Replies to Jackson

We can group the replies to Jackson's Knowledge Argument by their answers to the following two questions:
  1. When Mary comes out of the room and first sees something red, does she genuinely acquire some new piece of knowledge (knowledge of information)?

  2. If she does acquire some new piece of knowledge, is what she comes to know some new fact about our experiences, of which she was previously ignorant? Or is what she comes to know some old fact, now known in a new way?

Jackson says that Mary does acquire some new piece of knowledge, and that this is knowledge of a new fact.

Lewis denies that Mary acquires a new piece of knowledge. On Lewis's view, what Mary acquires when she leaves the room are just some new abilities. (If we want, we can call this new "know-how." But it is not new knowledge of information.)

Horgan agrees with Jackson that Mary acquires a new piece of knowledge. But he says that Mary is merely coming to know in a new way some fact about our experiences which she already knew. She is not coming to know a new fact.

Horgan

Recall our formulation of the Knowledge Argument:

  1. While still in the room, Mary knows all the physical information there is to know about other people and their experiences.

  2. When Mary leaves the room and sees something red for the first time, she learns something new about other people. She learns what their experiences of red are like. She acquires knowledge about what qualitative features their experiences have.

  3. So when she was still in the room, there was some knowledge that Mary lacked. She did not know everything there is to know about other people's experiences.

  4. So there is some fact about other people's experiences of which Mary was ignorant when she was still in the room. This fact was not captured by the physical information about those other people. It is a non-physical fact.

Horgan challenges the step from (3) to (4). (3) says that Mary acquires a new piece of knowledge. (4) says that Mary acquires knowledge of some new fact. This doesn't follow, Horgan says. It might be that Mary has just come to know an old fact, a fact she already knew, via some new mode of epistemic access.

This is a familiar situation. Lois might know that Superman has a birthmark, but fail to know that Clark Kent also has that birthmark. Yet the fact that Superman has the birthmark is plausibly the same fact as the fact that Clark has the birthmark. One fact, two different ways of knowing it. Similarly, you might know that sugar dissolves easily in water, but fail to know that sugar dissolves easily in H20. But since water just is H20, these too are plausibly the same fact. One fact, two different ways of knowing it.

What explains the difference between Lois's knowledge that Superman has a birthmark and her knowledge that Clark has that birthmark? One explanation is this: Her Superman-beliefs employ a different concept of the man than her Clark-beliefs employ. She has two concepts picking out a single individual. These concepts are different because they are associated with different properties. Under the Superman-concept, she associates certain hero-type properties with Superman. Under the Clark-concept, she associates certain meek-reporter-type properties with him.

Perhaps the same account can be given of Mary's knowledge of what it's like to see red. Mary may have known all along that our experiences of red things have certain neurophysiological features N. After she leaves the room, and sees something red for herself, she then knows that experiences of red have the peculiar "reddish" qualitative features they do. Perhaps these are just two ways of describing a single fact about our experiences. Perhaps having "reddish" qualitative features just is having neurophysiological features N. It's just that, like in the Superman case, Mary has two concepts of a single phenomenon.

Horgan presents this account in the following passages:

Consider Mary at the moment when she finally has her first color-experience--say, the experience of seeing ripe tomatoes. Jackson maintains, and I agree, that Mary obtains new knowledge at this moment, and thus new information: she finds out what it is like to see ripe tomatoes... ("Jackson on Physical Information and Qualia," p. 150)

[But when Mary expresses this new knowledge, by saying "Seeing ripe tomatoes has this property,"] the referent of "this property" may very well be a physical property. This possibility is not ruled out by the fact that Mary learns something new from her experience. (p. 151)

...Even though Superman is Clark Kent, nevertheless we must distinguish between the information that Superman can fly and the information that Clark Kent can fly. Similarly, even if [the qualitative feature of Mary's experience that Mary calls "this property"] is a physical property, nevertheless we must distinguish between (i) the information that the given property, as physicalistically described, is possessed by ripe-tomato experiences, and (ii) the information which Mary expresses by [saying "Seeing ripe tomatoes has this property."] (pp. 151-2)

It may be that Mary is employing two different concepts of a single physical property: a neurophysiological concept of that property, and a concept of it which she first acquires when she leaves the room and has an experience of red for the first time. So when she says "Seeing ripe tomatoes has this property," she is expressing new knowledge, but it is not knowledge about any new properties--properties which she didn't have concepts of before. She is now employing a new concept which picks out an old property.

