Knowledge Argument:
Annotated Bibliography
   by Torin Alter 


Back to Home Knowledge Argument Entry

Under construction 


Alter, T. (1995a), "Mary's New Perspective", Australasian Journal of Philosophy 73, pp. 582-4.

Because this is my own paper, I will not evaluate.

I explain and criticize Pereboom's (1994) objection to the KA, which is based on a version of the old-fact/new-guise theory. He argues that the physicalist should reject premise 2, the premise that Mary learns new facts about color experiences when she leaves the room. On his view, the physicalist can argue that the phenomenological representations Mary gains when she leaves the room are inaccurate: they distort the true, physical nature of color experiences.

I argue that Pereboom's objection is based on a false assumption. The assumption is that if Mary's new phenomenological representations distort their true, physical nature (i.e., if they make the experiences appear non-physical) then she fails to learn a new fact about them when she leaves the room. That assumption is false, I argue, because whether Mary's new representations are misleading is not relevant to 2's truth or falsity. If she learns how color experiences appear from the first-person perspective, then she learns a new fact about them. To explain that point, I draw an analogy to learning how a rectangle appears from a certain, distorted perspective. It is a fact that, from a certain perspective, a rectangle will appear rhomboidal. That is so even though that perspective might incline one to deny that the shape in question is a rectangle.

I also make a few points of general interest. First, I note that the question of whether Mary learns new facts when she leaves the room can be posed without introducing the troublesome notion of a physical fact. We can ask whether, given that Mary learns everything that can be learned through watching black-and-white lectures, she learns something new when she leaves the room. Second, I note that the KA does not depend on the assumption (plausible though it is) that how color experiences appear from the first-person perspective is essential to what they are. Third, I note that the KA poses no threat to physicalism unless it can be established that if physicalism is true then pre-release Mary would know every fact about the nature of color experiences.

I would like to add four points. First, in the paper I implied it is irrelevant to the KA's soundness whether color experiences really are as they appear to the experiencing organism. I now think that way of putting my point is potentially misleading, since how something appears is a fact about how it really is. Second, although I maintain that my argument constitutes a challenge to some of the distinctive aspects of Pereboom's objection to the KA, I should have said that more argument would be required to defeat the old-fact/new-guise theory. Third, I cite the wrong publication year of "Epiphenomenal Qualia": it is 1982, not 1980. Fourth, at one place I use 'important' instead of 'essential', which is an error; essential properties (if such there be) might or might not be important and vice versa.

Alter, T. (1995b), Knowing What It Is Like, UCLA Doctoral Dissertation.

Because this is my own work, I will not evaluate.

This is a general investigation of the notion of knowing what it is like, with particular attention paid to the KA and to the related arguments of Nagel's. Most of the material on the KA is expressed in Alter 1998, in a more polished, albeit condensed, form. One thing I regret is the title, which is hackneyed. And I now think the chapter on Nagel is uncharitable in numerous respects. A much better version of that chapter is forthcoming in the Journal of Philosophical Research, under the title "Nagel on Imagination and Physicalism."

Alter, T. (1998), "A Limited Defense of the Knowledge Argument", Philosophical Studies 90, pp. 35-56.

Because this is my own work, I will not evaluate.

I consider and reject several common objections to the KA, including the Lewis-Nemirow Ability Hypothesis, the Acquaintance Analysis, the old-fact/new-guise theory, and a semantic objection. Against the Ability Hypothesis, I argue that someone could understand what seeing red is like while they are seeing red, even if they lack the ability to imagine, remember, and recognize red experiences. Against the Acquaintance Analysis I argue that it is implausible to reduce Mary's new knowledge to acquaintance. I offer a diagnosis of why advocates of the acquaintance analysis like Conee (1994) believe otherwise. I suggest that the mistake derives from confusing two kinds of cases: cases in which gaining acquaintance with x provides one with knowledge of what x is like and cases in which gaining acquaintance with x does not provide such knowledge. Against the old-fact/new-guise theory I argue that the theory is motivated by misleading analogies, such as the case of seeming to learn a new fact about Cassius Clay, when you already know the fact under the 'Ali'-guise. The analogy breaks down, I argue, because the relation between phenomenological features and the relevant facts is more intimate than the relation between a name and pugilistic facts about Ali (a.k.a. Clay). The semantic objection to the KA is the idea that pre-release Mary can exploit her communicative relationship to those outside the room to gain access to the facts that, according to the KA's proponents, she does not know. I criticize that idea by arguing that, although pre-release Mary can refer to color qualia, that is not sufficient for understanding the facts in which those qualia figure.

My defense of the KA is limited, however, because I point out that the KA depends on two assumptions that Jackson fails to establish. The first is the assumption, which Dennett (1991) rejects, that we should trust our intuition that Mary learns something when she leaves the room. The second is the assumption that all physical facts are discursively unlearnable. I point out that both assumptions may be rooted in the ambiguous and vague notion of a physical fact.

I cite the wrong publication year of "Epiphenomenal Qualia": it is 1982, not 1980. And I cite the wrong title of Tye's article (see below).

Alter, T. (unpublished_a), "Ability, Know-How, and the Ability Hypothesis", paper presented at the 2000 Central Division APA

This is an objection to the Lewis-Nemirow Ability Hypothesis. The Ability Hypothesis is the claim that knowing what an experience is like is a kind of know-how and that know-how is ability. Lewis (1988, 1983) and Nemirow (1990) use their Ability Hypothesis to block the KA (Nemirow 1980 discusses a related argument by Nagel). In terms of the Field Guide version, they reject premise 2. They argue that upon release, Mary gains knowledge, but that the knowledge she gains is know-how, not knowledge of facts/information. Their contrast between know-how and knowledge of information (knowledge-that) is based on equating know-how with abilities. In particular, knowing what experiences are like is, they hypothesize, possessing the abilities to remember, recognize, and imagine those experiences.

Based largely on Noam Chomsky's ideas (1994, 1980), I illustrate various ways in which knowing how and ability can come apart. Those considerations, I argue, undermine the claim that the know-how relevant to knowing what it's like can be identified with abilities. I argue that the Lewis-Nemirow strategy for objecting to the KA fails even on the assumption that knowing what an experience is like is know-how.

Alter, T. (unpublished_b), "What A Vulcan Couldn't Know." University of Alabama ms.

Most discussions of the KA focus on Jackson's example of color experiences. I raise the question of whether and to what extent the KA applies to emotions and propositional attitudes. I argue that Jackson's reasoning, to the extent that it is sound, can be extended to all emotions and to certain propositional attitudes, including, e.g., hoping that p and dreading that p, but not to purely cognitive states, like believing that p and contemplating that p, or to the cognitive aspects of emotions. I argue that applying the KA to belief would be self-defeating.

Anchustegui, A. (1999), "The Knowledge Argument and the Multiple Route Doctrine." American Philosophical Association Colloquium, Pacific Division, Spring 1999.

An attack on the old-facts/new-guise theory. Anchustegui argues that that theory falls prey to Kripke's (1972) arguments against the identity theory. [not published, to my knowledge]

Bealer, G. (1994), "Mental Properties", Journal of Philosophy 91, pp. 185-208.

Bealer attempts to give an argument against the identity thesis, which combines the strengths of a traditional Cartesian argument and Kripke's modal argument. The article is rich with ideas, but the discussion of the KA occurs only in passing. Bealer argues that the KA is not sufficient to undermine the identity theory, since the identity theorist can admit that pre-release Mary fails to know certain facts about brain processes, which for them are "facts of scientific definition." That point is made rather briefly.

