Antiphysicalists have the intuition that phenomenal concepts—the conceptions we have of phenomenal qualities or qualia—neither a priori imply nor are a priori entailed by physical or functional concepts. This is the intuition raised by the kind of examples offered by Jackson. Antiphysicalists attempt to use this intuition to show that phenomenal properties must be different from physical properties. Loar accepts the intuition that phenomenal concepts are irreducible, but denies that this entails that phenomenal properties are distinct from physical properties:
So if the antiphysicalist wants to use the version of the argument that doesn’t beg the question, she must use some additional premise. What might that premise be?
We may start by noting that the descriptions involved in the acquired information are contingent. Let’s assume that the concept of alcohol is something like “ingredient of wine and beer that makes you feel intoxicated” and that the concept of gold is something like “the stuff that causes this visual experience”. But it is only a contingent property of gold that it causes such and such visual experience, and only a contingent property of alcohol that it makes you feel intoxicated.
Contrast this with the properties involved in the knowledge argument. The property of feeling that way that we attribute to a particular visual experience is supposed to be essential to the visual experience. It’s not like someone could have had the same visual experience and felt it in a different way. What could have happened is that someone had seen the same object and had a different visual experience, but this need not worry us.
Defenders of the knowledge argument may claim that it is because Mary’s new knowledge is not just of a contingent property of the visual experience that the knowledge argument goes through. Loar states this premise thus:
Loar’s aim is to offer an account of concepts that doesn’t require us to accept the semantic premise, and so, that renders the knowledge argument ineffectual. In other words, he will try to show that a phenomenal concept can pick out a physical property directly (not via a contingent property of the physical property), and still be conceptually independent of physical concepts.
Most of his points rely on defending the view he presents from possible objections.
There are two parts to Loar’s account of how phenomenal concepts may pick or refer to physical properties. The first explains how phenomenal concepts directly refer to whatever properties they refer to. The second explains why phenomenal concepts are conceptually irreducible to physical concepts.
First, Loar claims that phenomenal concepts are “type-demonstratives that derive their reference from a first-person perspective: ‘that type of sensation’, ‘that feature of visual experience’.” (p. 295) If phenomenal concepts really are type demonstratives, this would partially give us reason to explain why they directly refer.
What is a demonstrative? Demonstratives are referential expressions like ‘that’, ‘this’ and the like. On some accounts of these terms, it is not required to appeal to any particular property in order to account for their reference, unlike with other expressions such as definite descriptions.
Suppose you are in a supermarket looking at different fruits and you say “That looks good”, an utterance you accompany by a demonstration with your finger. Some philosophers would say that the expression ‘that’, as you used it, refers to a particular fruit—a peach, let’s suppose. Moreover, it does so without any need to appeal to a particular property that individuates the fruit.
Compare with the case in which you say “The peach on the upper left corner that is touching a pear looks good”. In this case, some philosophers would say that you referred to exactly the same fruit as before, only this time by means of a definite description. This time, they would say, the reference was not direct. Instead, it was achieved by means of identifying the fruit that had the properties described by the definite description. In this case, the definite description referred to that particular peach because that peach was the only thing in the context that was a peach, was in the relevant corner and was touching a pear.
Some people think that when someone refers to something by description, as in the second case, and moreover, when the description appeals to essential properties of the object, then unless the two descriptions are conceptually the same or reducible to each other, one is committed to the claim that the objects so referred are different.
In a way, Loar doesn’t attack this claim. Instead, he thinks it’s not relevant to the issue at hand, since phenomenal concepts refer to phenomenal properties (which, he thinks, are themselves physical properties) directly, we don’t need to care about the properties used to refer, or whether they are contingent or essential. We just don’t use any properties in order to refer!
The second part of his account consists in explaining why phenomenal concepts are irreducible to physical concepts, in a way compatible with the claim that phenomenal concepts directly refer. Loar claims that phenomenal concepts are of the kind we may call recognitional.
In general, recognitional concepts are the kind that we learn by how the referred things look, sound, etc. For instance, some philosophers think that color concepts are recognitional: it’s not like we have some sort of description of the color that we then use to categorize things by color. For instance, it’s not like we have a description of yellow and then we go out and see which objects satisfy the description. Instead, we recognize yellow things because they look a particular way. We also recognize shapes because of how they look, or pieces of music because of how it sounds.
Recognitional concepts and theoretical concepts are in general independent from each other. The reason, Loar thinks, is something like this:
Concepts of the two sorts have quite different conceptual roles [...] Basic recognitional abilities do not depend on or get triggered by conscious scientific analysis. If phenomenal concepts reflect basic recognitions of internal physical-functional states, they should be conceptually independent of theoretical physical-functional descriptions. That is what you expect quite apart from issues concerning physicalism
If phenomenal concepts are recognitional concepts, this would allow us to explain why they are irreducible to physical concepts. If they are direct demonstratives, they would allow us to explain how they could refer to physical things even though they are conceptually irreducible to physical concepts.
Usually, we find out which identity statements are true by means of something like an a priori procedure. How, for instance, may we identify liquidity with a particular physical property? We start with the assumption that there is a certain functional characterization of the concept of liquidity. Then we show that some physical theory gives that particular role to a certain physical property. Discovering that physical property is an empirical process, but the reduction itself is a priori: the identity between a given physical property and the property of liquidity is entailed by the definition of liquidity in conjunction with the physical theory in question.
Unfortunately, if phenomenal concepts are irreducible to physical concepts, this strategy is not available. The strategy requires at least that the phenomenal concepts be functionally described, but it doesn’t look like we can do this with phenomenal concepts.
Instead of trying to solve the problem, Loar attempts to explain why this explanatory gap shouldn’t worry us. One of the virtues of his account, he says, is that it can explain why there is an explanatory gap in a way that is compatible with physicalism:
[T]he idea that phenomenal concepts are recognitional concepts of a certain sort does account for the explanatory gap in a way compatible with physicalism. Phenomenal concepts, as we have seen, do not conceive their reference via contingent modes of presentation. And so they can be counted as conceiving phenomenal qualities directly.
What generates the problem is not appreciating that there can be two conceptually independent “direct grasps” of a single essence, that is, grasping it demonstratively by experiencing it, and grasping it in theoretical terms. The illusion is of expected transparency: a direct grasp of a property ought to reveal how it is internally constituted, and if it is not revealed as physically constituted, then it is not so. The mistake is the thought that a direct grasp of essence ought to be a transparent grasp, and it is a natural enough expectation. (p. 305)
1Perhaps closer to the original formulation of the argument would be: facts about CH3CH2OH do not exhaust facts about alcohol.