Against Horgan

Jackson agrees that Mary's concept of "reddish" qualities is different from her concept of the neurophysiological features N. He does not accept that they are two concepts of a single property. More importantly, he argues that the physicalist is not entitled to say that Mary has two concepts of a single property, like Lois had two concepts of Superman.

Why not? Jackson argues that the best explanation of the Superman case turns on the possibility of Lois knowing some of Superman's properties without knowing all of them. She associates certain properties with her Superman-concept, and other properties with her Clark-concept.

But if physicalism is true, then when Mary is in the room, Mary should know about all of the properties our experiences have. If she is ignorant of some of our experiences' properties, then Jackson wins and the physicalist loses. For Mary is supposed to know about every physical property our experiences have. If there are certain properties she is ignorant of, then they can't be physical properties. (Jackson gives this argument in BMJ, Ch. 8, pp. 130-31).

Lewis

Lewis concedes that "knowing what an experience is like" requires actually having the experience; and that no amount of scientific information about how that experience is produced, or what neurophysiological processes take place in you when you're having the experience, will by themselves put you in a position to know what the experience is like.

But Lewis denies that coming to "know what an experience is like" is a matter of learning new information.

Lewis's argument has two parts.

The first part is to argue that this idea, that "knowing what an experience is like" is a matter of learning new information, is "more peculiar than it may at first seem." Lewis offers several reasons for thinking that this idea is peculiar (on pp. 511-514). The most important reason is the following.

Jackson points out that the physical information that Mary has when she's in the room isn't sufficient to give her knowledge of "what it's like" to see red. But he neglects to point out that it's hard to see how any non-physical information could give her this knowledge, either. Even if Mary learned all the non-physical facts about our experiences that she could (by studying parapsychology or consulting with mediums), that would not teach her "what it's like" to see red. Only going out and actually having experiences of red would teach her that.

Lewis writes:

Black-and-white Mary may study all the parapsychology as well as all the psychophysics of color vision, but she still won't know what it's like... If there is such a thing as phenomenal information, it isn't just independent of physical information. It's independent of every sort of information that could be served up in lessons for the inexperienced. For it is supposed to eliminate possibilities that any amount of lessons leave open. ("What Experience Teaches," p. 511)
According to Lewis, the real moral of Jackson's thought-experiment is not that there's some special deficiency that physical information has, when it comes to knowing "what it's like" to see red. The real moral is that no information--neither physical information nor non-physical information--can give Mary knowledge of "what it's like" to see red. Jackson's argument threatens to prove much more than Jackson wants it to prove. It doesn't only prove that "knowing what it's like" to see red is not a matter of knowing physical facts about our experiences of red. It threatens to prove that it's not a matter of knowing any non-physical facts about our experiences, either.

The second part of Lewis's argument is to offer an alternative account of "knowing what an experience is like," which doesn't make it a matter of knowing special information. Lewis offers such an account on pp. 514-18. On his account, "knowing what an experience is like" is a matter of having certain abilities:

  1. the ability to remember experiencing red
  2. the ability to imagine or visualize the color red
  3. the ability to recognize objects as red by sight
  4. the ability to recognize experiences of red (as experiences of that sort, which you've had before)

Acquiring lots of physical information won't give Mary these abilities; just as acquiring lots of physical information about her ears needn't give her the ability to wiggle her ears.

Lewis writes:

The Ability Hypothesis says that knowing what an experience is like just is the possession of these abilities to remember, imagine, and recognize. It isn't the possession of any kind of information, ordinary or peculiar. It isn't knowing that certain possibilities aren't actualized. It isn't knowing-that. It's knowing-how. Therefore it should be no surprise that lessons won't teach you what an experience is like. Lessons impart information; ability is something else. Knowledge-that does not automatically provide know-how. ("What Experience Teaches," p. 516)
In summary: Lewis agrees that Mary does come to "know what it's like" to see red, when she leaves the room. But on his view, this is not acquiring knowledge of any new information, or knowledge about any new fact. It's a matter of acquiring abilities. Since Mary's coming to "know what it's like" is not a matter of her coming to know about new facts, we have no reason to believe there were facts that Mary was ignorant of, when she was still in the room. There were merely abilities she didn't have yet.

