Consciousness and the Meta-Hard Problem (2001)
By Andy Clark
In Mindware, Appendix II, pp. 171-87
I. Introduction
A. By the “meta-hard” problem of consciousness, Clark means the problem whether there really is a “hard” problem of consciousness, as Chalmers says (185)
B. like Chalmers, Clark points out the ambiguities in the notion of consciousness, which has many meanings, including: (171)
1. being awake
2. self-awareness
3. availability for verbal report
4. availability for control of intentional action
5. raw feel or qualia
C. it’s only the last one that threatens to present a problem for our usual notion of scientific explanation and understanding (172)
1. all the rest have to do either with:
a. informational content, or what is represented
b. how it’s poised either to control action or for use by other systems
2. the question is whether these theories of informational content and poise can explain everything there is to explain (172)
II. Cognitive neuroscience research
A. Clark turns to some research in cognitive neuroscience to see just how far scientific investigations of consciousness can go
B. Blindsight
1. due to cortical damage, patients have a blind spot where they claim to see nothing
2. yet they perform well above chance
a. in guessing whether there’s light flashing in this area
b. orienting hand and wrist to grab something in it
3. two different explanations have been proposed:
a. standard view is that primitive, mid-brain processing has been preserved, while phenomenal consciousness depends on more recently evolved parts of the brain
b. a competing view explains it in terms of preservation of visual “hot spots” within blind region
4. either way, the phenomenon shows that visually guided experience can take place without the accompanying experience (172)
5. hence, it is thought that blindsight can lead us to the neurophysiological basis for qualia (173)
C. Binding
1. Crick and Koch’s work on the 40-hertz oscillations
2. linking, for instance, motion and color of a face that’s talking to you
D. Dorsal versus ventral processing
1. old view: the “what” versus the “where” stream
a. the ventral stream was thought to be involved in being able to recognize and identify objects
b. the dorsal stream was thought to be involved in guiding action
2. Miller and Goodale hypothesis:
a. dorsal supports fine motor control
b. ventral supports perception involved in visual awareness
3. one could be damaged without the other being damaged:
a. if the ventral pathway is damaged, one can manipulate things without being aware of what they are doing
b. if the dorsal is damaged, we can say what and where things are but have trouble reaching out and grasping them (173)
4. difference shows up in normal people, also: the dorsal system doesn’t seem affected by optical illusions
5. what’s important for this discussion of consciousness is that it’s the ventral system that’s identified with conscious visual awareness (173-74)
III. Clark now asks what this sort of research can tell us about consciousness (174)
A. again, it depends on what we mean by this term
B. some philosophers would argue that science tells us nothing about the subjective aspects of experience
1. Ned Block
a. distinguishes “access consciousness” from “phenomenal consciousness” (q.v.)
b. doesn’t think that neuroscience research helps to explain the second sort. Only helps with access consciousness
c. on this view, we could build a robot with both a ventral and dorsal visual stream, and it still wouldn’t necessarily achieve phenomenal consciousness. Similarly if we gave it 40 herz oscillations. (174-75)
2. David Chalmers (175)
a. science sheds light only on functions like access
b. it is logically possible for there to be zombies that have these functions
c. so whatever explains these functions does not explain phenomenal awareness
C. But does science really leave something out? Clark discusses two possible ways to doubt this conclusion, representationalism and narrationism
D. representationalism
1. on this view, there is nothing in the mind that cannot be explained in terms of representations
2. for instance, pain is just a way of representing tissue damage
3. two sorts of representationalism: (176)
a.) simple, as described above
b.) “higher order thought theory”: on this theory, to be phenomenally conscious of pain is to have a representation of a representation of tissue damage (q.v.)
4. advantages of this approach:
a. phenomenally conscious states must be about something – even orgasms could be considered this way (? – see below)
b. we have a much better understanding of content-carrying mental states than we do of raw feels (176)
E. narrationism
1. this view derives from Daniel Dennett, according to whom our consciousness is something we’ve constructed using things taken from our language and culture
2. summarizes Dennett’s full story
First move: the intentional stance. We say that something has beliefs if this helps us predict its behavior (178, q.v.)
Second move: multiple drafts. There are multiple, semi-independent processes going on in the brain at once with no central processor. There is no single, ultimate judgment by the brains about any particular input
Third move: the narrative twist.
1.) Consciousness is where one story about the input is fixed. But it is made possible not by some sort of special inborn neural circuitry, but by the cultural imprinting of the “user-illusion” (178-79)
2.) it is our immersion in culture and language that allows us to make up a narrative about who we are, what we’re doing, etc. that fixes the contents of our thoughts (179, q.v.)
3.) meanwhile, there are still these multiple drafts going on
4.) but we report a specific stream of experience picked out by this high-level system that comes from our culture, in which there seems to be a clear fact of the matter concerning the character of our subjective experience
5.) this what leads to our belief in qualia
6.) but qualia are really nothing but this string of judgments made possible by language and culture
F. another, pessimistic possibility, which he finds in McGinn and Pinker
1. there is a possible scientific explanation of consciousness
2. but we’re not equipped in the right way to figure it out
IV. Now it’s time to take stock
A. Block and Chalmers, we recall, distinguish access from phenomenal consciousness, or the easy from the hard problems of consciousness, and say that science cannot explain the latter (179-80)
B. But is there really something left over when science has done its job? Dennett suggests that where Block and others see a difference in kind, there is really only a difference in degree regarding (180)
1. richness of content
2. degree of influence or control
3. for instance, blindsight involves less richness of content and a lesser degree of control than full consciousness
C. Clark sees two things in favor of Dennett’s view:
1. verificationism, that is, the right premises about access, poise, etc., will allow us to predict the behavior of someone in a way that is scientifically indistinguishable from treating her as the seat of phenomenal consciousness
2. so then why posit anything else? explanatory simplicity
D. on the other side, it could be argued that the first person point of view cannot be so easily put aside.
1. If you really doubt the existence of qualia, there’s nothing that can be said to convince you
2. “If you got to ask, you ain’t never gonna know” (180-81)
E. However, Clark argues: (181)
1. nowhere else in the sciences would such a move be allowed (181)
2. Dennett does not ignore the first person point-of-view. Indeed, he explains the illusion of a unified consciousness in terms of culture and language (q.v.)
