Is Frank Jackson’s Knowledge Argument about Knowledge?
Frank Jackson’s “Knowledge Argument” against physicalism, also known as the “Mary’s Room Argument,” has been much discussed. Jackson first put it forward in an article of 1982 called, forbiddingly, “Epiphenomenal Qualia.” There later developed a huge literature about it (books, articles, videos), including a number of other essays by Jackson. The bibliography to the Stanford Encyclopedia entry on it has dozens of entries. Recently it has appeared prominently in the work of philosopher Philip Goff (see, for example, his books Consciousness and Fundamental Reality and Galileo’s Error, as well as various articles and videos). I’ll discuss Goff’s use of the argument later after introducing Jackson’s version.
The Knowledge Argument purports to demonstrate that physicalism is false. What Jackson seems to mean is that the world as described by “the physical, chemical and biological sciences … leaves something out” (“Epiphenomenal Qualia,” 127, 129). This part of the argument seems to me obviously true. Physics, chemistry, and biology don’t address culture and individual behavior and psychology, which would seem to be important facts about our world. In a later article Jackson adds “functionalist psychology” to his list of physicalist sciences, so perhaps he would include certain kinds of anthropology and sociology. Anyway, the extent of what is covered by the term physicalism is not totally clear, but this probably doesn’t matter for what I’m going to say.
What’s often forgotten about the Knowledge Argument is that it is about knowledge of other minds. The problem of whether we can know that other people have minds is an old one which I’ve discussed before. How can I know for sure that other people have experiences like I do? The point of Jackson’s original Knowledge Argument is that physicalist science doesn’t tell us something important about other people’s experiences. Physicalist science leaves something out of our knowledge of other people, and the object of this knowledge is then by definition non-physical.
So the Knowledge Argument is about other minds. But is it about knowledge? I’m going to suggest that the Knowledge Argument does not actually demonstrate knowledge. I’m not sure whether this exact argument has been made before; as I said, there is a very large literature on the Knowledge Argument. But in any case I want to be clear that my intention is not to defend physicalism, whatever that is. I simply want to show that the Knowledge Argument is not about knowledge, or at least that it relies on a peculiar concept of knowledge that would need to be specifically defended. The Knowledge Argument, therefore, doesn’t work in its commonly given forms.
Jackson’s Knowledge Argument: Fred
Jackson’s original Knowledge Argument contained two thought experiments: one about Fred and one about Mary. Let’s start with Fred. Fred experiences finer color gradations than anyone else in history. He can distinguish colored objects in a way that no one else can. So where a bunch of ripe tomatoes all look pretty much the same to us, Fred can see a stark difference in color. According to this thought experiment, we know everything physical there is to know about Fred — about his anatomy and biology and physiology and behavior. But according to Jackson “none of this tells us what we really want to know about his colour experience” (“Epiphenomenal Qualia,” 129). I suppose what Jackson means is that Fred’s subjective experience of color is left out of our physical knowledge about him.
Then, let’s say that one day scientists figure out how to change other people’s physiologies to be “like Fred’s in the relevant respects” (129). For ease of explanation I’m going to introduce Jane into the thought experiment. Let’s say that scientists perform an operation on Jane so that her visual apparatus is now like Fred’s. Now Jane can (she and we suppose) experience color the way Fred experienced it. According to Jackson, Jane will have learned something non-physical about Fred: She “will know more about Fred and especially about his colour experiences” (129). Jane will, Jackson says, know more about “the special quality of [Fred’s] experience” (132). This “special quality,” as opposed to the behaviors that demonstrate it or the neurophysiology that explains it, was left out of the physical facts that we had before.
But I question whether in fact Jane will “know more.” Jane, with her altered physiology, will presumably have a new kind of experience. But in what sense does this count as knowledge about the quality of Fred’s experience? Jane will be able to do things now that only Fred could do before. But scientists already knew that if they changed Jane’s visual apparatus to be like Fred’s, she would also be able to do these things. Jane herself may suppose or believe that she is now experiencing color as Fred experienced it. But is this supposition or belief about the sameness of experiences knowledge? If Jane sees the color distinctions that she supposes Fred saw, does she necessarily know more about Fred? Or does she simply have a new experience?
My argument is that Jane is not warranted in claiming knowledge of Fred simply on the basis of her new individual experience. We are warranted in claiming knowledge of Fred if we can show that people with physiologies like Fred can do things that Fred can do but which other people cannot do. A scientist could study Fred and Jane and say, “Before, only Fred could consistently differentiate among these ripe tomatoes. But now Jane can do it too. Therefore, Jane and Fred must see color in a similar way.” The scientist could be said to know something, in the sense that the scientist has evidence for making a comparison between two people. It’s true that this scientist does not have knowledge of the “special quality” of either Jane’s or Fred’s experience; but I’m arguing that Jane and Fred also do not have knowledge of the “special quality” of each other’s experience. All they really know, like the scientist, is that they can both do some special things that regular people can’t do.