So Jackson's thought-experiment does not show physicalism to be false.

Against Lewis

  1. Jackson agrees that Mary acquires the abilities Lewis cites, when she leaves the room and sees red for the first time. But is that all she acquires? Or does she also acquire knowledge of some facts about experiences of red (namely, what qualitative features they have)?

    Remember: the knowledge that Mary lacked which is important for Jackson's argument is knowledge about other people's experiences--knowledge about what it's like for them to see red.

    Jackson tells the following story:

    Suppose [Mary] received a lecture on skepticism about other minds while she was incarcerated. On her release she sees a ripe tomato in normal conditions, and so has a sensation of red. Her first reaction is to say that she now knows more about the kind of experiences others have when looking at ripe tomatoes. She then remembers the lecture [on skepticism] and starts to worry. Does she really know more about what their experiences are like, or is she indulging in a wild generalization from one case? In the end she decides she does know, and that skepticism is mistaken... What was she to-ing and fro-ing about--her abilities? Surely not; her representational abilities were a known constant throughout. What else then was she agonizing about than whether or not she had gained factual knowledge of others? There would be nothing to agonize about if ability was all she acquired on her release. ("What Mary Didn't Know," p. 394)
    In this story, Mary knows that she has the ability to visualize red, to recognize red by sight, and so on. Her abilities aren't in question. What's in question is whether or not she knows certain facts about other people's experiences--or whether she only thinks she knows them.

    Comment: It's not really important for Jackson's argument that Mary ever does come to know what other people's experiences are like. Perhaps the skeptic is right and she can't know that. All that Jackson needs is that, when Mary is in the room, she lacks knowledge of what other people's experiences are like--even though she has knowledge of all the physical information about other people, and their experiences. This is enough to show that there is some information about other people's experiences which isn't captured by the physical information.

  2. Mike Thau also argues that coming to have the abilities Lewis describes will not suffice to explain what Mary learns. For merely having an ability does not entail that you know what it's like to exercise that ability.

    Thau imagines the following sort of case:

    Leon has seen red, like the rest of us, but at the age of 20 he shuts himself up in the black and white room for 30 years. Leon retains the ability to visualize red, to remember what experiences of red are like, and so on. But we can suppose that, while he's in the room, he never exercises these abilities. He hasn't visualized red or had any memories of seeing red for all that time. At the age of 50, Leon comes out of the room again, and looks at a ripe tomato. He says, "I'd forgotten what it's like to see red! Now I know again."

    What this example shows, Thau argues, is that one can forget what it's like to see red (i.e., lose knowledge of what it's like to see red), while still having the abilities Lewis describes. So why should the mere acquisition of those abilities, by itself, give Mary knowledge of what it's like to see red?

    One move Lewis might make in response is to say that knowing what it's like doesn't just require having the abilities; it also requires exercising them. When one stops exercising the abilities, then one no longer knows what it's like.

    But this is surely a desperate response: for it would force Lewis to say that Mary--and the rest of us--lose our knowledge of what it's like to see red, as soon as we stop seeing, or visualizing, or remembering seeing, red. That's a very implausible thing to say.

  3. Much of Lewis's discussion is devoted to arguing that "knowing what it's like" is a strange phenomenon, which can't be explained in terms of possessing any information. This is why he proposes to interpret "knowing what it's like" as having certain abilities, rather than as knowing certain information.

    But Jackson's argument can be run without ever introducing the expression "knowing what it's like." Just run the argument as follows:

    1. While still in the room, Mary knows all the physical information there is to know about other people and their experiences.
    2. When Mary leaves the room and sees something red for the first time, she learns something new about other people. She learns what qualitative features their experiences of red have.
    3. So when she was still in the room, there was some knowledge that Mary lacked. She did not know everything there is to know about other people's experiences.
    4. So there is some fact about other people's experiences of which Mary was ignorant when she was still in the room. This fact was not captured by the physical information about those other people. It is a non-physical fact.
    This captures the spirit of Jackson's original argument while avoiding the confusing and controversial notion of "knowing what it's like." Lewis's account of "knowing what it's like" doesn't show us how to diagnose and resist this formulation of Jackson's argument.