F. but Clark still finds a tension in Dennett’s position
1. on the one hand, Dennett says that the qualia freaks are inflating differences of degree into differences in kind
2. but on the other hand, Dennett thinks that there is a difference in kind between humans and other animals, in that humans have the sort of culture that makes it possible “to be the sort of thing it is like something to be” (see quotation, p. 181)
3. Clark finds this sharp division between us and other species hard to reconcile with Dennett’s view that even within human beings responses and discriminations can vary in degree
4. Clark would prefer to think that beliefs, etc. are more widespread in the animal kingdom, differing from ours only in “degree of richness of content and poise for control” (q.v.)
G. Zombies (see quotation, p. 181)
1. Chalmers uses the logical possibility of zombies to make the argument that after you’ve told the whole scientific story, there’s still something left out, since the zombies could be identical to a person from the scientific point of view (182)
2. Clark first questions whether zombies really are logically possible or conceivable, given the facts about the world we live in
3. but he also sees two deeper problems:
a. violates verificationist principle insofar as zombies would be undetectable
b. even Chalmers admits that in our actual world, it is not likely that zombies are naturally possible
c. but Clark argues
1.) that if in our world, the connection between what science tells us and subjective experience is this tight, then it is not clear why we couldn’t have the whole story of consciousness from a scientific point of view (q.v.)
2.) and if we can do it for the actual world, why should scientists be concerned about possible worlds? (182-83)
4. Clark also worries about whether these issues can be settled by appealing to our intuitions about what is conceivable (183)
H. Chalmers, as we’ve seen, tries to solve the problem of consciousness by making it a fundamental property, but Clark thinks that this is premature. Also, consciousness seems to be enjoyed only complexly-organized matter and is not evenly spread out in the universe
I. McGinn and Pinker’s pessimism
1. given the sorts of things that human brains evolved to do, there is no reason to think them capable of figuring out consciousness
2. but Clark doesn’t see any obvious limits to what we can do, or any reason to think that the problem of consciousness is beyond those limits (183-84)
J. Representationalism (184)
1. two reasons for doubt:
a. it is not clear that all mental states are about something: unease, depression, orgasm
b. if we allow that there are some representational states that have content but no phenomenological dimension, why should we think that it is representational content and not “something extra” that gives this phenomenological dimension to others? (184)
2. higher-order representationalism looks like a better bet, but since thinking about one’s thoughts seems to call for higher-level abilities, it leads to the following dilemma:
a. either deny that babies and animals actually have phenomenal consciousness, as Dennett does
b. or find some way to make higher-level states less rationalistic, for example, to regard them as perceptions of mental activity rather than thoughts about thoughts
K. The Meta-Hard problem (185)
1. for Clark, this is the question as to whether there really is a hard problem of consciousness.
2. That is, is it true that there is something about phenomenal consciousness that places it outside the realm of science?
V. The psychology of the hard problem
A. one alternative is, as Price suggests, to ask the psychological question as to why it seems that there is a hard problem
1. Price treats it as an illusion based in our psychological make-up.
2. Because of this make-up, we ignore perfectly good explanations that are staring us in the face (185)
3. after all, why don’t we find hard problems all around?
a. as philosophers since Hume have argued, all we ever really get in the sciences is one thing correlated with another. Causes are elusive
b. so if there are always such explanatory gaps, why make a big deal about explaining consciousness?
4. according to Price, the answer is psychological (186)
a. there are certain tricks we play on ourselves that give us a feeling of understanding, but none of these tricks work in the case of explaining consciousness
b. these tricks include:
1.) when we think we “see” the effect in the cause, as in billiard balls imparting their motion to others
2.) seeing our actions as caused by our intentions
3.) seeing the outcome as an acceptable transformation of its cause, as in puppy to dog
4.) seeing the outcome as belonging to a certain kind, as in acids cause burns
5.) perhaps others
5. according to Price, consciousness may present a case like modern physics: it will just take time and familiarity for us to come to feel comfortable with and accept certain explanations (186)
B. Chalmers replies to Price:
1. that there is an explanatory gap in the case of consciousness is precisely his point
2. but we don’t always have such gaps: Chalmers thinks that the connection between statistical mechanics and thermodynamics is necessary
3. it’s because we don’t have such necessary connections in explaining consciousness that the kind of explanatory gap Price points to actually supports Chalmers’s position
C. But Clark is not convinced by Chalmers’s reply (187)
1. if we can give an explanation of consciousness that’s just as good as our explanation of how the remote works the TV, then Chalmers’s “hard problem” is an illusion
2. we would not let the mere possibility of remote control zombies – remote controls in another, possible world that did not change the channel – undermine our faith in the explanation of how they work in our world
3. Clark is not persuaded that the problem of explaining consciousness is any different than this.