Bealer compares the KA to a propositional-attitude version of the paradox of analysis. Suppose being a computable function = being a recursive function, and suppose also that I am absolutely certain that computable functions are computable functions. It does not follow that I am absolutely certain that computable functions are recursive functions. Likewise, red sensations might be identical to certain brain processes, even if Mary learns something when she leaves the room, because she does not know the relevant scientific definition. I am not so sure the comparison is apt; it sounds a bit like Churchland's accusation that Jackson commits an intensional fallacy. But it is interesting.

Bigelow, J. and Pargetter, R. (1990), "Acquaintance with Qualia", Theoria 61, pp. 129-147.

Bigelow and Pargetter argue that there is a close connection between Mary's situation and a lack of indexical knowledge, and that idea forms the basis of an objection similar to the one based on the acquaintance knowledge analysis (see Conee, 1994). In fact, they use the term 'acquaintance'. But, unlike Conee, Lewis, and Nemirow, their view is not that, when Mary leaves the room and learns what it is like to see in color, her new knowledge is not propositional in kind. Rather, they argue that her knowledge-gain is comparable to, and no more puzzling than, the absent-minded U.S. historian who learns that today is July 4th, America's Independence Day-propositional knowledge that may seem new but is not knowledge of a fact distinct from those previously known.

This paper is clearly written and well argued, and it anticipates some ideas that others have since expressed (myself included). One problem is that, once we distinguish the position from the old-fact/new-guise theory, it loses much of its intuitive appeal. On the other hand, their position could be seen as a version of the old-fact/new-guise theory; in fact, that may be what the authors had in mind (see section 4).

Bigelow and Pargetter frame much of their discussion in terms of a possible-worlds framework. They write, "In the present context, an acquaintance relation is one which enables a person to hold a belief which is true in all and only the possible worlds containing a specific structure" (139). It is not clear to me that the framework of possible worlds is necessary or especially helpful in the context of the KA. That is a matter of stylistic preference and should not necessarily be taken as a reason not to read the article.

Toward the end of the article, Bigelow and Pargetter provide two arguments against the Ability Hypothesis. The first is similar to an argument given by Conee (1994) and Alter (1998): someone might know what it is like to see red while staring at a red object, and yet lack the associated abilities. (I discovered this when preparing this annotated bibliography.) The second argument is, as far as I can remember, unique to Bigelow and Pargetter. They argue that someone might acquire the associated abilities through blindsight and yet lack knowledge of what it is like to see in color.

Chalmers, D. (1996), The Conscious Mind, New York: Oxford University Press.

Chalmers' discussion of the KA occupies only about ten pages of his 414-page book, but it plays an important role in his argument against physicalism. He provides a spirited and unqualified defense of the KA.

Chalmers takes on almost all of the objections to the KA (he does not discuss Conee's acquaintance analysis), and he provides clear and powerful criticisms. Many of his criticisms are similar to those made elsewhere, but his presentations are forceful. His discussion of the old-fact/new-guise view is nuanced. In particular, he explains why Loar's (1990) version of the view is especially sophisticated and difficult to refute, but then he argues that not even Loar's version ultimately succeeds.

From the perspective of someone interested specifically in the KA, the fact that Chalmers defends the KA as part of his argument against physicalism, and for his own view (naturalistic dualism), has both advantages and disadvantages. Among the advantages is that the KA is placed within a larger context and related to other anti-physicalist arguments. Here Chalmers follows Jackson's (1982) lead in distinguishing the KA from Kripke's modal arguments and other related arguments. But Chalmers goes into more detail than Jackson when discussing those other arguments. Among the disadvantages is that, in some cases, fully appreciating what Chalmers has to say about the KA requires understanding the distinctive features of Chalmers' own general approach to the mind-body problem. For example, his discussion invokes technical notions of primary and secondary intentions, which exploit David Kaplan's dthat operator. It is not clear to me that that apparatus is helpful for the discussion of the KA (though it makes sense in the context of the book). Sometimes he restates his arguments without using the apparatus. But he does not always do that, and at places the discussion is somewhat difficult to follow, despite the fact that his writing is generally clear and elegant.

Chalmers, D. (1997), "The Puzzle of Conscious Experience", Scientific American Special Issue, Mysteries of the Mind, 30-37.

Although this adds nothing to the discussion of the KA in Chalmers 1996, it contains an excellent photograph depicting the Mary case. It's awesome

Churchland, P. (1985), "Reduction, Qualia, and the Direct Introspection of Brain States", Journal of Philosophy 82, pp. 8-28.

This is the article that inspired Jackson's (1986). In it, Churchland defends physicalism against various attacks. The KA is not Churchland's only target, but it is one of the main ones. The article is rich with argument and is well worth studying, although it contains some errors.

The errors are noted in Jackson (1986). For example, Churchland implies that the KA depends on assumptions about what humans can and cannot imagine, given the limitations of our experiences. As Jackson notes, the KA depends on the assumption that pre-release Mary does not imagine color experience. It does not depend on the assumption that she cannot imagine color experiences. See Jackson (1986) for other mistakes of interpretation.

One helpful aspect of Churchland's article, along with his other writings on the KA (e.g., Churchland 1995), is that he refutes objections to physicalism that are related to the KA (e.g., ones that commit the so-called intensional fallacy). That forces critics of physicalism to distinguish their objections from those that Churchland refutes and show why his arguments do not refute their criticisms. Another helpful aspect of the article is that it contains a fairly clear expression of the old-fact/new-guise theory. Churchland also suggests that the KA depends on an equivocation between two senses of 'know', one indicating propositional knowledge and the other indicating know-how. That suggests the Ability Hypothesis of Lewis (1983, 1988) and Nemirow (1980, 1990), which is different from the old-fact/new-guise theory. Churchland acknowledges this, but it remains unclear to me why he puts the point in terms of an ambiguity in 'knowledge'. See Lewis 1988, 516-17.

Churchland's discussion toward the end of the article also suggests the objection that Jackson (1998a) has recently accepted, that we should not trust our intuition that Mary gains knowledge at all when she leaves the room. Dennett (1991) credits Churchland with being the only philosopher (at that time), aside from himself, who questioned that premise.

Conee, E. (1994), "Phenomenal Knowledge", Australasian Journal of Philosophy 72, pp. 136-150.

This is the clearest articulation and defense of the acquaintance analysis. Other critics of the KA use the term 'acquaintance knowledge' to describe the kind of knowledge Mary gains upon leaving the room (see Bigelow and Pargetter). But usually what those other critics mean is that Mary gains a new representation of knowledge she already had, i.e., the old-fact/new-guise theory. Conee's view is clearly not that, although that is a point he does not stress. In fact, in an appendix, he expresses regret over not having studied Bigelow and Pargetter 1990 prior to preparing his article. Nevertheless, Conee clarifies the acquaintance analysis by characterizing it in relation to the Lewis-Nemirow Acquaintance Hypothesis. As Conee explains, he adopts the same strategy as Lewis and Nemirow, but with one difference. Like them, he claims that knowing what an experience is like is not just a special case of propositional knowledge. Unlike them, he claims that knowing what it is like is acquaintance knowledge, not know-how.

Another helpful part of Conee's article is his presentation of (what I call in this Guide) a semantic objection to the KA. Pre-release Mary can refer to color qualia by using demonstratives: she can speak of that special character of others' experiences.

I believe that both Conee's acquaintance analysis and his semantic objection to the KA ultimately fail (Alter, 1995a). But both should be taken seriously. Again, his article is a clear source of both.

Incidentally, Conee's article is one of the few commentaries on the KA that discusses Jackson's (1982) Fred case.

Dennett, D. (1991), Consciousness Explained, Boston: Little Brown and Company.

