The Knowledge Argument Critiqued

Could Love be like a Heatwave?  (1986)

By Janet Levin

In Heil, pp. 539-52

I.  Introduction:  Levin’s goal

A. Nagel

1. argues that no objective knowledge of bats could tell us what it is like to be a bat

2. concludes that there are facts about the subjective character of experience that no physicalist, functionalist, or other objective theory of the mind could describe

B. Jackson

1. although Mary had all the physical information, her knowledge is incomplete

2. so there is more than physical information and physicalism is false

C. Philosophers (viz., Nemirow) have objected that both arguments contain an equivocation

1. knowledge of what it’s like to be a bat or to recognize colors is practical knowledge or an ability (539, q.v.)

2. a lack of such an ability is not the same thing as not having some factual or theoretical knowledge (q.v.)

a. not clear that there is any link between the theoretical and the practical, such that a knowledge of mental states should lead to these abilities (539-40)

b. in fact, Jackson himself makes this argument against Nagel (540n4)

c. hence, even though Mary does not know what colors are like and we don’t know what it’s like to be a bat, it doesn’t follow that there are some facts about experience that no objective theory can describe (540, q.v.)

D. but in spite of these objections, it nevertheless still seems plausible that without the ability to have certain kinds of experience, one cannot have knowledge of the facts about those sorts of experience (q.v.)

1. it would be “perverse” to argue that experience provides only practical abilities

2. experience could be the only possible source for some facts, such as the taste of a pineapple

3. there seem to be cognitive differences between us and those who cannot have the same experiences (540)

E. thus, a physicalist must still find some way to argue against the plausibility of the empiricist claim that one has to have certain sorts of experiences to have all the facts about mental states (540-41, q.v.)

F. Levin’s goal is to do just that (541)

1. she thinks that the plausibility of this claim derives from an argument that is more compelling than Jackson’s or Nagel’s, that is Richard Warner’s argument

2. this argument is based on the following premises (write on board):

a. if one lacks certain experiences, one lacks a certain discriminative or recognitional ability, which she calls “direct”

b. this ability to recognize or discriminate among mental states is needed for having full factual knowledge of them

2. she will argue that this argument relies on an ambiguity in the notion of a “direct recognitional capacity” (541-42)

a. neither meaning will make both premises true

b. source of the ambiguity is a failure to distinguish:  (542)

1.) having a concept

2.) having the ability to apply the concept

c. once this distinction is drawn, we can see that it’s the latter that objective descriptions fail to supply to someone who lacks the relevant experience

d. but this ability is not needed to have a complete knowledge of the facts about these experiences

3. the problem with the argument is not that it equivocates between having an ability and having some knowledge, but that it equivocates between two kinds of abilities, only one of which is required for having knowledge of all the facts in question

a. as we shall see below, she means an ability to apply concepts vs. an ability to answer questions

b. this distinction has been overlooked because of an implicit acceptance of an empiricist theory of concepts

4. once the physicalist has made this distinction, he or she will be in a better position to explain what makes our knowledge of colors different from Mary’s, etc.

II. In order to examine the notion of a direct recognitional ability, will consider Molyneux’s Question:  could a man born blind who was then made to see determine by sight alone which of two things was a sphere and which a cube?  (542)

A. Molyneux, Locke, and Berkeley all said no

1. for Molyneux, this constitutes proof that the idea of visible shape or contour could be acquired only by sight, and not by touch, reasoning, etc.

2. Molyneux thus took a person’s discriminative or recognitional ability as the test of his knowledge of the facts (542-43)

B. Suppose these philosophers were right that the man born blind would fail the test.  What does that tell us about his theoretical knowledge?  (543)

1. an empiricist philosopher would take it that this was evidence that the man lacked the ideas of the visual properties of cubes and spheres, because if he had “faint copies” of these properties in his mind, he’d be able to match them to what he’s currently seeing

2. but on some other view of concept formation, it is not clear why the failure of the recognition task tells us anything interesting about his conceptual knowledge

C. suppose this guy, sort of like Jackson’s Mary the neuroscientist, was all-knowing, but about the objective facts about geometry and psychology. 

