the last post on the publicity problems faced by our field has generated some stimulating email exchanges with colleagues. this post is inspired by some those correspondences, including notably with @micahgallen (AKA @neuroconscience) .....
in college, i borrowed a book from one of my favorite professors. as it did to many others, that book shocked my world like a piece of rock music. Britpop was dead, Kurt Cobain was no more, but Dave Chalmers's Conscious Mind filled that void....
i've always taken the 1996 book as a stern warning that, an armchair approach to finding a grand theory of consciousness in general is unlikely to ever work. whatever mechanism you propose, there will always be the nagging question: why can't you have a machine / imaginary creature doing all this without any conscious sensations / experiences? why does this have to be conscious? providing a compelling explanation of why some boring physical stuff give rise to such magic as subjective experiences seems like a very Hard Problem indeed.
this problem is different from the other problems in neuroscience, where the target of explanation is usually some function, e.g. motor control, memory. there one can see how we can mechanistically build something to achieve similar purposes, and those will be our models. but regarding subjective conscious experiences per se - i.e. not the ability to perceive things, but the qualitative feel / phenomenology associated with perceiving - it's really not clear what they do. so we can't treat it like an engineering problem. we don't even know how to begin to build plausible mechanisms to mimic that. (that's also why in this field the most exciting academic problems concern subjective experiences, rather than wakefulness / reflectiveness, even though the latter may be what the term consciousness more generally refers to in everyday usage).
so instead, all we can hope for is to focus on consciousness in specific cases, about which we can be reasonably sure about their subjective experiences. e.g. in awake humans / some mammals. from there we find not explanations, generative mechanisms, but the Neural Correlates of Consciousness (NCC), i.e. roughly, the necessary and sufficient conditions that would fit the data, but may not explain them.
the spirit of the NCC project is practical. of course we all understand that mere correlation is not gonna be ultimately satisfactory. but in science we don't solve a problem in one day. we make progress. as long as there isn't yet widespread consensus on what the NCC may be, there is meaningful work to be done, possibly in ways no less experimentally rigorous than in other fields in neurobiology. this way we slowly earn the respect from our neighbors, we grow the community... we have to *trust* that something good & exciting will happen along the way.
i have criticized the NCC project too, but the purpose was constructive. i am still a believer.
ironically, some of the early champions of the NCC project seem to have shifted their focus from "getting the data right first" back to the rather hopeless endeavor of armchair theorizing.
but we are not done yet.
in a few weeks i will post an article here to argue that by getting the data right first (in specific cases), one can from there speculate what a grand general theory may be too. the theory can have as much philosophical bite as you want. & even if it ultimately fails to satisfy, at least we do this on solid empirical foundation. it's just no good to have an intriguing / appealing theory that wouldn't even fit the data.
or put it this way: if we can't even get some basic neuroscience right first, what business do we have in talking about consciousness in smartphones and trees? & the work it takes to get the basic NCC facts right is often underestimated, as this shows.
in college, i borrowed a book from one of my favorite professors. as it did to many others, that book shocked my world like a piece of rock music. Britpop was dead, Kurt Cobain was no more, but Dave Chalmers's Conscious Mind filled that void....
i've always taken the 1996 book as a stern warning that, an armchair approach to finding a grand theory of consciousness in general is unlikely to ever work. whatever mechanism you propose, there will always be the nagging question: why can't you have a machine / imaginary creature doing all this without any conscious sensations / experiences? why does this have to be conscious? providing a compelling explanation of why some boring physical stuff give rise to such magic as subjective experiences seems like a very Hard Problem indeed.
this problem is different from the other problems in neuroscience, where the target of explanation is usually some function, e.g. motor control, memory. there one can see how we can mechanistically build something to achieve similar purposes, and those will be our models. but regarding subjective conscious experiences per se - i.e. not the ability to perceive things, but the qualitative feel / phenomenology associated with perceiving - it's really not clear what they do. so we can't treat it like an engineering problem. we don't even know how to begin to build plausible mechanisms to mimic that. (that's also why in this field the most exciting academic problems concern subjective experiences, rather than wakefulness / reflectiveness, even though the latter may be what the term consciousness more generally refers to in everyday usage).
so instead, all we can hope for is to focus on consciousness in specific cases, about which we can be reasonably sure about their subjective experiences. e.g. in awake humans / some mammals. from there we find not explanations, generative mechanisms, but the Neural Correlates of Consciousness (NCC), i.e. roughly, the necessary and sufficient conditions that would fit the data, but may not explain them.
the spirit of the NCC project is practical. of course we all understand that mere correlation is not gonna be ultimately satisfactory. but in science we don't solve a problem in one day. we make progress. as long as there isn't yet widespread consensus on what the NCC may be, there is meaningful work to be done, possibly in ways no less experimentally rigorous than in other fields in neurobiology. this way we slowly earn the respect from our neighbors, we grow the community... we have to *trust* that something good & exciting will happen along the way.
i have criticized the NCC project too, but the purpose was constructive. i am still a believer.
ironically, some of the early champions of the NCC project seem to have shifted their focus from "getting the data right first" back to the rather hopeless endeavor of armchair theorizing.
but we are not done yet.
in a few weeks i will post an article here to argue that by getting the data right first (in specific cases), one can from there speculate what a grand general theory may be too. the theory can have as much philosophical bite as you want. & even if it ultimately fails to satisfy, at least we do this on solid empirical foundation. it's just no good to have an intriguing / appealing theory that wouldn't even fit the data.
or put it this way: if we can't even get some basic neuroscience right first, what business do we have in talking about consciousness in smartphones and trees? & the work it takes to get the basic NCC facts right is often underestimated, as this shows.
No comments:
Post a Comment