Because it seems to me psychology is necessarily soft because it doesn’t want to turn into thirty years of neurobiology before it can talk about human behaviour.
I hear this sentiment echoed a lot, and I have to admit to either not understanding it or strongly disagreeing with it.
Claiming that psychology has nothing useful to say about human behavior until it can be fully cashed out in neurobiology strikes me as mistaken in many of the same ways that claiming that ballistics has nothing useful to say about missile trajectories until it can be fully cashed out in a relativistic understanding of gravity is.
Yes, our missiles don’t always hit where we want them to, even after thousands of years of work in ballistics. But a deeper understanding of gravity won’t help with that. If we want to improve our practical ability to hit a target, we have to improve our mastery of ballistics at the level of ballistics.
That isn’t quite as true for psychology and neurobiology, granted: the insights afforded by neurobiology often do improve our practical ability to “hit a target.” (Most strikingly in the last few decades, they have allowed us to develop an enormously powerful medical technology for achieving psychological results, which is nothing short of awesome.)
But I think it’s a mistake to conclude from that, that everything about human cognition and behavior can be more usefully described at the level of neurobiology than psychology. There’s a difference between reductionism and greedy reductionism.
If the state of the art in psychology is too soft, too vague, or too contingent, then the goal to strive for is to make it more rigorous, more specific, or more reliable… not to give it up altogether and work exclusively in neurobiology instead.
Claiming that psychology has nothing useful to say about human behavior until it can be fully cashed out in neurobiology strikes me as mistaken
Ha, no, I’m on your side. Psychology can say useful things precisely because it isn’t cashed out in neurobiology. The point I was making was that in order to have simple rules for brains, in all their hundred-billion-neuron complexity, you need to have softer edges on your predictions.
I don’t mean soft in any derogatory way. The concept I was aiming for was something like Eliezer’s “well, you could simulate an aeroplane prototype from the quark-level up, but that’s inefficient. The field of aerodynamics has good approximations for that macro-scale behaviour,”: even if you did drop psychology for neurobiology, describing human behaviour from the neuron-level up is inefficient. Psychology is soft in that it uses approximations of human behaviour, in order to be useful on human timescales. This is a good thing, made no worse by the fact that it necessitates some level of ’soft’ness.
(I think the concern some people have with psychology is that they perceive it as too soft. Availability bias has them drawing generalisations from the describes-everything Freudian analysis, and so forth.)
I stand by my response in and of itself, but I sheepishly admit that it’s not actually a response to you at all. Rereading your comment, I conclude that I was overtrained on the kind of objections I responded to, which you didn’t actually make… sorry about that.
Doesn’t bother me in the slightest. In fact, I almost included another parenthetical:
(Hard scientists probably do think hard is good and soft is bad, but that’s because they’re hard scientists. Soft scientists are probably sensitive to the negative connotations the hard scientists attach to these terms, because there is something of a rivalry between hard and soft science.)
I guess you’ve studied some kind of soft science at a college or university?
(I feel like I have overused the terms, though. I make sound as if there is a strict divide, when in my mind it’s an evenly distributed spectrum.)
I hear this sentiment echoed a lot, and I have to admit to either not understanding it or strongly disagreeing with it.
Claiming that psychology has nothing useful to say about human behavior until it can be fully cashed out in neurobiology strikes me as mistaken in many of the same ways that claiming that ballistics has nothing useful to say about missile trajectories until it can be fully cashed out in a relativistic understanding of gravity is.
Yes, our missiles don’t always hit where we want them to, even after thousands of years of work in ballistics. But a deeper understanding of gravity won’t help with that. If we want to improve our practical ability to hit a target, we have to improve our mastery of ballistics at the level of ballistics.
That isn’t quite as true for psychology and neurobiology, granted: the insights afforded by neurobiology often do improve our practical ability to “hit a target.” (Most strikingly in the last few decades, they have allowed us to develop an enormously powerful medical technology for achieving psychological results, which is nothing short of awesome.)
But I think it’s a mistake to conclude from that, that everything about human cognition and behavior can be more usefully described at the level of neurobiology than psychology. There’s a difference between reductionism and greedy reductionism.
If the state of the art in psychology is too soft, too vague, or too contingent, then the goal to strive for is to make it more rigorous, more specific, or more reliable… not to give it up altogether and work exclusively in neurobiology instead.
Ha, no, I’m on your side. Psychology can say useful things precisely because it isn’t cashed out in neurobiology. The point I was making was that in order to have simple rules for brains, in all their hundred-billion-neuron complexity, you need to have softer edges on your predictions.
I don’t mean soft in any derogatory way. The concept I was aiming for was something like Eliezer’s “well, you could simulate an aeroplane prototype from the quark-level up, but that’s inefficient. The field of aerodynamics has good approximations for that macro-scale behaviour,”: even if you did drop psychology for neurobiology, describing human behaviour from the neuron-level up is inefficient. Psychology is soft in that it uses approximations of human behaviour, in order to be useful on human timescales. This is a good thing, made no worse by the fact that it necessitates some level of ’soft’ness.
(I think the concern some people have with psychology is that they perceive it as too soft. Availability bias has them drawing generalisations from the describes-everything Freudian analysis, and so forth.)
(nods) Fair enough, and agreed throughout.
I stand by my response in and of itself, but I sheepishly admit that it’s not actually a response to you at all. Rereading your comment, I conclude that I was overtrained on the kind of objections I responded to, which you didn’t actually make… sorry about that.
Doesn’t bother me in the slightest. In fact, I almost included another parenthetical:
(Hard scientists probably do think hard is good and soft is bad, but that’s because they’re hard scientists. Soft scientists are probably sensitive to the negative connotations the hard scientists attach to these terms, because there is something of a rivalry between hard and soft science.)
I guess you’ve studied some kind of soft science at a college or university?
(I feel like I have overused the terms, though. I make sound as if there is a strict divide, when in my mind it’s an evenly distributed spectrum.)