Pryor (and Tye 1986)

Let's go back and look at the question "Do you know what it's like to X?" This question can be asked for a number of different kinds of X. For instance, we can ask "Do you know what it's like to be the CEO of a Fortune 500 company?" or "Do you know what it's like to drive a locomotive fast on a cold stormy night?" (this is an example Lewis discusses). These are good examples to think about, because, unlike Mary's knowledge of "what it's like to see red," they don't obviously have anything to do with qualia or knowledge of qualia.

Lewis suggests that knowing what it's like to drive a locomotive fast on a cold stormy night is just a matter of knowing what experiences you would have, when driving a locomotive in that way. (You could know this, even if you've never yourself driven a locomotive. All you need is some way of referring to or describing the relevant experiences. You might refer to them by their causes, or by what they represent, or by what their neural realizations are, etc. You don't need to have them.)

I don't think that Lewis is right about this. It seems to me that knowing what it's like to drive a locomotive fast or to be CEO of a Fortune 500 company requires more than just having knowledge by some means or other of what experiences you would have when driving the locomotive or when being CEO. Let's concentrate on the CEO case. It seems to me that you'd only count as knowing what it's like to be CEO of a Fortune 500 company if you had direct first-hand knowledge of what it's like to be such a CEO. That is, you have to have knowledge which derives from your being a CEO, rather than evidence about someone else's being CEO.

Maybe that's too strong. Perhaps you're not required to have been CEO, but only to have been CEO or something similar. Suppose you were president of a major university, and you have direct first-hand knowledge of what that is like. If you have it on good authority that being president of that university is similar to being CEO of a Fortune 500 company (perhaps you know someone who has occupied both jobs), then you might count as "knowing what it's like" to be CEO, in that case, even though you have never been CEO. Here your knowledge is partly based on testimony, but an important and central component of it is first-hand and direct. Maybe that's enough to count as "knowing what it's like" to be CEO. (I do not know how important and central one's first-hand knowledge has to be, for one to count as "knowing what it's like" to be CEO. There are limits. Not anything goes. I doubt that any first-hand knowledge I have is sufficient for me to count as "knowing what it's like" to be CEO, no matter how much additional, second-hand evidence I may acquire about being a CEO.)

This inquiry is instructive, because it shows us what "knowing what it's like" amounts to in cases where "knowing what it's like" doesn't involve having any special knowledge about qualia. Presumably, to know what it's like to be CEO of a Fortune 500 company, you don't have to know what qualia a CEO has. You just have to know that the job is stressful in certain ways, and that your schedule is very full, and so on. And on my account, you also have to know these things first-hand, rather than on other people's say-so.

This may help us figure out what's going on with Mary. When we say that Mary doesn't know what our experiences of red are like, presumably we're saying this because she has no first-hand acquaintance with experiences of red--and hence, no first-hand, introspective knowledge of what qualitative features those experiences have.

It does not matter whether Mary has knowledge by some other means of what qualitative features experiences of red have. Even if she does, she won't count as "knowing what it's like" to see red. Just as I don't count as "knowing what it's like" to be CEO of a Fortune 500 company, even if I have excellent second-hand evidence about what it's like to be CEO.

My proposal is this: What Mary lacks, when she's in the room, isn't knowledge of some new fact about our experiences. She may know all the relevant facts about our experiences. What she lacks is knowledge of those facts by a certain means. In order to know what our experiences of red are like, Mary has to have some experiences of her own with the same "reddish" qualitative features that our experiences of red have. And she has to have first-hand introspective knowledge of her experiences that they have those "reddish" features. She also has to have some good reason to believe that our experiences of red have the same "reddish" features. (Perhaps she believes this because of the biological similarity between her and us.)

On my view, when Mary leaves the room, she does acquire new knowledge about what qualitative features others' experiences of red have. This may be a fact she already knew. But unlike before, this fact is now known, in part, on the basis of her own introspective knowledge of certain of her experiences, that they have those qualitative features, too.

This is a version of the "Old Fact Known in a New Way" reply to Jackson, of the sort we saw Horgan offering. As such, it denies the step from (3) to (4) in Jackson's argument.

The objection Jackson made to this reply earlier was that it relied on the possibility of knowing some of a thing's properties without knowing all of them. However, my version of this reply does not rely on that possibility. It only rests on the possibility of knowing facts via different modes of evidence.

