Facing up to the Problem of Consciousness (1995)
By David J. Chalmers
In Heil, pp. 617-39
I. Introduction
A. in this paper, begins by separating what he calls the “hard” problem from the easier problems of consciousness
B. explains why reductive approaches fail to solve the hard problem
C. proposes his own, non-reductive but naturalistic theory of consciousness
II. The easy problems and the hard problem
A. there is not just one problem of consciousness, as “consciousness” is an ambiguous term
B. Begins by distinguishing the “easy” problems from the “hard” problem
C. the easy problems include things like being able to react to environmental stimuli, report on mental states, etc. See p. 618 for full list
1. there is no real question about whether these things could be explained scientifically
2. when he calls these “easy” problems, he means this in a relative sense, as we don’t actually have answers to all of them at present (618)
D. “The really hard problem of consciousness is the problem of experience”
1. by which he means the subjective aspect of experience, or as Nagel puts it, what it is like to be conscious (619)
2. although it is widely agreed that it arises from some physical basis, there is no explanation how
E. Chalmers chooses to reserve the term “consciousness” for subjective experience, referring to the phenomena more easily explained as “awareness”
F. Chalmers finds that many philosophers and scientists who pretend to have a solution to this problem are really solving only one of the “easy” problems
III. Functional explanation (620)
A. What makes the easy problems easy is that they concern the explanation of cognitive functions or abilities
1. to explain a cognitive function, all one needs to do is to specify a mechanism that can perform the function
2. to explain the easy problems of consciousness, we can turn to the methods of the cognitive sciences and the neurosciences, which are well-suited to this task (620, 622)
3. makes an analogy with genetics: to explain the gene, we need to specify the mechanism that stores and transmits information (621)
a. it turns out that DNA does this
b. once we’ve explained how DNA does this, the job is done, and there is no further question to answer about the gene
4. similarly, to explain life, we explain life functions such as reproduction, metabolism, etc.
5. same goes for much of cognitive science, like explaining learning
B. What makes the hard problem of consciousness hard is that it goes beyond questions about the performance of functions
1. it asks why the performance of these functions is accompanied by experience
2. once one gives the functional explanation, there is still something left over to explain, unlike, say, the case with genetics and DNA (q.v.)
3. why don’t all these information-processing functions take place “in the dark,” so to speak?
C. Chalmers is not denying that experience has a function (622)
1. perhaps it will turn out to play some cognitive role
2. but Chalmers thinks that to explain experience, we will have to do more than explain its function
3. and to explain it, we will need some approach other than that of the cognitive sciences and neurosciences
IV. Some case-studies
A. here he gives two examples to back up his claim that the cognitive and neuroscience approaches to the study of consciousness are actually studying only one of the easy problems
B. as an example of a neuroscience approach, cites Francis Crick and Christof [typo] Koch’s theory that the basis of consciousness are 35-75 hertz oscillations in the cerebral cortex
1. to them, these oscillations suggest a mechanism that explains the binding of information
a. binding is the process by which separate pieces of information about a single entity are brought together, such as the color and shape of an object
b. they hold that binding is achieved by the synchronous oscillation of the neuronal cells representing the different properties
2. However, Chalmers argues that this theory does not tell us why these things are experienced (623)
a. even if we give them the claim that these oscillations are the neural correlates of experience, that doesn’t tell us how oscillations give rise to experience
1.) the only reason to think there is some connection is the role that these oscillations play in the binding and storage of information
2.) but their theory merely assumes some connection between binding and experience and doesn’t explain it
b. to be fair, Koch at least does not pretend that this is a solution to the hard problem – see quotation, p. 623
C. as an example of the cognitive psychology approach, Chalmers cites Baars’s global workspace theory of consciousness
1. consciousness is like a communal blackboard that various specialized non-conscious processes send information to for the rest of the system to use