Dennett's discussion of the KA appears as a part of his overall attack on qualia. On his view, we should not trust the intuition that, when she leaves the room, Mary acquires new knowledge in any sense. He attributes the intuition that she acquires new knowledge to a failure to appreciate how much she would know prior to her release. For example, one might expect that, when she leaves, a blue banana could trick her. She knows that yellow is the typical color of ripe bananas, and so, one might think, she would be fooled into thinking she is recognizing a yellow experience when she leaves the room and sees what is actually a blue banana. However, Dennett explains, pre-release Mary would know which human brain processes correlate with seeing blue and which correlate with seeing yellow. Therefore, if she brings along her auto-cerebroscope when she leaves the room, she will not be fooled by the blue-banana trick.

I agree with Dennett that pre-release Mary could thwart the blue banana trick. But, unlike Dennett, I do not think that his reflections succeed in undermining the KA. One problem is that he tacitly assumes that what the pre-release Mary lacks are missing abilities. That view is part of the Ability Hypothesis, and defenders of the KA reject it. A more serious problem is that Dennett fails to explain why Mary's passing the blue-banana test shows that she learns no new facts upon leaving the room. See Jacquette 1995 and Alter 1998. However, Dennett was right to question the intuition that Mary gains knowledge when she leaves the room. Jackson himself (1998a) now denies that she does, and Dennett was one of the first to focus criticisms specifically at that premise.

I should also mention that Dennett writes in an entertaining style. Consciousness Explained was written not only, and probably not primarily, for philosophers. Whatever your philosophical views may be, the book is useful if only because it summarizes significant findings and theories in cognitive science (even though it dates from 1991). However, Dennett has strong opinions and, as his discussion of the KA indicates, his biases frequently affect his presentations of opposing positions.

Flanagan, O. (1992), Consciousness Reconsidered. Cambridge: M.I.T. Press.

Flanagan's discussion of the KA is only about five pages long, and it is worth reading. He distinguishes between linguistic and metaphysical physicalism. Metaphysical physicalism says that physical stuff and its relations exhaust what there is. Linguistic physicalism says that "everything physical can be expressed or captured in the languages of the basic sciences" (98). He then suggests that the KA may refute linguistic physicalism, but that it does not even threaten metaphysical physicalism. Pre-release Mary knows everything that can be expressed or captured in the languages of the basic sciences. What she does not know are still physical facts in the metaphysical sense of 'physical fact': some metaphysically-physical facts are not expressible in the languages of the basic sciences.

Flanagan's discussion is simple, clear, and concise. Viewed charitably, his argument could be seen as shifting the burden of proof on those who wish to defend linguistic physicalism or argue that metaphysical physicalism entails linguistic physicalism.

But there are several problems. First, he claims that linguistic physicalism is a stronger thesis than metaphysical physicalism. But, as he formulates the views, that is far from obvious. Someone might consistently believe that every physical thing can be captured or expressed in the languages of the basic sciences and also that some aspects of reality are non-physical. Yet the idea that everything is metaphysically physical and that there are some facts that cannot be captured or expressed in the languages of the basic sciences is, while not inconsistent, not obviously true. Second, it is not so clear that his point is best put in terms of whether all facts about color experiences can be captured or expressed in the languages of the basic sciences. In the first place, as Chomsky (1980, 6) points out, the languages of the basic sciences change to suit our explanatory needs. Moreover, an advocate of the KA can admit that all facts can be captured or expressed in the languages of the basic sciences, but still deny that pre-release Mary understands such facts (Alter 1998, section V, Bealer, 1994, fn. 9). Third, although Flanagan succeeds in shifting the burden of proof, he himself seems to think that he has accomplished more than that. He writes, "The argument is seductive, but easy to defeat" (98). I do not think he defeats the argument, even though I am sympathetic to his general approach.

Warning: Flanagan's book contains passages like the following: "Unless your friends are radical reductionists or epiphenomenalists, no one is trying to distance you or to disconnect you from your experiences. Don't worry. Be happy. Have your experiences. Enjoy them" (101).

Foss, J. (1989), "On the Logic of What It Is Like to be a Conscious Subject", Australasian Journal of Philosophy 67, pp. 305-20.

Foss attempts to do two things. First, he attempts to explicate Subjectivism-the view that "the subjective aspects of consciousness, which is its essence, cannot be objectively studied or characterized; that non-emergentist forms of physicalism are misconceived; and that the physical sciences are necessarily incomplete" (205). Second, he argues that Subjectivism is false.

This is a clear and well-written paper, which is generally fair to the views under review. But the criticisms of Subjectivism are, to me, not persuasive. In connection to the KA, he argues that pre-release Mary would know everything that those on the outside of her room would, or even might, report about color experiences. It seems doubtful that that claim entails that pre-release Mary would know everything there is to know about what color experiences are like. Perhaps Foss thinks otherwise because he tends to characterize Subjectivism as entailing the thesis that there is ineffable information about experiences. However, Subjectivism carries no commitment to an ineffability thesis. If pre-release Mary fails to know some information, that is because she does not understand the information. The information may nonetheless be expressible, albeit in a language she cannot understand (Alter 1998, section V, Bealer, 1994, fn. 9).

Furash, Gary (1989), "Frank Jackson's Knowledge Argument Against Materialism", Dialogue 32, pp. 1-6.

Gertler, Brie (1999), "A Defense of the Knowledge Argument," Philosophical Studies 93, pp. 317-36.

Gertler defends the KA against various prominent objections. She also criticizes the claim that the KA’s conclusion entails epiphenomenalism. Her goal is to challenge "the viability of the prevailing view of the phenomenal, which weds a "top down" approach to the mind with a materialist ontology" (317). (For the definition of "top down approach", see Churchland 1984, Matter and Consciousness, Cambridge: MIT Press, revised (edition, 1988), 1984 p. 96.)

This is an excellent paper. It’s nuanced, clear, and makes several original points. Gertler elaborates Conee’s objection to the Ability Hypothesis, that the abilities to recognize, remember, and recognize red are neither necessary nor sufficient for knowing what seeing red is like. She then argues that Conee’s acquaintance analysis lacks explanatory power, which is exactly the right point to make (cf. Alter 1998). She also criticizes the analogies typically used by old-fact/new-guise theorists.

Her discussions of the Ability Hypothesis and Conee’s acquaintance analysis are especially strong, but the paper has no weak parts. The discussion of epiphenomenalism is helpful. She doesn’t take on every objection to the KA. In particular, she doesn’t consider the idea that the KA’s conclusion seems stronger than it really is because it trades on an equivocation between metaphysical and epistemological senses of ‘physical fact’ (Horgan 1984, Alter 1998). And unlike Chalmers (1996), she doesn’t distinguish Loar’s (1990) sophisticated old-fact/new-guise theory from other versions. But her paper does everything it tries to do—which is a lot—convincingly.

Holman, E. (1987-8), "Qualia, Kripkean Arguments, and Subjectivity", Philosophical Research Archives 13, pp. 411-429.

Horgan, T. (1984), "Jackson on Physical Information and Qualia", Philosophical Quarterly 34, pp.147-52.

Horgan claims that pre-release Mary does not know all the physical facts, even though she does know all the explicitly physical information-all the information expressed by true sentences that are formulated in explicitly physical language. On his view, the physicalist is committed only to the view that all information is ontologically physical, not that it can be expressed in explicitly physical language. The KA acquires its apparent plausibility from an equivocation on the two sense of 'physical information'.

Horgan's paper is clear, concise, and well argued. The basic idea, that the KA trades on an equivocation, is important. Others make similar claims (Alter 1998, Flanagan 1992), but Horgan made it first. The paper is often read as expressing the old-fact/new-guise theory, and that reading has some basis in what Horgan writes. For example, he illustrates his distinction between two senses of 'physical fact' with standard examples that suggest an old-fact/new-guise theory; he says that 'Superman can fly' and 'Clark Kent can fly' express different information even though they both attribute the same property to the same individual.

