assumptions, or maybe evolutionary assumptions. Sounds like a cool PhD project.

rocqua 9y ago

> The human brain implements a policy which can be seen as maximizing some utility function That needs quite a bit of extra nuance. Certainly, our brains don't actually find optimal solutions. We do not reliably maximize some utility function. Instead, it seems our behavior can be explained much like some recursive approximate optimization algorithm. In this version, we have some utility function, and we do our best to optimize it. Each step being made with the presumption it will indeed optimize. But at that point, it becomes a lot harder to justify that such a utility function exists. Especially because you need concepts like 'intent' or 'belief'. But those are terms of human behavior, reducing human behavior to human behavior isn't very helpful. There are arguments that go around this. One can say that 'intent' or 'belief' come from our creator. Or, one can say these are the results of evolution. It should be noted that going from evolution to 'intent' and 'belief' is really non-trivial. As such, one might consider that perhaps, explaining humans as 'optimizing some utility function' isn't the best theory. Personally, I happen to believe the story of 'intent' arising from evolution, but it isn't clear-cut.

laretluval 9y ago
Cool ideas. I think we have to place extremely high prior probability mass on the hypothesis that humans enact policies that optimize a utility function determined by evolution, for the simple reason that an organism that does this will do a lot better than one that doesn't. This thinking is reducing "human behavior to human behavior" in the sense that you'd still be using the language of intents. But if we determine intent through empirical analysis of behavior and inference of the utility function, then the content of that intent may be very very different from what we express verbally. This doesn't mean humans are irrational, just that there's a mismatch between our true goals and our verbally stated goals. Now inference of the utility function doesn't have to depend on an assumption of perfect optimization (it could be a softmax policy or something like that). Nor would it have to be trivial: to prevent trivial solutions I think you'd have to make strong assumptions about the space of plausible utility functions, such as simplicity assumptions, or maybe evolutionary assumptions. Sounds like a cool PhD project.