2. According to Chalmers, what this really is is a theory of cognitive accessibility or awareness.
3. Nothing in the theory explains why information in the global workspace should be consciously experienced (624)
D. Chalmers says that similar critiques could be mounted against other theories of consciousness in the cognitive and neurosciences
E. he describes five different strategies that researchers take with regard to the hard problem of consciousness:
1. solve one of the easy problems instead – which is a worthwhile thing to do
2. deny the phenomenon: that is, to argue that once we’ve explained such things as accessibility and reportability of sensations, there’s nothing left to explain
3. claim to be explaining experience in the full sense, but actually passing over the problem quickly, leaving it to look like magic, after going into detail about information processing, but not explaining how these processes give rise to experience (625)
4. explain the structure of experience
a. for instance, the geometrical structure of the visual field, or the relations among the colors
b. although this, too, may be very useful, it takes the consciousness of this experience for granted, and does not tell us why there should be experience in the first place (625)
5. isolate the substrate of experience
a. what Crick and Koch have done in proving the neural correlate
b. at best, this tell us only which processes give rise to consciousness, but not why or how they do it
V. The extra ingredient (626)
A. We have seen that the neurosciences and cognitive sciences fail to explain the hard problem of consciousness
1. they are the wrong sorts of methods
2. an “extra ingredient” is needed
B. there have been no shortage of candidates for this extra ingredient:
1. chaos and nonlinear dynamics
2. nonalgorithmic processing (Penrose)
3. future discoveries in neurophysiology
4. quantum mechanics
C. Chalmers rejects them all
1. mathematical models provide only an account of the functions involved in reasoning
2. from dynamical models, we get only more dynamics
3. neurophysiological accounts explain only brain functions (626)
4. quantum mechanics:
a. people seem to think that since QM is mysterious, and consciousness is mysterious, there must be some connection
b. but when it comes to explaining the mysteries of consciousness, QM is no better than anything else: it is not clear why any of the processes in QM should give rise to experience (627)
c. it’s no help to point to the interpretation of QM according to which conscious experience collapses the wave function, since this assumes the existence of consciousness and does not explain it (627 n. 2)
D. any physical account of consciousness leaves unanswered the question as to why this physical process should give rise to experience
E. Purely physical explanations can account for structures and functions, but not consciousness
1. physical facts can entail only more physical facts
2. although experience may arise from physical things, it is not entailed by them
a. reductive methods work elsewhere because structures and functions are the sorts of things that a physical account can entail (627)
b. my comment: why should we care about entailment? Reductions rarely involve simple entailment relationships. And they explain functions only in a limited sense.
F. some might object that Chalmers’ position is analogous to the old discredited vitalist position in biology (628)
1. according to the vitalists, no physical explanation could explain life
2. but Chalmers sees this case as disanalogous:
a. in the case of the vitalists, they doubted that there could be any physical explanation of the complex functions of life. These doubts disappeared as science explained these functions.
b. but in the case of conscious experience, physical explanations of the underlying functions are not in doubt. The problem is that these physical functions are not sufficient
3. another disanalogy
a. the vital spirit was a posit, intended to explain life functions
b. but conscious experience is not a posit; it’s a fact (628)
G. one might also object that all sorts of puzzling phenomena were eventually given physical explanations. But Chalmers argues
1. that these were all physical phenomena, and their explanations all boiled down to problems in the explanation of structure and function
2. while the problem of consciousness is not a problem of structure and function
VI. Nonreductive explanation
A. At this point, some philosophers, such as Colin McGinn, say that we should give up, and that the problem of consciousness is too hard for our limited minds. We are “cognitively closed” to it.
B. But Chalmers finds this sort of pessimism premature: if reductive explanation fails, try nonreductive explanation
C. In physics, it has sometimes turned out that certain phenomena could not be reduced and so were taken to be fundamental (629)
1. for instance, in the 19th century, physicists decided that electromagnetism could not be explained in terms of mechanical processes so it was taken to be fundamental
2. other things, like space-time and matter, are also taken to be fundamental in physics (629)
D. so Chalmers proposes that we take conscious experience as fundamental in the same way, alongside mass, charge, space-time
1. everything in physics is compatible with the absence of consciousness (q.v.)
2. along with fundamental properties, there will be fundamental laws
3. but Chalmers does not see that these laws will conflict with physical laws in any way: they will simply be in addition to them
E. of course, to take conscious experience as fundamental is not to explain why it exists in the first place. But then nothing in physics tells us why there is matter, either (630)
F. Chalmers admits that his position is a form of dualism, but insists that it is a perfectly naturalistic and not a spiritual or mystical dualism (q.v.)
G. curiously, though, he says that we need to add “bridging principles” to explain how experience arises from physical processes (q.v.).
1. this term is often used to explain reduction in the physical, chemical, and biological sciences
2. I find this curious because if subjective experiences arise from physical processes, they are not fundamental
3. for instance, we do not explain how matter arises from space-time
H. I also find it surprising that he readily admits that biology does not need to postulate any fundamental entities, and simply says that this means that a theory of consciousness is more like physics than biology (630)
1. go back to p. 629, where he appears to argue for his fundamental property dualism on the grounds that everything in physics is compatible with the absence of consciousness.
2. We could also argue that everything in physics is compatible with the absence of tigers. Does that make tigers fundamental?
3. In all seriousness, I doubt very much that we can explain the concept of a species, or indeed many concepts in evolutionary biology, in terms of physics.
4. while we’re on the subject of biology, a worry I have about Chalmers’s paper is that he keeps looking for entailment relationships and asking why various physical processes “should” give rise to consciousness
a. that is, he appears to assume that natural selection must work logically
b. but does it? For instance, why are there two visual pathways in the human brain?
c. so perhaps it is a mistake for look for a reason that the functions associated with what he calls the “easy” problems should be associated with conscious experience.
1.) there’s no necessity for natural selection to have done it this way
2.) it could have taken another path
3.) but it didn’t
VII. Toward a theory of consciousness (631)
A. The first problem is lack of data for constructing such a theory
B. Chalmers suggests we can turn to:
1. our own experience
2. indirect evidence, such as verbal reports from others
3. philosophical analysis
4. thought experiments
5. such scientific values as simplicity, internal coherence, coherence with the rest of science, etc.