One problem with the paper is that Horgan does not separate the old-fact/new-guise idea from the charge of equivocation. The ideas are separable (Alter 1998). One might also quibble with his tendency to formulate his points semantically, in terms of what certain sentences express-one should be able to accept his criticism of the KA while remaining neutral on the semantic issues he raises.

Jackson, F. (1982), "Epiphenomenal Qualia", Philosophical Quarterly 32, pp. 127-36. avalaible online

This is where it all began: here Jackson states his KA for the first time. To illustrate the argument, he uses both the Mary case and the case of Fred, who sees two colors where we see one. In addition to giving the KA, he distinguishes it from Nagel's related argument in Nagel 1974 and from modal arguments associated with Kripke (1972). He also defends epiphenomenalism against three common objections (section IV). One is that it is "just obvious that the hurtfulness of pain is partly responsible for the subject seeking to avoid pain, saying 'It hurts' and so on." Another is that, based on the theory of natural selection, "we should expect qualia to be conducive to survival. The objection is that they could hardly help us to survive if they do nothing to the physical world." The third is that epiphenomenalism would make the inference to other minds harder than it actually is: "how can a person's behavior provide any reason for believing he has qualia like mine, or indeed any qualia at all, unless this behavior can be regarded as the outcome of the qualia"? He ends the article by arguing that physicalism is based on an overly optimistic view about our powers of understanding. He writes, "The wonder is that we understand as much as we do, and there is no wonder that there should be matters which fall quite outside our comprehension. Perhaps exactly how epiphenomenal qualia fit into the scheme of things is one such."

It hardly needs saying that reading this article is "a must" for anyone interesting in the KA. Accept no substitutes: read the original. In addition to clearly articulating the KA, Jackson puts it in context and usefully distinguishes it from similar, well-known arguments. It is a splendid article, despite Jackson's (1998a) recent change of view and despite all the criticisms that been leveled against the KA.

One problem with the article is that Jackson does not say much about why all physical facts would be known by the pre-release Mary. The latter premise is crucial to the KA, because Jackson is trying to draw metaphysical conclusions based partly on epistemic premises. Jackson should have said more about why he thinks it is true, or at least acknowledged the need for doing so. He says more in Jackson 1995.

Another problem concerns the relationship between epiphenomenalism and the KA. Evidently, Jackson assumes there is a close connection between the two, but he does not explain it. Instead, after presenting the KA and distinguishing it from other arguments, he begins defending epiphenomenalism, as though it were obvious that irreducibly subjective qualia could have no effects on the physical world. But that is not obvious; see Searle 1992. For an argument that the KA entails epiphenomenalism, see Lewis 1988.

Jackson's defense of epiphenomenalism is adequate as far as it goes. The second two objections, relating to natural selection and knowing other minds, are based on rather unsophisticated mistakes, as Jackson's replies make clear. The first objection, that qualia obviously have physical effects, is somewhat underdeveloped. For that reason, Jackson's response, that our intuitions about causality can mislead, is rhetorically adequate. But epiphenomenalism is still generally regarded as an implausible theory, and Jackson makes it especially attractive.

Jackson's closing remarks about how physicalism derives from an overly optimistic view about our powers of understanding are intriguing, if inchoate. That is also one of Nagel's main themes in his writings about the mind-body problem. One could also imagine Jackson's remarks appearing in Colin McGinn's writings on the topic. McGinn argues that we are constitutionally incapable of understanding how brain processes give rise to subjective consciousness (1989, "Can We Solve the Mind-Body Problem?" Mind 98: 349-66).

Jackson is somewhat uncharitable to Nagel in his brief discussion of Nagel's argument. Jackson claims that, unlike his own KA, Nagel's argument against physicalism depends essentially on claims about what humans can and cannot imagine. While Nagel's argument does involve such premises, a careful examination of his reasoning reveals that the reliance is inessential. I explain this in "Nagel on Imagination and Physicalism", forthcoming in the Journal of Philosophical Research.

Jackson, F. (1986), "What Mary Didn't Know", The Journal of Philosophy, 83, pp.291-95.

This is Jackson's reply to Churchland's (1985) criticisms of the KA. The short paper begins with three useful points, which Jackson describes as "clarifications", although the second is really more than a clarification. First, the KA does not depend on claims about what pre-release Mary is and is not able to imagine. Second, Churchland's charge that Jackson commits the intensional fallacy is implausible, because it is implausible that pre-release Mary's lack of knowledge "could be remedied merely by her explicitly following through enough logical consequences of her vast physical knowledge" (292). Third, the knowledge that pre-release Mary lacks is about the experiences of others, not herself. This section should be studied carefully by anyone thinking seriously about the KA.

Although Churchland refers to two shortcomings in the KA, Jackson divides Churchland's objections into three. The first and third objections fail, Jackson argues, because they are based on the misunderstandings of the argument that Jackson clarifies at the beginning of his article. In the course of discussing the point about the intensional fallacy, Jackson provides a summary of the argument that those discussing the argument have since followed: Mary (before her release) knows everything physical there is to know about other people; Mary (before her release) does not know everything there is to know about other people (because she learns something about them upon her release); therefore, there are truths about other people (and herself) which escape the physicalist's story. That summary has proven helpful. However, the summary is misleading in one respect. In the summary I provide in this Guide, I describe the Mary case by appealing to the notion of what can be learned through watching black-and-white lectures. Doing so allows us to raise the question, "Can everything physical be learned by watching lectures on black-and-white television?" In Jackson's summary, what the pre-release Mary learns is described with the phrase 'everything physical'. That obfuscates what in the KA can be stipulated (that pre-release Mary knows everything that one can know by watching black-and-white lectures) and what cannot be stipulated (that Mary's pre-release knowledge includes all physical information).

Churchland's second objection, unlike his first and third, is not based on a misunderstanding about how the KA should be formulated. Churchland gives a "parity of reasons" objection: pre-release Mary might also get black-and-white lectures from a dualist, explaining the laws governing ectoplasm. Therefore, if the KA refutes physicalism, then it refutes dualism too. However, Jackson replies, the cases are not analogous. It is more plausible that pre-release Mary has the full story according to physicalism than that she has the full story according to dualism. Jackson might have said more about why the full dualist story cannot be delivered by black-and-white lectures. Why would the assumption that mental events are non-physical make any difference to what pre-release Mary knows (see Nagel 1986, 29)? The answer is presumably that only the physicalist is committed to denying that qualia are subjective. But that answer requires explanation and defense.

In a footnote, Jackson replies to Horgan's (1984) criticisms of the KA. The reply is cryptic and, to my mind, unpersuasive. He says simply, "The claim here is not that, if physicalism is true, only what is expressed in explicitly physical language is an item of knowledge. It is that, if physicalism is true, then if you know everything expressed or expressible in explicitly physical language, you know everything." However, it is unclear, to me at least, that Jackson's remark defeats Horgan's objections.

Jackson, F. (1995), "Postscript" In Contemporary Materialism. Ed. P. Moser and J. Trout. New York: Routledge, pp. 184-9.

