Can we explain human experience using the terms of brain physiology? Well, it depends what you mean by "explain." Our experience has a qualitative character: the feeling of red, the smell of methane, the feel of a cat's scratchy tongue. We can do research and figure out what parts of the body gather this sensory data, how they get to the brain, and which parts of the brain activate when we have these experiences, but according to Chalmers, that still doesn't actually explain why a particular brain state should have the feel to us that it does. We still don't understand how a physical event can cause a mental phenomenon, much less how (or whether) a mental event like an "intention" then causes our physical movements. Physical accounts employ a vocabulary of structure and function, these "feels" (the technical term is "qualia") don't seem amenable to that kind of explanation, therefore there's a conceptual gap between science and the consciousness, and bridging this is what Chalmers famously refers to as "the hard problem of consciousness." For more motivation about why this might be a problem, see our discussion of Thomas Negel's "What Is It Like to Be a Bat" in our episode 21.
"Consciousness and Its Place in Nature" first goes through the different forms of the argument for such a gap, then gives a typology of the kinds of solution that have been proposed, arguing as he goes that the materialist solutions are unsatisfactory in various ways but that other solutions hold more promise. Where appropriate in the discussion, we bring up the two papers that we'll be focusing on for ep. 219 (they were actually planned for this one, but the Chalmers paper was so rich that we just ended up taking all of our time on it): Ned Block's "The Harder Problem of Consciousness" (2002) and David Papineau's "Could There Be a Science of Consciousness?" (2003).
The three arguments he gives against materialism are:
- The explanatory argument. This is the one I just gave: explaining structure and function does not suffice to explain consciousness. Some prominent philosophers like Daniel Dennett deny this.
- The conceivably argument. This is the one that brings up "zombies." It's probably not physically possible that some creature identical in behavior to you could nonetheless not have any of your qualia, but conceptually, it is possible, and maybe such creatures could have existed in a very different sort of universe (it's metaphysically possible).
- The knowledge argument. This is the one that comes out of Frank Jackson's papers about Mary the color scientist who has lived her whole life in a room without color but nonetheless has read every book about what color vision is, how it works in the brain, etc. (Imagine that these books contain even the physiological details that current scientists don't know.) When she emerges from the room and sees color for the first time first-hand, this argument claims that she learns a new fact over and above all the physical facts that were contained in the books. There therefore must be "phenomenal facts" that are distinct from all those physical facts that she knew."
Chalmers then goes through the various positions about consciousness in philosophy of mind literature with an eye to how they respond to these arguments:
Type A Materialism is materialism like Gilbert Ryle's (see our ep. #21) or Pat Churchland's (see our ep. 41) that just insist that consciousness can be fully explained by considerations of function and structure, that zombies as described above aren't legitimately conceivable, and that Mary emerging from the room does not learn any new facts, because there aren't any new facts to learn.
To understand this, it might help to think about Wittgenstein's private language argument: No word in our (or any) language can refer to something that only a single individual can identify. The relation between sign and signified, to be meaningful, needs to stay relatively fixed, and if it's just a word you made up to stand for some private thing in your head, there'd be no way to ensure that its meaning wouldn't drift all over the place. Yet we talk about "pain" and "red" and other seemingly phenomenal experiences all the time. Well, says the argument, then we must be referring to pain-behavior: we see people writhing around in pain and so identify pain, and really, we only secondarily figure out that this sharp sensation now is also pain, largely because our parents, etc., treat us as if we are in pain. So "pain" doesn't refer to a phenomenal quality at all. There really are no words for such qualities, and so they can play no role in our theories: Saying someone did something from an "intention" and some "beliefs" is to make a functional model of their mind (more on this in ep. 220) to try to explain their behavior. We don't learn about intentions and beliefs by introspecting, pointing inwardly, and saying, "Look, an intention! A belief!"
Dan Dennett goes so far as to deny (in "Quining Qualia," discussed briefly in ep. 21) that there are such things as qualia. Certainly there can't be "facts" about these qualia. Chalmers thinks this point of view is very strange and denies the obvious. Whether or not our experience of ourselves is actually informative, i.e.,, whether we can know by contemplating our qualia of red what that qualia actually is in some intrinsic sense, whether or not we can therefore put it into words and say anything really informative about it, much less use the term in an explanation of something else, clearly we do have these experiences, and they are the thing to be explained.
Type B Materialism is the more common scientific type of materialism that doesn't claim that there's anything conceptually wrong with claiming that there could be zombies or phenomenal facts, but claims that as a matter of scientific fact, a phenomenal state is equal to a brain state. It's not just that they're correlated events, but that one simply IS the other, and that this is something we've discovered through science, much like we discovered that water is H2O. So the difference between Type A and Type B is that for Type A, the identity is a priori (a matter of the concepts involved) whereas for Type B, the identity is empirical (something science discovered, not just a matter of concepts).
An argument relevant to this position that we've considered is from Saul Kripke (alluded to in ep. 126), who argues that even though "water" and "H2O" don't mean the same thing, once science discovers that they are identical, we know that they're metaphysically identical. If we consider an alternate possible world where the stuff that serves the function of water (runs in rivers and streams, is for drinking, looks and tastes like water) had a different chemical composition, Kripke argues that actually, that stuff they have still wouldn't be what we call water. He's appealing to our linguistic intuitions to prove this. "Water" is essentially defined by its chemical structure, not its function. However, imagine a comparable case with pain: Let's say we've discovered that pain in us is always accompanied by brain state P. In an alternate world, there are creatures who also phenomenally feel pain, but they have different types of brains, and so they're in brain state Q. Would they (according to our language) actually be feeling pain? Yes, according to Kripke. Unlike water, "pain" is defined according to its phenomenal character, not its underlying material composition.
Type B materialists might like Kripke's analysis of water, but would argue that his analysis of pain is wrong: that pain is exactly like water in this respect. People in the other world with brain state Q might be feeling something functionally similar to pain, and we might have practical reasons if we meet these people to treat their "Q-state pain" as the social equivalent of our "P-state pain," but assuming science really has discovered that P is the thing that pain matches up with in us, then we're going to have to say that, like water, what pain really is is its underlying structure.
The other positions Chalmers covers are interactionism (aka Type D Dualism, from Descartes), where mind and body are separate substances that somehow mysteriously interact; epiphenomenalism (Type E Dualism), where mind is a mere byproduct of brain action, so brain states cause mental states, but it's just an illusion that mental states in turn cause physical changes (the neuroscientific data used to support this that Mark refers to is the Libet Experiment); and "Type F Monism" which posits that there's just a single substance that's beyond what (Newtonian) physics currently understands the world to be. This includes pansychism, the idea that all matter is in some sense conscious or proto-conscious. The panschist answers the hard problem by saying that it's not the burden of the brain to explain why this fundamentally different seeming stuff (qualia) arises out of its particular structure, because qualia are in some sense everywhere. This is a very weird view, but it's one of Greg's specialties, so we let him fill us in on some of the possible details of this view near the end of our discussion. One of the papers he referred to is Pat Lewtas's "What Is It Like to Be a Quark?"
Read Chalmers's paper online, or go ahead and pick up the wonderful collection that David edited, Philosophy of Mind: Classical and Contemporary Readings, which we'll be drawing on for a few of the next episodes.
Image by Genevieve Arnold.