And this knowledge is by Jackson’s definition physical knowledge, not non-physical knowledge. As Jackson himself says, “No amount of knowledge about Fred, be it physical or not, amounts to knowledge ‘from the inside’ concerning Fred. We are not Fred” (132). But Jane is also not Fred; she is not looking “from the inside” of Fred; she just happens to have a similarly unique visual apparatus. Jane still cannot directly compare her subjective experience, her qualia, with Fred’s subjective experience. So she cannot claim to know, in a relevant way, any “special quality of his experience” (132). Jane has no evidence for making the comparison other than her own behavior and Fred’s behavior. But this is appealing to physical evidence.
This is the point I’m going to keep coming back to: Why is one individual’s experience of color necessarily knowledge of another person’s experience of color? What is the exact link between the two things? In the modern world, knowledge, as opposed to faith or conviction or supposition or hypothesis or belief, requires evidence. We cannot simply assert that my experience provides knowledge of your experience and leave it at that. This is at best a hypothesis. There needs to be a theory of evidence for getting from one thing to the other. What is this theory?
Jackson’s Knowledge Argument: Mary
The Mary thought experiment approaches this from a slightly different angle. Mary, as you may have heard, is stuck in a black and white room. There she learns everything physical (physical, chemical, biological, behavioral) there is to know about color and other people’s color experience. Then, one day she is let out of the room. Suddenly, as the saying goes, she’s not in Kansas anymore. Jackson says, “It just seems obvious that she will learn something about the world and our visual experience of it” (130). (When he says “our” he obviously means people with typical color vision.) But if this is the case, if Mary really learns something about “us,” then her physical knowledge was incomplete. Hence, physicalism is false.
As Jackson made clear in a follow-up article (“What Mary Didn’t Know,” 1986), the key point is not that Mary learns something about herself — about her own ability to experience color. Jackson says that physicalists and non-physicalists will both agree that Mary will have new experiences after she leaves the room and that she may learn something from them. The key point is that Mary supposedly gains “knowledge about the experiences of others” (292). In the Fred experiment, the key point was that other people (Jane, in my version) gained knowledge about Fred; likewise in the Mary experiment, Mary supposedly gains knowledge about others. Jackson explains: “she will realize how impoverished her conception of the mental life of others has been all along” (292). Her conception of other people, other minds, will change to take account of something non-physical.
But to me, I repeat, the link between Mary’s experience and her “knowledge” of other people’s experiences is not clear. Mary, according to the argument, has learned everything science can tell us about color. Then she experiences color for the first time. She says something like, “Oh, this quality must be what people are talking about when they say ‘the sky is blue.’” But does this count as knowledge of other people, whether physical or not? How do we get from Mary’s own experience of color to knowledge about other people’s experience of color?
Mary will still see other people behaving in the ways she has come to expect from her studies in her black-and-white room. Her own behavior will no doubt be different; for example, she’ll more easily be able to distinguish between red and green apples, between overcast skies and clear skies. Mary may suppose or speculate or hypothesize that now her experience of a blue sky or a red apple is like other people’s experiences of those things. But to call this knowledge of other people’s experiences seems to me an abuse of language. Just because I try your favorite food for the first time, do I now know something about your experience of that food? I may know something of my experience of it; I may say, “You’re right, this flan is really good” or “Actually, I don’t care for it.” I may have some guess about why you like it or how it tastes to you. But it does not follow that I now know something non-physical about your experience. (Locke, for instance, considered our intuitions about others’ experiences to be opinions rather than knowledge; Fechner considered them beliefs rather than knowledge.)
Goff’s Use of the Knowledge Argument
As I said, the Knowledge Argument has been much discussed. And someone who’s made special use of it recently has been the philosopher Philip Goff, probably best known for his support of panpsychism and his arguments against physicalism.
Goff changes Jackson’s Knowledge Argument so that it’s no longer an argument about other minds but is simply about Mary’s experience. I reproduce here the steps of Goff’s version from his book Galileo’s Error (p. 74):
- If materialism is true, then Mary in her black-and-white room has a complete and final theory of color experience.
- If Mary in her black-and-white room has a complete and final theory of color experience, then it shouldn’t be possible for her to learn about some new essential features of color experiences.
- And yet, when Mary leaves her room, she does come to learn about new essential features of color experiences: she learns about what it’s like to have color experiences.
- Therefore, Mary in her black-and-white room can’t have had a complete and final theory of color experiences and materialism is false.
Again, in Jackson’s version it was accepted by both physicalists and non-physicalists that Mary would learn something new. After all, her physical circumstances changed which would cause physical changes to her physiology. Jackson wrote, “It is not … an objection to physicalism that [Mary] learns something on being let out” such as “facts about her experience of red” (“What Mary Didn’t Know,” 292). Why? “Before she was let out, she could not have known facts about her experience of red, for there were no such facts to know” (292).