Here Jackson defends his assumption that if physicalism is true then the pre-release Mary should be able to deduce the nature of color experiences. This is important because some of the KA's critics argue that, while the facts pre-release Mary knows fix the facts about human color experiences, physicalism does not entail that she would be able to deduce the latter facts. Those objectors sometimes appeal to an analogies to H20 and water: water is H20 even if that fact is not deducible a priori. Jackson argues, however, that "a rich enough story about the H20 way things are does enable the a priori deduction of the water way things are" (188). Thus he believes the following argument is valid: H220 covers most of the planet; H20 fills the water role; therefore, water covers most of the planet. Likewise, he claims, if pre-release Mary's physical knowledge is really complete, then she should be able to deduce what it's like to see in color. However, the most Jackson shows is that the analogies to water and H20 do not provide the physicalist with an adequate model for objecting to the KA. But that should come as no revelation. The burden still seems to be on the KA's defenders to show that physicalism entails that pre-release Mary knows everything about color experiences. I do not believe that Jackson quite meets that burden in this postscript. Also, the points about a prioricity are potentially misleading in the present context, since Mary is entitled to use empirical methods, so long as she does not cheat by having color experiences (e.g., by directly stimulating the relevant areas of her brain).

Jackson, F. (1998a), "Postscript on Qualia" in his Mind, Method and Conditionals. London: Routledge.

Here is Jackson's official announcement that he now rejects the KA and believes physicalism to be true. He believes that Mary does not learn anything when she leaves the room, and he thinks that the real puzzle is to explain why the intuition to the contrary is so strong. He suggests the following explanation. Learning a physical fact often involves making inferences; it is often a long and complex process. By contrast, when Mary leaves the room, her gain in knowledge is almost immediate. We therefore infer, wrongly but naturally, that the knowledge gained cannot be knowledge of physical facts.

Jackson's conjecture is incomplete at best. Barbara Montero, in her article reporting Jackson's change of mind, explains why: "I can come to know that Sally, the adult Elephant is larger than Henry, the baby goat, at a mere glance. And so whether an explanation of the sort Jackson proposes [is correct] will, of course, be open to debate" (Montero, 1999). One point Jackson makes in passing is worth noting. It is incorrect to say that one cannot know what it's like to see in color without having color experiences: memory traces are enough, and one might in principle be given the memory traces without actually having the experiences.

Jacquette, Dale (1995), "The Blue Banana Trick: Dennett on Jackson's Color Scientist", Theoria 61, pp. 217-30.

Jacquette argues that Dennett's color-recognition test is invalid: it cannot be used to decide whether Mary learns something new when she leaves the room. That point is plausible, and his discussion is thorough and persuasive. It is not entirely clear to me that Dennett's position warrants such a detailed reply; the basic point is quite simple. But given Dennett's stature, perhaps it is needed.

Levin, J. (1985), "Could Love Be Like A Heatwave? Physicalism and the Subjective Character of Experience", Philosophical Studies 49, pp. 245-261.

Levin begins by noting that the Ability Hypothesis provides a challenge to the Nagel-Jackson arguments. She then says that she will discuss an argument similar to those arguments but which is "considerably harder to refute" (480), which is presented in Warner (1986). Warner's argument has two premises: "if one lacks certain experiences, one will lack a certain recognitional or discriminative ability-an ability to know that one is in a particular state without making inferences, or consulting instruments, but simply by applying one's concept of that mental state to the experiences at hand. The second premise is that this capacity to recognize or discriminate among mental states is require for having full and complete factual knowledge of them" (480). Levin argues that even Warner's argument fails, because it trades on an equivocation between two senses of "direct recognitional capacity", the source of which is "a failure to distinguish between having a concept and having the wherewithal to apply it." In terms of Jackson's example, pre-release Mary lacks the ability to apply certain concepts, not the concepts themselves.

The first thing to say about this paper is that, despite its clever title, emotions are mentioned nowhere in the body of the paper. Second and more significantly, her decision to discuss Warner's argument instead of the Nagel-Jackson arguments is not well motivated. Although she regards Warner's arguments as stronger than the Nagel-Jackson arguments, she does not defend that view adequately. And the result is that it is somewhat difficult to tell whether her arguments actually refute the KA. Some of her points are clearly relevant, but beyond that their implications are not clear. For example, she maintains that, although experiences of specific kinds contribute to knowledge, "What makes it special, however, is not that experience contributes a chunk of knowledge that could not be gleaning in any other way, but that it contributes such knowledge as it does so efficiently" (489). If she could establish that position, then perhaps she could argue that the pre-release Mary lacks no knowledge about color experiences. But Levin does not, as far as I can tell, provide a strong argument for her conclusion. Also, the way Levin formulates the problem comes close to presupposing that some version of the Ability Hypothesis is true; otherwise, why would an ambiguity in the notion of a "direct recognitional capacity" be so central?

Lewis, D. (1983), Postscript to "Mad Pain and Martian Pain." In his Philosophical Papers, vol. 1. New York: Oxford University Press, pp. 130-32.

This is the first place where the Ability Hypothesis is applied specifically to the KA, and where Lewis first introduces his Vegemite example. (Lewis gives credit for the general idea to Nemirow (1980), who applies it to related arguments by Nagel.) This is an excellent little paper. Lewis gives a more thorough treatment in Lewis 1988, but this presentation is short and sweet. Not a word is wasted. For criticisms of the Ability Hypothesis, see Raymont 1999, Alter 1998, Conee 1994, Bigelow and Pargetter 1990, Loar 1990, Lycan 1996, and Chalmers 1996.

Lewis, D. (1988), "What Experience Teaches" In Proceedings of the Russellian Society. Sydney: University of Sydney. Rpt. in Mind and Cognition. Ed. W. Lycan. Oxford: Blackwell, 1990, pp. 499-518.

Here Lewis expands on his presentation of the Ability Hypothesis in Lewis 1983. This article is rich, clear, and entertainingly written. If one merely wants to know what the Ability Hypothesis is and how it would block the KA, then reading Lewis 1983 is sufficient. But the present paper contains much more than a statement of his view. For example, Lewis brings together various cases, including his Vegemite case, the case of bat-echolocation experiences that is associated with Nagel 1974, and the Mary case. He also discusses six different ways to miss the point of the KA. These should be studied closely. In particular, the "fifth way" of missing the point is, in effect, an argument against the old-fact/new-guise theory. Lewis points out that the thesis that Mary gains information upon leaving the room is not undermined by the idea that what she gains is analogous to learning a new word or language. He also explicates the notion of gaining information in terms of eliminating possibilities. It is not clear to me that his notion is equivalent to Jackson's, but Lewis' explication is precise and interesting. Also, he argues that if we accept that Mary gains new information when she leaves the room then epiphenomenalism is inescapable. Jackson (1982) suggests such a view, but he provides no argument for it. For that reason alone, argument is well worth studying.

Loar, B. (1990), "Phenomenal States" In Philosophical Perspectives IV: Action Theory and Philosophy of Mind. Ed. J. Tomberlin. Atascadero: Ridgeview, pp. 81-108.

This is, among other things, an explanation and defense of the old-fact/new-guise theory. Loar cites Jackson 1982 in a footnote, but he does not discuss the KA specifically. Nevertheless, it is relatively clear how Loar's view, if correct, would apply to (and undermine) the KA. Loar distinguishes between phenomenal and functional concepts, but argues that the properties they pick out may be identical. So far, that is a fairly standard formulation of the old-fact/new-guise theory. But Loar takes it one step farther. He argues that, not only do phenomenal and functional concepts co-refer, they express the same properties, even though we cannot know that a priori. That is a stronger claim than other old-fact/new-guise theorists make.

Along the way, Loar provides two brief arguments against the Ability Hypothesis. One is that it cannot account for the meaning of expressions like 'feels like such and such' in conditionals such as "if pains feel like such and such, then Q". I believe there is something in this objection, but it would have helped to explain why it works, if it does. His second objection is less interesting. He states, "For many conceptions of phenomenal qualities, there simply is no candidate for an independently mastered term instances of which one then proceeds to learn how to recognize: my conception of a peculiar way my left knee feels when I run (a conception that occurs predicatively in various judgments) is not my knowing how to apply an independently understood predicate" (86). However, I do not see why Lewis and Nemirow are committed to saying that there is an ability term (in English?) for every alleged phenomenal quality.

