Distinction between thoughts and sensations
Perry introduces some useful distinctions and terminology for thinking about the relation between thoughts and sensations.
It's one thing to have a sensation; it's another thing to know about or think about that sensation (46).
In general, we can know about an object (i.e. an object is "epistemically accessible") only if the object has some sort of causal effect on us, for instance by affecting our visual, auditory, or olfactory experiences. Sensations are different: they seem to become epistemically accessible simply by virtue of being experienced (48).
Mental representations
individual mental representations: ideas (as opposed to compound mental representations: thoughts, beliefs, desires, etc.)
Two kinds of ideas: notions are ideas of individual things (particulars); concepts are ideas of universals such as properties and relations (50).
Ideas are sort of like file folders stuffed with information of various kinds (51). What an idea is of (what it represents) is determined by where the information in the file came from, not by the content of the information. For concreteness, Perry uses the example of a folder kept by an academic advisor for a student.
Four kinds of relations between idea (file folder) and object:
1. object is origin of the file (what led one to open
the file in the first place)
2. object is dominant source of information in the
folder (difference between 1 and 2: one is origin of file; 2 is origin of what's
in it)
3. object is denoted by content of information in
folder (i.e. the information correctly characterizes the object)
4. object is applicandum of folder (the think we use
the folder to guide our interactions with)
When things are going smoothly, the same object will have all four of these relations to the idea. But things don't always go smoothly, and they can come apart! Perry takes 1 to determine what the concept is "of," and this can end up being something different from the other three. (If information about the wrong thing gets into the folder, 2 and 1 are different. If inaccurate information gets into the folder, 3 and 1 are different. If one uses the folder to deal with the wrong object, 4 and 1 are different.)
Concepts of experiences
We have concepts of our experiences (53). These can contain all sorts of information about the experience: what it is caused by, what it tends to cause, etc. One crucial component of the concept is "demonstrative/recognitional" (54): we formed the concept as a result of having the sensation; we remember what it's like; and we can recognize later occurrences of the same sensation.
There are self-directed methods of knowing the subjective character of one's experiences (55). Some of these methods are subjective, in the sense that they are available only to the person having the experience.
Remembering and anticipating sensations is something like having them (57).
The three above features of concepts of experience constitute their "Humean core," which is likely to be central (58). This isn't to say that you can't have a concept of a sensation you haven't had: you can, but it's missing something that is central to the concepts of people who have had the experience. So for example a color-blind person can have a concept of the subjective sensation of red, but it will be missing an important feature of the concept of this sensation possessed by people who have the sensation.
Mental States as Physical States
If physicalism is to be a live option, our concepts of mental states such as sensations must be "topic-neutral" -- that is, it must be neutral on the issue of whether they are physical or not.
Two ways this could be true: (1) our understanding of mental states is functional. Then it's an open question whether the realizers of the functional roles are physical or not. Perry doesn't think this will work for sensations, though it may work for "psychological" mental states. (2) our understanding of mental states is demonstrative: pain is this experience.
Perry's terminology:
Perry's view is that although zombie worlds (in his sense) are possible, Chalmers zombie worlds aren't.
Zombies and Epiphenomenalism
Perry makes an interesting point: he suggests that the question of whether Chalmers zombie worlds are possible is really not the question of whether physicalism is true or false, but rather the question of whether epiphenomenalism is true or false. If you think epiphenomenalism is false, then you won't find zombie worlds possible, regardless of whether you're a dualist or a physicalist. On the other hand, if you think epiphenomenalism is true, then if you're a dualist you will think Chalmers zombie worlds are possible, and if you're a physicalist you'll think that (almost) Chalmers zombie worlds are possible.
He illustrates the point like this (table reproduced with small changes from p. 78):
| Epiphenomenalism | Efficacy of the Conscious | |
| Physicalism | Conscious states are physical nomological danglers |
Antecedent Physicalism |
| Dualism | Conscious states are nonphysical nomological danglers (Chalmers' view) | Commonsense (and Cartesian) dualism |
His point is that Chalmers zombie worlds (or (almost) Chalmers zombie worlds) are logically possible if you're in column 1, regardless of whether you're a physicalist or not, while the aren't logically possible if either view in column 2 is correct, regardless of whether you're a physicalist or not.
What kind of properties are Sensations?
Possibilities:
The identity theory says that sensations are properties of type B. Functionalism says they're properties of type C. Chalmers' property dualism says they're of type D.
Perry thinks that the zombie thought-experiment is intended to support D over C. (He seems to think that both C and D are in effect epiphenomenalistic. The argument that C is epiphenomenalistic would be essentially Kim's problem of causal exclusion.) He thinks it's ineffective against his own view, which is that sensations are of type B.
Last update:
November 4, 2009.
Curtis Brown | Philosophy of Mind | Philosophy Department
| Trinity University
cbrown@trinity.edu