We can distinguish pieces of knowledge not just by what's known but also by how it's known: that is, what kind of evidence one has for what's known. When you're taught in grade school that phosphorus burns in water, you may thereby come to know that phosphorus burns in water. When you actually see phosphorus burning in water, though, you come to have a different kind of knowledge that phosphorus burns in water. You've now seen it happen with your own eyes. What you know is the same, but now you have a different kind of justification for it. Your current justification depends centrally on the fact that you have had visual experiences representing a certain substance as burning in water. This was not true before.

My diagnosis of Jackson's thought-experiment turns on the possibility of distinguishing knowledge in this way. There is a difference between visually representing burning phosphorus and being told about burning phosphorus. Similarly, there is a difference between introspectively representing some experiences as having qualitative features Q, and being told that those experiences have Q. This difference does not have anything centrally to do with what properties you associate with the representations, in each case. (So it is not the same sort of difference as exists between Lois' beliefs about Superman and her beliefs about Clark Kent.)

I think Lewis's response to Jackson has part of the truth. What Mary needs, to "know what it's like" to see red, is not merely knowledge of some information. On my account, she has to know that information in a certain way, if she's to count as "knowing what it's like" to see red. (Knowing the information in that way will go hand-in-hand with acquiring the abilities that Lewis describes.)

I think Horgan's response to Jackson also has part of the truth. Although Mary does acquire some new piece of knowledge, this may be knowledge of an old fact, a fact she already knew when she was in the room. What she acquired was knowledge with an important first-hand introspective component.

Unlike Horgan, though, I do not insist that Mary's knowledge involves some new concept of an old property. (Mary may come to have a new concept as a result of seeing red for the first time. But this is not necessary for my diagnosis of Jackson's thought-experiment.) What makes Mary's knowledge new is that it's justified in a new way. Before it was justified solely by testimony and neurophysiological experiment. Now it's justified (centrally) by introspection.

The view I've just laid out shares much with the view defended in Michael Tye's 1986 paper "The Subjective Qualities of Experience." (A few years ago, Tye changed his mind about what to say about the Knowledge Argument. We will see his current views later, when we study Intentionalism.)

In this 1986 paper, Tye tries to define "knowing what it is like to have an experience with qualitative features Q" in a way that is compatible with physicalism about qualitative features Q. Ignoring certain complications, Tye's definition is this:

x knows what it is like to have an experience with qualitative features Q =
  1. x is having, and is introspectively aware of having, an experience with Q; or
  2. x can remember having an experience with Q.
Condition (i) here makes Tye's proposal very similar to my proposal. Condition (i) doesn't just say that x is aware--by some means or other--that one of his experiences has Q. x has to know this fact via introspection. Tye writes:
Knowing what it is like, according to [my account] is grounded on factual knowledge which is obtained in the appropriate manner, namely by introspection." ("The Subjective Qualities of Experience," p. 9)

With this account of "knowing what it is like" in hand, Tye argues that we should diagnose the scientist in Jackson's thought-experiment as acquiring, not knowledge of a new fact, but rather new knowledge--introspective knowledge--of a fact he or she may already have known. Tye writes:

After the operation, Super-Jones does learn things about his color experiences. For he acquires knowledge as to what they...are like. But, according to the physicalist, he does not acquire knowledge of any new facts. Rather by introspection he acquires a new way of knowing certain facts he already knew by other means. ("The Subjective Qualities of Experience," p. 14)
Earlier in the paper he writes:
When [Jones] does undergo these experiences [of seeing red, etc.] and he identifies them as the experiences they are via introspective awareness, then he will know what it is like to have the experience of seeing red, blue, green, etc. So after the operation he will indeed learn something. But in learning something he will not come to know facts of a new sort different in kind from those he knew before, or so the physicalist can insist. Rather he will come to undergo experiences of a sort he has not undergone before, and by introspectively responding to those experiences, he will come to know facts of an old sort, facts just like those he already knew about the experiences of his sighted fellows, but in a new way. Before the operation he only knew facts of that sort by external physical observation of his fellows' bodies and brains; after the operation he knows them by introspective awareness. ("The Subjective Qualities of Experience," p. 9)

This is the same diagnosis of Jackson's thought-experiment that I laid out above.

 


URL:  http://www.courses.fas.harvard.edu
Last Modified: 
Copyright © The President and Fellows of Harvard College