C. for Chalmers, a non-reductive theory of consciousness will consist of “psychophysical principles, principles connecting the properties of physical processes to the properties of experience.” (631-32)
1. These principles will concern the way experience arises from the physical (q.v.)
2. they should tell us what sorts of physical systems will have experiences and what sort of experiences they will have, etc. (632)
D. He then proceeds to explain three principles:
1. the first two concern relationships at a high level between processing or “awareness” and “consciousness” (cf. 636)
2. the third one operates at a more basic level
VII.1 The principle of structural coherence
A. this principle says that the structures of consciousness and awareness are isomorphic (633)
1. Chalmers is using “awareness” to refer to the functional processes that make information available for the control of behavior and verbal reporting (632)
2. wherever there is conscious experience, there is awareness, and wherever information is available for control and reporting, there is consciousness (q.v.)
3. this isomorphism extends to the details of awareness and conscious experience
a. same geometry
b. same relations of sameness and difference (632)
c. even color: three-dimensional structure of phenomenal color space corresponds to that of visual awareness (633)
d. same goes for other kinds of experiences besides the visual, including emotions
B. this principle reflects the fact that although facts about cognitive processes do not entail facts about conscious experience, consciousness and cognition are intimately connected (q.v.)
C. limits to the principle:
1. not all properties of experience are structural properties, such as the intrinsic redness of red
2. the very intelligibility of the inverted spectrum argument, where colors are inverted but the structure remains the same, shows that structure alone does not suffice to explain experience (633-34, q.v.)
D. this principle of structural coherence is useful and even assumed by many researchers studying consciousness (634)
1. if we accept the principle, the underlying neural processes that explain awareness will at the same time be the basis for consciousness
2. for example, if we can explain the structure of our awareness of colors through neural processes, we can indirectly explain the structure of phenomenal color space (634)
3. similarly, Crick and Koch’s 40-hertz oscillations and Baar’s “global workspace” can be thought of as mechanisms that explain awareness, and these mechanisms will correlate with consciousness
VII.2 The principle of organizational invariance (635)
A. “this principle says that any two systems with the same fine-grained functional organization will have qualitatively identical experiences”
1. On this principle, if, say, your neural organization were duplicated in silicon, it would have the same experiences as you
2. Chalmers realizes that some philosophers such as Searle would disagree with this
B. Proposes a thought experiment in support of this principle:
1. imagine that the principle is false, and that there could be two functionally isomorphic systems with different experiences. Say for instance one is biological and the other is silicon, and that where one sees red, the other sees blue
2. now imagine gradually transforming one into the other by replacing parts
3. but now imagine that instead of replacing one of the neural circuits, we simply install its silicon counterpart next to it, with a switch in between (635)
a. what would happen when we flip the switch? Would something change from red to blue?
b. but if the functional organization of the system remains the same, the system should not be able to notice (be aware of, a functional notion) any difference!
4. however, the idea of the qualia changing without our noticing anything is absurd (636)
a. whenever our experience changes and we are paying attention we notice the change, otherwise, there’s the possibility of “dancing qualia” (q.v.).
b. although this may be logically possible, it’s not plausible. It’s like the possibility that the world was created 5 minutes ago
c. thus the original assumption behind the thought experiment is false: there couldn’t be two functionally isomorphic systems with different conscious experiences
d. hence, he thinks that philosophical thought experiments involving absent or inverted qualia, although they may be logically possible, are not empirically or nomologically possible (636)
e. n.b. that here he does not take mere logical possibilities seriously. There’s a tension here with things he’s said elsewhere.
VII.3 The double-aspect theory of information
A. the first two were non-basic, high-level principles. In addition, basic principles are needed to explain the high-level principles
B. borrows the notion of information from Shannon’s mathematical theory of communication
1. involves information states embedded in an information space, which has a structure that reflects differences among the elements
2. Although this idea is abstract and mathematical for Shannon, we can see information as physically embodied when there is a space of distinct physical states, which have different physical effects
C. his double-aspect principle starts with noticing the isomorphism between physically embodied information spaces and phenomenal information spaces (636-37 q.v.)
D. leads to the hypothesis that information has both a physical and a phenomenal aspect (637)
1. a basic principle that might explain how experience emerges from the physical (637, q.v.)
2. in support of this principle, briefly mentions a few considerations, such as that physical changes that correspond to changes in experience involve informational changes
3. admits that this principle is highly speculative
C. one question that pops up is whether all information has a phenomenal aspect
1. if it does, then perhaps a thermostat has some minimal experience (638)
2. indeed, if experience were a fundamental property, it would have to be fairly widespread
D. perhaps physical entities have intrinsic or phenomenal properties in addition to the extrinsic or relational properties studied by physics
1. this could help with the puzzle about the causal relevance of experience
2. the worry is that all causation occurs in the physical realm
3. but on Chalmers’ view, experience could be causally relevant in virtue of the fact that it is the intrinsic counterpart of the physical
VIII. Conclusion
A. he thinks the first two principles will be part of any satisfactory theory of consciousness, but he’s less sure of the third one (638)
B. he hopes to have shown that it’s possible to make progress on the hard problem of consciousness