1. He might then be able to answer questions about spheres and cubes correctly (543)

2. suppose further that he’d first been shown some other shapes and was told what they were, and then was able to go on and correctly identify the cube and the sphere. 

a. for Levin, this shows that what he was previously missing was not some facts or concepts, but the ability to apply to new experiences the concepts he already had

b. he just needed a little practice

D. on this interpretation of “direct recognitional ability” (having the ability to apply concepts), the man born blind would have all the factual knowledge, without having the relevant recognitional capacities

1. this would make the first premise in the argument from part I (F.2.a) plausible (544; cf. p. 541)

2. but the second one false (I.F.2.b)

E. However, there’s another way to interpret this same thought experiment

1. again, the man born blind is shown some other shapes, and then correctly identifies the cubes and spheres, by sight, without having touched them

2. here we might say that the recognitional ability can be taken as a test for having the concept of a sphere or cube (544)

a. if he fails the test, then it seems he was lacking some concept that only the relevant visual experience could supply.

b. But if he passes it, then the man born blind would have a certain concept without having had the certain sort of experience

c. that is, on this interpretation, the second premise is plausible but the first one is false

F. Could someone born blind pass this test? 

1. today, Molyneux’s question is no longer just a thought experiment

2. results depend on many variables, including how good an education the person had while blind

3. so there is empirical support for the view that theoretical knowledge can give you recognitional abilities

G. two problems:

1. that if the man born blind does recognize the sphere and the cube, it is by inference, and not directly.  (544-45)

a. Then Levin’s version of this example would not be a counter-example to the claim that recognitional abilities depend on prior experiences of a certain sort (545)

b. so perhaps we need a different sort of example

2. there’s another reason, too, that we need a better example: 

a. the ideas of spheres and cubes are rather complex, unlike colors, tastes, pains, such that there could be features common to the visual and the tactile experiences of each

b. or, in other words, for each shape, the ideas of them that we get from sight and touch are not completely different

III. Warner’s Alien

A. same issues can arise for other sensations

B. Warner argues that neither functionalism nor physicalism can capture all the facts about the experience of pain

1. imagine an all-knowing Alpha Centaurion who had never had an unpleasant experience

2. now imagine he (it?) invents a device that allows him to feel a stomach cramp for the first time (545-46)

3. but an unforeseen result was that this device also makes him feel nauseous (546)

4. Warner argues that this alien would not be able to distinguish which feeling was the pain and which was the nausea without making inferences or consulting additional instruments (546)

C. but Levin thinks that the same sorts of ambiguities in the notion of “direct recognition” arise for this case, too

1. first of all, it’s hard to see why the alien would not be able to distinguish cramps from nausea.  If he knows the functional difference between them, he knows their causes and effects

2. one might object that the alien is then only drawing an inference and not making a “direct” identification

3. but to raise this objection is to beg the question against those who would define mental states in terms of their functional or other objective properties

a. the question was whether these concepts were sufficient for making the identification

b. to conclude that these concepts were not sufficient to this task, one has to have some way of showing that the alien lacked the recognitional ability other than simply saying he had to draw an inference

c. furthermore, even we humans may make inferences in sorting one pain from another (546-47)

D. hence, it looks like we have another case where we can have recognition without prior experiences of a particular sort (547)

E. but this could just be another bad example. 

1. Nagel’s and Jackson’s arguments are about experiences very different from ones we’ve had

2. but even if the alien had never felt pains before, he might be able to relate them to other sensations he’s had in his stomach

F. but let’s say he’s never had any sensations in his stomach.  Then he might very well fail to distinguish cramps from nausea.

1. but it’s not clear why this failure should tell us that there’s a gap in his knowledge of the facts about stomach pain and nausea

2. if he could correctly answer questions about pains and nausea, and if, after feeling certain other sorts of pains, he could correctly identify these, we’d say that what he had been missing was not certain concepts, but the ability to apply them

3. thus, just as in the Molyneux case, in the sense in which “direct recognitional ability” is needed for factual knowledge, it wouldn’t depend on experiences of a certain sort, and vice versa (547)

G. Similarly, we can say that what Jackson’s Mary lacked was not the color concepts, but the ability to apply them

1. Mary will have the relevant color concepts if she can answer questions about them (547-48, q.v.)

2. also, if she can go on to identify other colors, or shades of them, after having seen a few (548)

H. but in the bat case, we not only lack the ability to apply concepts of sonar perception to experience, but we lack the concepts

1. however, there is no reason to conclude that one must have some specific set of experiences in order to acquire those concepts

2. all sorts of different experiences could contribute to forming such concepts

a. hence, it could be helpful to tell a blind person that red is like the sound of a trumpet, if he’s told what other colors are like as well

b. or to tell a prepubescent that love is like a heat wave

3. Levin suspects that what makes the bat case difficult is our lack of objective knowledge about bat experience

IV. Facts and Abilities (548)

A. So far, she’s argued

1. that a failure to identify a mental state could be due to either a lack of knowledge or to a lack of an ability to apply certain concepts