The style of the passage quoted in the previous paragraph is typical of the article. The article is dense in style and content. For criticisms of Loar’s sophisticated old-fact/new-guise theory, see Chalmers (1996, pp. 142-3) and Chalmers "Materialism and the Metaphysics of Modality", Philosophy and Phenomenological Research, September 1999, section 3.4).

Lycan, W. (1996), Consciousness and Experience. Cambridge: M.I.T. Press.

Chapter 5 is basically the same material as Lycan 1995. He gives no less than ten concisely-presented arguments against the Ability Hypothesis. Not all are original, as he acknowledges. For example, he uses Loar's (1990) idea involving embedding "what it's like" in conditionals. Some of the ten arguments seem like pot shots; some are more substantive. It might have been helpful for Lycan to reflect a bit more on the relative merits of each and how they relate to one another. Still, Lycan's presentations of objections are crisp and precise. He also offers an interesting positive account of phenomenal information, which comports with his old-fact/new-guise theory. And he offers criticisms of Lewis' attacks on phenomenal information, based in part on his account of the notion. It is a useful chapter.

Lycan, W. (1990), "What is the 'Subjectivity' of the Mental?" In Philosophical Perspectives IV: Action Theory and Philosophy of Mind. Ed. J. Tomberlin. Atascadero: Ridgeview, pp. 109-130.

This paper is similar to Loar (1990): it is, among other things, an explanation and defense of the old-fact/new-guise theory. The paper is chock-full of arguments. Its official goal is to provide a materialist account "of the subjectivity or perspectivialness or point-of-view-iness of the mental" (110). He keeps his promise. Lycan suggests that the alleged subjectivity of the mental amounts to nothing that creates the least problem for the materialist. He writes, "I know my pain by introspection, and my representation of disorder is directly formed by introspection and has obvious immediate inferential and other functional properties" (125). He claims that such facts encourage "the illusion of an ontologically special kind of state of affair"-of intrinsically perspectival or subjective facts. But, he says, we should not accept such claims at face value: "There are only states of subjects that both function in a particularly intimate way within those subject and have the subjects themselves and their other states as inevitable referents. And that, I think, is all there is to 'subjectivity'."

Many readers, I suspect, will not be convinced that that is all there is to subjectivity. And those interested in the KA will not find a detailed discussion of it here. Also, Lycan's discussion of Nagel's argument is not very charitable. He criticizes Nagel at length for "act-object jargon"-that is, for assuming that experiences are themselves (typically) objects of perception. I do not see why Nagel's argument (or Jackson's) depends essentially on any such assumption, and the phrases that Lycan quotes from Nagel can and should be understood in more charitable ways.

Lycan, W. (1995),"A Limited Defense of Phenomenal Information." In Conscious Experience. Ed. T. Metzinger. Tucson: University of Arizona Press, 1995.

This is basically the same as Chapter 5 of Lycan 1996.

McMullen, C. (1985), "'Knowing What It's Like' and the Essential Indexical", Philosophical Studies 48, pp. 211-33.

"Perry meets Jackson": McMullen likens knowing what it's like to indexical knowledge. He argues that Mary's gain in knowledge upon leaving the room is comparable to the absent-minded U.S. historian who learns that today is July 4th, America's Independence Day. There is in the example no reason to regard the latter fact as non-physical; likewise, McMullen says, for Mary's new knowledge. This article is the clearest expression I have read of the indexicalization view. It is a clearer expression of the view than Bigelow and Pargetter 1990, since (a) McMullen focuses explicitly on indexicals, and (b) Bigelow and Pargetter also suggests an old-fact/new-guise theory and an acquaintance knowledge theory. However, I do not find the comparison of the Mary case to the case of indexicals especially enlightening, especially if detached from the old-fact/new-guise theory. One of McMullen's main moves is to distinguish the cognitive significance of a referring expressions from its reference. It is, in my view, doubtful that that distinction gets at the core of the problem raised by the KA. But the article is interesting and clever.

Mellor, D. (1993), "Nothing Like Experience" Proceedings of the Aristotelian Society 93, pp. 1-16.

This is a defense of the Ability Hypothesis. What distinguishes it from Lewis' and Nemirow's papers is the absence of physicalist underpinnings. Mellor's primary reason for accepting what he prefers to call the know-how theory is that "that theory alone" (9) explains (a) why we cannot put that which Mary learns when she leaves the room into words and (b) why we cannot have such knowledge without imagining the relevant experiences. It is not clear to me what Mellor's discussion of the Ability Hypothesis adds to Lewis' and Nemirow's papers. Crane and Mellor (1990, "There is no Question of Physicalism," Mind 99: 185-206) is a much more important piece, which is also directly relevant to the KA. Even so, the present paper is useful because it detaches the Ability Hypothesis from physicalism. The final section on the nature of experience is definitely worth reading, because it expresses in a particularly forceful way the idea that science need not tell us what experiences are "in themselves".

Nagel, T. (1974), "What is it Like to be a Bat?" Philosophical Review 83, pp. 435-50. Rpt. in his Mortal Questions. Cambridge University Press, 193-214.

Several philosophers have suggested that Jackson’s KA is a more austere version of the argument against physicalism presented in Nagel 1974. The phrase ‘the Nagel-Jackson argument’ is not uncommon (see, for example, Pereboom 1994). And although Jackson (1982) draws distinctions between the KA and Nagel’s argument, Jackson acknowledges a considerable debt to Nagel’s article.

Nagel’s article is almost certainly more influential than either Jackson 1982 and 1986 is.

Dennett (1991, p. 441) writes, "The most widely cited and influential thought experiment about consciousness is Thomas Nagel’s ‘What Is It Like to Be a Bat?’" It is a remarkably rich and nuanced work. If it is an early expression of the KA, it is also much more than that. (I discuss Nagel’s argument in depth in "Nagel on Imagination and Physicalism", forthcoming in the Journal of Philosophical Research.)

Nagel presents his arguments primarily as a challenge to the "several analyses of mental phenomena and mental concepts designed to explain the possibility of some variety of materialism, psychophysical identification, or reduction" (165) that were popular in 1974, and which remain popular today. Very roughly put, the problem he identified was that such analyses fail to deal adequately with phenomenological consciousness; they ignore or badly misconstrue the very aspect of the mind that "makes the mind-body problem really intractable" (165). The KA is an articulation of that basic idea.

Nagel illustrates the problem with the example of bat consciousness. In order to understand what it’s like to be a bat, we would have to understand what bat-echolocation experiences are like for the bat—what they are like from the bat’s perspective. In our present state of knowledge, the only way to acquire such understanding would be to adopt the bat’s point of view, e.g., by accurately imagining its experiences; having complete objective physical knowledge would not suffice. But, he argues, we cannot adopt the bat’s point of view, since it is too different from our own. So, he concludes, there appear to be facts (information) about the bat’s experiences that are not captured by the objective physical facts about its brain, behavior, etc.

Nagel’s point about bat consciousness generalizes. There would appear to be subjective facts—facts that cannot be understood except by adopting the viewpoint of the experiencing organism. That constitutes a challenge to physicalism, since, according to Nagel, physicalism entails that all facts are objective, in the sense that (roughly put) they can be understood without adopting the viewpoint of the experiencing organism. And that argument is very similar to the KA. The main problem can be put concisely as follows: complete physical knowledge about the bat’s experiences (alternatively: about color experiences) does not entail knowledge of the facts about what those experiences are like.

