Notes on John Searle, The
Rediscovery of the Mind
Searle argues that unconscious mental states must be potentially conscious. (The "connection principle.")
My comments are indented (and in blue, if you're reading this on the web).
Here's the big picture of why this is important.
One point Searle has been arguing for is closely related to arguments for property dualism that we have previously considered: conscious experience is not reducible to more basic properties. On this point he agrees with dualists like Chalmers. However, it's common for people who agree with Searle about the irreducibility of (phenomenal) consciousness to think that nevertheless, intentional or representational states such as beliefs and desires, and reasoning or problem solving involving such states, are reducible (to biological or functional properties). This is closely related to Chalmers' distinction between the easy problem(s) and the hard problem: explaining intentionality is (closely related to or part of ) the easy problem; explaining (phenomenal) consciousness is the hard problem.
Searle now wants to take a considerable further step. Searle argues that if the hard problem cannot be solved, then neither can the easy problem, because intentionality requires consciousness.
|consciousness is reducible||consciousness is not reducible|
|intentional states and processes are reducible||physicalism: identity theory, functionalism, etc.||Chalmers' property dualism; for Searle this is not a coherent view, because a state can't be intentional without being (potentially) conscious|
|intentional states and processes are not reducible||???||Searle|
The argument for the connection thesis, as Searle presents it (in chapter 7, section II):
1. distinction between intrinsic and as-if intentionality. There must be such a distinction because if the only intentionality is as-if intentionality, then everything is intentional.
This is so only, it seems to me, if we stick with a pretty primitive idea of as-if intentionality. Searle's label helps with this: makes it seem as though anything we could pretend is intentional would really be intentional. But a more sophisticated version of the view might insist that further conditions must be met. For instance, not just that we can pretend that something is intentional, but that doing so is actually more useful than alternative ways of regarding it. (Regarding water flowing downhill as trying to reach the bottom is no more useful in prediction than regarding it as obeying the law of gravity. Regarding water as calculating the size of rocks etc. seems actually less useful than just assuming that it obeys basic physical laws. Searle seems to make no effort whatsoever to reply to more discriminating versions of the view that there is no in principle distinction between intrinsic intentionality and some version of Dennett's view that something is intentional if it is predictively useful to attribute intentionality to it.
Searle treats any naturalistic account of intentionality as really only an account of "as-if" intentionality. But that term seems appropriate mainly for Dennett's view. An alternative naturalistic view: a physical structure represents an abstract structure iff (very very roughly!) the physical structure has states corresponding to the abstract states of the abstract structure, and the causal relations between the physical states mirror the logical relations between the abstract states. On this sort of view, water can't calculate the size of a rock unless the water has recurring states that can be regarded as representing rocks, physical magnitudes, mathematical relations, etc.
2. Unconscious intentional states are intrinsic [i.e. have intrinsic rather than as-if intentionality].
OK, if the distinction makes sense, but see response to 1.
3. Intrinsic intentional states always have "aspectual shape." (Other authors have used the Fregean term "mode of presentation" for this notion. Another related idea is the idea that we think about things only "under a description.") Roughly speaking, I never think about something, period; I always think about it in some particular way, that is, I always think about it as something. For instance, I might think of water as water rather than as H2O.
4. Aspectual shape cannot be completely characterized in terms of third-person predicates, including behavioral or neurophysiological predicates.
Why not? Consider behavior. S claims that there is no sure behavioral test that will determine whether someone is thinking of water as water or as H2O, since the two are always found together.
Well, you can almost see what he has in mind here. That someone tries to get some water whenever he is near it is good behavioral evidence that the person wants water, but you can't try to get water without trying to get H2O, so it seems to be equally good behavioral evidence that the person wants H2O.
But come on! If there is any difference between thinking of water in these two ways, mustn't there be some way this difference could be behaviorally manifested? For instance, if someone gives me a glass of stuff and says "here's the stuff you wanted," if I was thinking of what I wanted as water, I may check the glass to see whether the substance it contains is a colorless, odorless liquid; if I was thinking of it as H2O, I may try to see whether I can convert it to hydrogen and oxygen.
Searle considers another kind of behavioral evidence: how the subject responds to verbal questions involving the words 'water' and 'H2O'. But, he says, "there is no way just from the behavior to determine whether the person means by 'H2O' what I mean by 'H2O' and whether the person means by 'water' what I mean by 'water'" (158). But he doesn't argue for this, as far as I can see, and offhand it doesn't seem that it would be all that difficult! I ask the subject, "does the term 'H2O' logically imply that a molecule of what it refers to contains two atoms of hydrogen and one atom of oxygen?'" If they say "yes," that's a pretty good sign that by 'H2O' they mean 'H2O'.
Searle appeals to Quine's thesis of indeterminacy of translation a bit later on (163-64). But even Quine wouldn't say that it's indeterminate whether to translate someone's term 'H2O' as 'H2O' or 'water'! His examples are much more esoteric (e.g. 'water' vs. 'undetached water part', 'instance of wateriness', etc.).
5. The ontology of unconscious mental states is purely neurophysiological. (At least, as long as they are unconscious.)
So, as long as a mental state is unconscious, there is no way to determine its aspectual shape. The only way we can meaningfully speak of it as having an aspectual shape (and therefore the only way we can regard it as a mental state at all) is if we speak of the aspectual shape it would have if it were conscious. And this in turn is only meaningful if it is possible for it to become conscious. So,
6. The notion of an unconscious intentional state is the notion of a state that is a possible conscious thought or experience.
7. "The ontology of the unconscious consists in objective features of the brain capable of causing subjective conscious thoughts."
Note the term "causing" here. I'm not sure why S calls unconscious mental states "mental states" at all, actually, since his view seems to be not that they are mental states but that they are capable of causing mental states. In his discussion of Freud, he seems to think that this is just a verbal quibble, but I'm not so sure. On his view an intentional state has to have aspectual shape. But also on his view unconscious mental states don't really have an aspectual shape at all, they are just capable of causing something else that does have an aspectual shape.
(A digression on micro-macro causation and ontological reduction. Note p. 162: "characteristically in the sciences we define surface phenomena in terms of their micro causes; we can define colors in terms of wavelengths . . .": this suggests that he does think that even in contemporary science macro properties are caused by micro properties. ?? Earlier it seemed that ontological reduction was incompatible with causal reduction. So it looked like S thought that if micro properties cause macro properties, then the macro properties cannot be (ontologically) reduced to the micro properties. Then it turns out that he thinks reduction is possible in the sciences after all. But this is because we "carve off" the subjective features of the macro properties, leaving only ontologically reducible stuff. I took this to mean that the only macro properties that were caused by micro properties were properties of subjective experience, so that when these are carved off, only ontologically reducible stuff is left. But in that case color, heat, etc. as currently understood, i.e. as properties of the objective world, not features of our experience, would not be caused by micro level properties. So I'm puzzled about exactly what S's position is here.)