2. the lack of this ability does not imply a lack of factual knowledge

B. but one might think that it’s impossible to draw this distinction between having a concept of a mental state and being able to apply it (548-49)

1. it may seem that the only alternative to the Empiricist’s faint copy theory of concepts is a theory that identifies having concepts of mental states with the ability to make distinctions among them (549)

2. even Levin herself seems to be committed to a necessary connection between factual knowledge and recognitional abilities, as she has argued that someone with complete factual knowledge of a kind of mental state would be able to recognize new instances of them with the right sort of priming

C. however, she thinks that such a person would develop these recognitional abilities quickly only under the following conditions (write on board):

1. the experiences in each field can be individuated by an objective description (549)

2. the dimensions along which they are distinguished are perceptually salient for human beings

D. failures on the part of the highly intelligent, primed subject may be traced to one of these conditions

E. if the second condition did not hold, then someone’s failure to recognize a mental state wouldn’t imply a lack of factual knowledge and thus a problem for physicalism

F. if the first condition did not hold, on the other hand, then there would be a gap in the person’s conceptual knowledge

1. but if a physicalistic theory does not provide a way of objectively distinguishing mental states, it has failed as a physical theory (549-50, q.v.)

2. Nagel’s and Jackson’s arguments, however, tried to show that even a physical theory that was able to objectively discriminate among our mental states would leave out some facts about mental states, that is, what it is like to have them (550)

3. hence, the only cases in which someone’s failure to recognize a mental state would call a physicalist theory into question are those in which Nagel’s and Jackson’s arguments are completely irrelevant

V.  But is there a necessary connection between concepts and abilities?  (550)

A. for Levin, the connection is not necessary, but only epistemic: 

1. that is, the recognitional ability simply provides evidence that someone has mastered the relevant concepts

2. and the inability to recognize or discriminate shows a gap in one’s knowledge

B. but this is true not only for our knowledge of mental states. 

1. For example, we would expect an expert on dogs to be able to recognize different sorts of dogs, or a physicist to recognize the track of an electron in a bubble chamber

2. our concepts of dog and electron are tied in with a network of other concepts, so that even if a person lacked the specific recognitional abilities associated with their use, there would be some roundabout way of showing that he understood the concepts

C. Levin suggests that perhaps because our concepts of mental states are not as networked as our concepts of dog or electron, that we rely more heavily on recognitional ability as evidence that people have mastered the concepts.  Hence, we are tempted to say that the recognitional abilities are necessary to the concepts (550-51)

D. however, she thinks the difference between our concepts of mental states and our concepts of dog or electron is only one of degree (551)

1. hence, recognitional abilities are not the only kind of evidence that we have for mastery of concepts

a. to be sure, sometimes a recognition failure is evidence of a lack of conceptual knowledge, as in the case of the man born blind who thought the quarter moon should be wedge-shaped

b. but Warner’s alien and Jackson’s Mary were described as being highly knowledgeable

2. she sees no reason to think that any failures to recognize a kind of experience should be taken more seriously than failures by the dog expert or the physicist

E. one could grant all this and still insist that there is some special sort of contribution that experience makes to knowledge, a contribution that could not be made in any other way

F. Levin argues that it is not that experience contributes some piece of knowledge that could not be obtained in any other way, but that it provides it in the most efficient way (551)

1. the function of experience is both causal and evidential

a. it’s unlikely that one would know about the causes, effects, and similarity relationships regarding a kind of experience unless one has had it

b. having had the experience is the best guarantee that one knows all there is to know about it

2. however, it is not necessary to have the experience to obtain this knowledge.  One could instead go through the laborious procedures illustrated by Mary and the other examples to obtain it

G. and again, this applies not only to knowledge of mental states but to all sorts of knowledge:  having the experience is the best or most efficient way to acquire the knowledge.  E.g., what it’s like to be a parent (551-52)

H. the argument that there is some necessary connection between experience and knowledge of mental states:  (552)

1. ignores the generality of this connection.  It’s not just about mental states

2. also takes the connection too seriously: 

a. our expectations about what sort of experiences a knowledgeable person has have little to do with whether this connection is a necessary connection (552)

b. thus these expectations are not a threat to physicalism, functionalism, or any other objective theory of mind

I. she concludes that it is not the insufficiency of the objective point of view, but the vestiges of empiricism, that suggest that objective theories of the mind are inadequate (q.v.)