Nagel’s conclusion about physicalism is more guarded than Jackson’s conclusion. Nagel concludes only that we cannot presently understand how physicalism might be true. In fact, he suggests that we may eventually come to understand how physicalism might be true. He compares the present status of physicalism to "that which the hypothesis that matter is energy would have had if uttered by a pre-Socratic philosopher" (177). Thus, physicalism might be true, but we have to admit that at present we cannot understand how that could be—because it would entail something that seems obviously false, namely, that all facts about consciousness can be rendered objective. However, in Nagel 1986, Nagel’s conclusion about physicalism is not similarly guarded. There he concludes that physicalism is false, and he is no longer hesitant to endorse the doctrine of subjective facts (he says that there are such facts, not merely that there appear to be).

One point about the bat case should be noted, even though the point does not ultimately damage Nagel’s reasoning: Nagel’s discussion of bat consciousness is oversimple. Allen and Bekoff (1997, p. 155) note that the bats that use echolocation include "about 30 species from the genus Rousettus in the suborder Megachiroptera, and all of the approximately 660 species in the suborder Microchiroptera…." Bats echolocate in at least three different ways, constant frequency, frequency sweep, and short burst. Therefore, as Allen and Bekoff conclude, "Presumably there is no one thing it is ‘like to be a bat,’ any more than there is one thing it is ‘like to be a primate.’"

Nemirow, L.(1980), Review of Mortal Questions, by Thomas Nagel. Philosophical Review 89, pp. 473-77.

This is the first published statement of what Lewis later dubbed ‘the Ability Hypothesis’ and what Nemirow (1990) later dubbed ‘the ability analysis’. The review pre-dates Jackson’s KA; Nemirow makes the point in connection to Nagel’s version of the argument. Nemirow clearly states that to know what an experience of seeing red is like is to have abilities, such as the ability to remember or imagine the experience (he says "to remember a similar experience," presumably because he has token experiences in mind). In addition to presenting the Ability Hypothesis, he claims that Wittgenstein’s private language argument shows that there are no subjective facts, and he criticizes Nagel for failing to recognize this.

The review is clearly and concisely written. One could quibble with his scholarship, since in Mortal Questions Nagel is more tentative about the existence of subjective facts than one would gather from reading Nemirow’s review. For example, in "What is it Like to be a Bat?" (Nagel 1974), Nagel’s claim is only that there would seem to be subjective facts; he leaves open the possibility that apparently subjective facts could one day be rendered objective. In his later writings (e.g., Nagel 1986), however, Nagel claims unequivocally that some facts are subjective.

One more quibble. Nemirow dismisses Nagel’s "understanding of understanding" as "naïve", and he thereby implies that his (Nemirow’s) own view is not similarly naïve. Lewis (1988) gets it right when he presents the Ability Hypothesis as merely an alternative theory, not necessarily a less naïve one.

Nemirow, L.(1990), "Physicalism and the Cognitive Role of Acquaintance" In Mind and Cognition. Ed. W. Lycan. Oxford: Blackwell, pp. 490-99.

Here Nemirow explicates the ideas he expressed more inchoately in Nemirow 1980. His presentation of the Ability Hypothesis and how it would refute the KA is similar to Lewis’. There are some differences of detail, but it is not clear (to me) that the differences are terribly significant. For example, Nemirow presents the Ability Hypothesis in connection to Feigl’s discussion of Nagel-Jackson type problems in Feigl 1967, in addition to Nagel’s and Jackson’s arguments. Also, Nemirow asserts that "Knowing what it’s like may be identified with knowing how to imagine"; Lewis (1988) also includes the abilities to remember and recognize. (Nemirow 1980, however, mentions the ability to remember.) Later in the article (section V), Nemirow discusses recognitional abilities, but gives reasons to think that imaginative abilities are more central to knowing what it’s like.

The article includes a discussion of objections. Unfortunately, the objections he considers are not those that many have found convincing. The latter have appeared in works published since 1990. (See, for example, Conee 1994.) Therefore, Nemirow cannot be blamed for omitting discussion of them.

The paper is not as rich as Lewis 1988. It bears repeating, however, that Nemirow had the idea first; Lewis is very clear about this.

Nida-Rumelin, M. (1998), "On Belief about Experiences: An Epistemological Distinction Applied to the Knowledge Argument against Physicalism", Philosophy and Phenomenological Research 58, pp. 51-73.

Papineau, D. (1993), Philosophical Naturalism Oxford: Blackwell.

Pereboom, D. (1994), "Bats, Brain Scientists, and the Limitations of Introspection", Philosophy and Phenomenological Research 54, pp. 315-29.

This is a defense of (what I call) the old-fact/new-guise analysis. Pereboom points out a shortcoming of the analysis as typically presented. He then introduces a revised version of the analysis, and argues that the revised version succeeds in blocking the KA. He argues that the KA is based on the doubtful assumption that, if materialism is true, then one has cognitive aspects to everything about one’s own mental states.

One difficulty with the article is that Pereboom does not make a compelling case for his claim that if materialism is true, then one has cognitive aspects to everything about one’s own mental states. I give another criticism of Pereboom’s particular version of the old-fact/new-guise analysis in Alter 1995a. But even if my criticisms are correct, they do not undermine many of the ideas in Pereboom’s article. My criticisms do not undermine every version of the old-fact/new-guise analysis. And Pereboom makes a number of interesting moves. For example, Pereboom explains the connection between Nagel’s (1974) argument and Jackson’s in a very natural way. And he points out a problem for the analogy sometimes invoked between knowledge represented with indexicals (e.g., the difference between representing the same fact with "I weigh 185 pounds" and "You weigh 185 pounds) and the knowledge Mary is supposed to gain when she leaves the room.

Pereboom kindly showed me the article prior to its publication. I learned a lot from it and from discussing the KA with him. Working on his article was what got me interested in the subject in the first place.

Pinker, S. (1997), How the Mind Works, New York: W. W. Norton & Co., 1997.

Raymont, P. (1995), "Tye's Criticism of the Knowledge Argument", Dialogue 34, pp. 713-26.

Raymont, P. (1999), "The Know-How Response to Jackson's Knowledge Argument", Journal of Philosophical Research 24, pp. 113-126.

This is a criticism of the Ability Hypothesis and a defense of the KA. Before presenting his criticisms, he notes a problem for a different criticism. The criticism he finds problematic is one offered by Bigelow and Pargetter (1990) and Seager (1991), It runs as follows. Blindsight may provide one with the ability to, e.g., recognize red objects without providing knowledge of what it’s like to see red objects; therefore, the Ability Hypothesis is false. Raymont claims that that objection misses the mark, because "although the blindsight subject is able to recognize things in her environment without consciously experiencing them she might not be able to recognize conscious experiences of those things (were she to have any)" (115). Raymont finds other grounds for concluding that recognitional abilities and the corresponding knowledge of what an experience is like can come apart. He cites "empirical data in support of the view that one can have the ability to directly (i.e., noninferentially) recognize a certain type of visual experience without ever having had it, and thus without knowing what it is like to have it" (115). The empirical data concern the recognitional and inferential abilities of infants less than one month old (Meltzoff and Borton, "Intermodal Matching by Human Neonates," Nature 282: 403-4). Raymont gives several other arguments. He gives an argument similar to Conee’s (1994), intended to show that imaginative abilities do not entail the relevant knowledge. And he argues that combining various abilities conjunctively or disjunctively does not answer the Conee-like objections. He also provides another objection to the Ability Hypothesis, based on some views of Papineau (1993).