In other words, in Jackson’s view, it precisely doesn’t matter for the argument whether Mary learns (in Goff’s words) “what it’s like [for her] to have color experiences.” The physical facts that she had before were not about her experience of color, but about the experiences of color for people with color vision living in a colorful world. The fact that Mary learns something about herself when she enters a colorful world is irrelevant.
But in any case, Goff also doesn’t make the connection between experience and theory clear. Why should Mary’s theory of color change just because she has a new experience? We have new experiences all the time without necessarily changing our theories about anything.
Goff thinks that if Mary’s theory really was complete then she should already “know what it’s like to have [say] a yellow experience.” But I don’t think this is actually an argument maintained by physicalists. I may be mistaken, but I don’t think many of them would argue that having enough propositions about a topic is equivalent to experiencing that topic for yourself — that, for example, if I read enough about skydiving this will become equivalent to my actually experiencing skydiving. There are different physical facts involved in reading about skydiving compared to actually skydiving. I think a physicalist would predict that there are different kinds of experiences involved. This is why Jackson doesn’t make Goff’s version of the argument. No matter how much I read about skydiving, this won’t tell me what it will feel like for me to actually skydive; no matter how much Mary reads about color in her black-and-white room, this won’t tell her what it will be like for her to actually experience color. As far as I can tell, every sane person, physicalist or not, agrees on this point.
Jackson’s point, remember, was that through experiencing something we gain facts about other people which cannot be gotten through physicalist propositions. In other words, I may know all the physical propositions about other people’s experiences that can be stated, but when I actually experience something that they’ve experienced, I gain some new knowledge about their experience, in addition to the facts given by all these physicalist propositions. And I’ve argued that we may gain new convictions or hypotheses about others in this way, but we don’t gain any new non-physical “facts” about them. If you want to call Jane’s new beliefs about Fred, or Mary’s new beliefs about normal-color-visioned people, knowledge, then you need to make explicit the theory of evidence being used. You need to explain why we should count this as knowledge. Saying “it just seems obvious,” as Jackson did, isn’t good enough.
What Do Anti-Physicalists Really Want?
What many anti-physicalists seem to want when they appeal to the Knowledge Argument is to literally have other peoples’ experiences. They complain that physicalism cannot give them other people’s experiences; therefore, physicalism is incomplete. But, as the philosopher Grace de Laguna once said, this is to criticize science by appealing “to a sort of knowledge which only a God might enjoy, or perhaps a mortal blessed with a magic power” (see her important article of 1918, “Dualism in Animal Psychology”).
To wit: At the end of a recent debate between Philip Goff and Sean Carroll (Carroll being the token physicalist), someone in the audience brought up Jackson’s thought experiment about Fred. Then he said, “Here’s what I want: What the hell does Fred see?” (1:37:20). Anti-physicalists seem to want a language that perfectly “captures” or “characterizes” (to use Goff’s words) an experience or conscious state. But many philosophers have questioned whether this kind of verbal or theoretical “capturing” is possible or even desirable. (What would you do with an experience once you have perfectly “captured” it?) I would go so far as to argue that this kind of thing is prima facie impossible. Words (or equations, for that matter) can never fully “capture” any phenomenon. That’s not what they’re for.
Goff has said many times that physicalist science has never explained a single conscious state (Donald Hoffman has made similar statements). I can never figure out exactly what he means by “explanation” in this context.
But whatever he may mean, Goff here makes the perfect the enemy of the pretty good. As I said, I’m not strictly a physicalist, in that I think non-scientific/non-physicalist perspectives can give us important insights. But I think our physicalist knowledge of, say, perception is already pretty good in terms of what it has tried to do. The physiological and behavioral sciences have stated many of the conditions under which perception occurs — which, indeed, is the purpose of scientific explanation. That such explanations are not “full” is barely a criticism, since it is impossible for a finite being to exhaustively account for all conditions necessary for any phenomenon. What we want to know about are those conditions which will allow us to effectively control and predict (for example) perceptual experience.
In any case, as far as I can tell, Goff has also not “explained” a single conscious state. Is there not a double standard in anti-physicalists demanding that physicalists do something which non-physicalists also cannot do? For my part, I don’t believe either physicalists or non-physicalists can “fully capture” or “completely characterize” any phenomenon; I don’t require them to and I don’t expect them to (“knowledge which only a God might enjoy,” and all that). The world still awaits panpsychism’s “complete characterization” of a single conscious state. And it will have to keep waiting. The goal is impossible, given the infinite complexity of reality.
Enough of the Knowledge Argument. Are there better arguments out there against physicalism? Let me know. (But please, no zombies ;)