This is a nuanced paper, and those interested in the Ability Hypothesis should examine it. I have only one quibble. At places Raymont seems to assume that one must have an experience of (say) seeing red in order to know what that kind of experience is like (e.g., see the quotation in the previous paragraph beginning "empirical data"). In the context of his paper, he should not make that assumption without argument. Consider Jackson’s claim to the contrary: "Seeing red and feeling pain impact on us, leaving a memory trace which sustains our knowledge of what it is like to see red and feel pain on the many occasions where we are neither seeing red nor feeling pain. This is why it was always a mistake to say that someone could not know what seeing red and feeling pain is [sic] like unless they had actually experienced them: false ‘memory’ traces are enough." (Jackson 1998, 77)

Robinson, H. (1993), "Dennett on the Knowledge Argument", Analysis 53, pp. 174-77.

Robinson, H. (1996),"The Anti-Materialist Strategy and the Knowledge Argument." In Objections to Physicalism Ed. H. Robinson. New York: Oxford University Press, pp. 159-84.

Sacks, O. (1995), "The Case of the Colorblind Painter" In his An Anthropologist on Mars. New York: Knopf, pp. 3-41.

Seager, W. (1991), Metaphysics of Consciousness. New York: Routledge.

Searle, J. (1992), The Rediscovery of the Mind. Cambridge: M.I.T. Press.

Stemmer, Nathan (1989), "Physicalism and the Argument from Knowledge" Australasian Journal of Philosophy 67,pp. 84-91.

Defends physicalism against the KA. Stemmer’s main claim is that physicalists have the resources to explain all the data in the Mary case, by appeal to neurophysiology.

I find this paper unconvincing and unhelpful. Stemmer seems to assume that the data requiring explanation are such things as Mary’s utterances upon leaving the room. Not surprisingly, Stemmer finds a physicalistically consistent explanation of that data, in terms of "the direct or indirect effects of [a] neurological entity" (87). He fails to confront Jackson’s intuition that Mary gains knowledge when she leaves the room—the reasons for the utterances she might (or might not) make. There is also a minor terminological error worth noting. Stemmer characterizes what he does as establishing the "invalidity" of the KA, while in fact his complaints concern the truth of the argument’s premises, not its reasoning.

Teller, D. (1992), "A Contemporary Look at Emergence" In Emergence or Reduction? Prospects for Nonreductive Physicalism Ed. A. Beckermann, H. Flohr, and J. Kim. Berlin: De Gruyter.

Tye, M. (1986), "The Subjective Qualities of Experience", Mind 95, pp. 1-17; reprinted as Chapter 6 of his The Metaphysics of Mind. Cambridge University Press (1989).

This is a fairly standard presentation of (what I call) the old-fact/new-guise analysis. One distinctive aspect of the article is Tye’s use of the Kripkean (1972) apparatus of rigid designators. He argues that the physicalist should argue as follows. Let ‘R’ rigidly name the phenomenal property with which pre-release Mary is allegedly unfamiliar (actually, Tye does not use Jackson’s Mary example, but he uses similar examples). Even though Mary does not have the relevant phenomenal representation, ‘R’ rigidly names a property that is in fact purely physical; so, she does know the facts involving R, e.g., the fact that a certain experience e had by others outside the room has R.

I do not accept the old-fact/new-guise analysis, for reasons I explain in Alter 1995a. But I have little to say about this particular presentation of the view. Here is one quibble. The reason Tye invokes the notion of a rigid name is, presumably, that the alternative would be a name with descriptive content. If the name of the phenomenal property in question had descriptive content, then it would be less plausible to say "Experience e has R" expresses the same information as if we replace ‘R’ with its physicalistic counterpart (i.e., with a physical description of the physical property that ‘R’ names). But in that case it is not the rigidity of ‘R’ that is relevant, but rather the assumption that ‘R’ lacks descriptive content. Rigidity and lacking descriptive content are distinct notions. ‘The positive square root of the number 16’ is a rigid designator (it designates the same thing in all possible worlds), but it has descriptive content. See David Kaplan, "Demonstratives" (in Themes From Kaplan, Ed. J. Almog, J. Perry, H. Wettstein, New York: Oxford University Press, 1989: 481-563). As far as I can tell, rigidity is irrelevant to the discussion of the KA.

Van Gulick, R. (1993), "Understanding the Phenomenal Mind." In Consciousness. Ed. M. Davies and G. Humphries. Oxford: Blackwell, pp. 137-54.

Vidal, Javier (1995), "La Cuestion de los 'Qualia'", Anu Filosof 28, pp. 425-41.

Watkins, M.(1989), "The Knowledge Argument Against the Knowledge Argument", Analysis 49, pp. 158-60.

Watkins argues that Jackson’s epiphenomenalism undermines the KA. Briefly, the argument is this. By epiphenomenalism, qualia do not cause beliefs about qualia (since they are causally inefficacious). Therefore, Mary would have had the same beliefs regarding qualia even if she had no qualia. Therefore, all she gains when she leaves the room are unjustified beliefs. But "Surely physicalism can account for unjustified beliefs."

This is a clever argument, and the article is clearly and concisely written. But there are problems with the argument.

First, Jackson could respond by giving up epiphenomenalism, as Watkins allows. Although some (e.g., Lewis 1988) argue that Jackson is committed to epiphenomenalism, that is controversial. Second, one may doubt that physicalism can account for beliefs, justified or not, especially where the beliefs are purportedly about qualia. Third, in Jackson 1982 (section IV), Jackson states, "I will say nothing about two views associated with the classical epiphenomenalist position. …The second is that the mental is totally causally inefficacious. For all I will say it may be that you have to hold that the instantiation of qualia makes a difference to other mental states though not to anything physical. Indeed general considerations to do with how you could come to be aware of the instantiation of qualia suggest such a position." So, Jackson allows that qualia may have effects on mental states, presumably including beliefs about qualia. So, even if Watkins argument showed that some versions of epiphenomenalism are incompatible with endorsing the KA, it does not show that Jackson’s version of epiphenomenalism (the one he held in Jackson 1982) has that implication.

Yi, Byeong (unpublished), "The Nature of What Mary Didn't Know", University of Queensland (Australia) ms.

Other references

[not on the KA; just mentioned in passing in the discussion]

Allen, C., Bekoff, M. (1997), Species of Mind: The Philosophy and Biology of Cognitive Ethology, MIT Press.

Chomsky, N. (1980), Rules and Representations, New York: Columbia University Press

Chomsky, N. (1994), Language and Problems of Knowledge. Cambridge: M.I.T. Press.

Donnellan, K. (1979), "The Contingent A Priori and Rigid Designators" In Contemporary Perspectives in the Philosophy of Language. Ed. P. French, T. Uehling, and H. Wettstein. Minneapolis: University of Minnesota Press, pp. 45-60.

Feigl, H. (1967), The 'Mental' and the 'Physical'. The Essay with a Postscript. Minneapolis: University of Minnesota Press.

Geach, P. (1960), "Ascriptivism", Philosophical Review 69, pp. 221-5.

Jackson, F. (1982b), "Acting, Trying, and Essentialism", Inquiry 25, pp. 255-62.

Jackson, F. (1998b), From Metaphysics to Ethics. Oxford: Clarendon Press.

Kaplan, D. (1989), "Afterthoughts" In Themes from Kaplan. Ed. J. Almog, J. Perry, and H. Wettstein. New York: Oxford University Press, pp. 565-614.

Kripke, S. (1972), "Naming and Necessity" In Semantics of Natural Language. Ed. D. Davidson and G. Harman. Dordrecht: Reidel, pp. 253-355 and 763-9. Rpt. as Naming and Necessity. Cambridge: Harvard University Press, 1980.

Nagel, T., (1986) The View from Nowhere. Oxford: Clarendon Press.

Russell, B. (1910-11), "Knowledge by Acquaintance and Knowledge by Description" Aristotelian Society Proceedings 11, pp. 108-128.

Wilson, E. (1998), Consilience New York: Knopf.

                                                                                                                      Top of page