It's worth highlighting that this premise can't be right, for we can talk about things that do not causally affect us. We can even talk about things that don't exist, like unicorns, or God. (Or does Eliezer think that 'God' refers to the religious module of the brain, so that God exists after all?)
I'm reminded of Hilary Putnam's a priori semantic response to the skeptical scenario that I might be a brain in a vat (BIV). The idea is that if I were really a BIV then my terms like 'brain' and 'vat' would instead refer to the objects in the hallucinated 'image' (i.e. in the phenomenal world). So, the argument goes, whatever the words end up meaning, 'I am a brain in a vat' is guaranteed to come out false. I can "know" that I'm in the fundamentally real world, just because it's (allegedly) impossible for my term 'fundamentally real world' to refer to anything other than the phenomenal world that I'm presented with.
Clearly, something has gone wrong in these arguments. The appropriate response, I think, is to say "Stop redefining my words!" I can understand the BIV scenario perfectly well, and it's a scenario I cannot rule out with absolute certainty; hence 'I am a BIV' is not guaranteed to be false at all. Putnam's argument to the contrary is a mere semantic trick, playing with words.
The same is true of Eliezer. We know perfectly well what we mean by the term 'phenomenal consciousness'. We most certainly do not just mean 'whatever fills the role of causing me to make such-and-such utterances'. By suggesting this, he is playing Humpty Dumpty, redefining words to mean whatever he wants them to mean. It's simply changing the subject. (I never claimed that 'whatever fills the role of causing me to make such-and-such utterances' is physically irreducible. So to argue against this claim is not to address my claim that 'consciousness is irreducible'. It's just like arguing against atheism by redefining 'God' to mean 'the universe', or some such silliness.)
Eliezer has recently repeated the mistake in arguing for reductionism about identity (a view I actually share, though for different reasons). He writes:
Whatever-it-is which makes me feel that I have a consciousness that continues through time, that whatever-it-is was physically potent enough to make me type this sentence. Should I try to make the phrase "consciousness continuing through time" refer to something that has nothing to do with the cause of my typing those selfsame words, I will have problems with the meaning of my arguments, not just their plausibility.
Whatever it is that makes me say, aloud, that I have a personal identity, a causally closed world physically identical to our own, has captured that source - if there is any source at all.
And we can proceed, again by an exactly analogous argument, to a Generalized Anti-Swapping Principle: Flicking a disconnected light switch shouldn't switch your personal identity, even though the motion of the switch has an in-principle detectable gravitational effect on your brain, because the switch flick can't disturb the true cause of your talking about "the experience of subjective continuity".
This is a terrible argument, for the simple reason that whatever it is that makes me say 'X', is not necessary what I mean by 'X'. (Again: unicorns, God, etc.)
Cf. Linguistic Anti-Paternalism: each individual is the final authority on what they are using their words to express. To offer a brief sketch of how this might work: We may have a particular idea in mind, a bunch of descriptive platitudes about what it is to be X (which need not invoke causal relations at all, though it often will). Arguably, this is what decides the meaning of 'X', though there's no guarantee that it will successfully refer to anything in the world; that instead depends on whether there is any worldly thing which satisfies the platitudes.
Anyway, that's just a very rough and inadequate sketch of how an alternative view might go. But the purpose of this post is not to give a fully-fledged theory of meaning. It is simply to highlight one adequacy constraint on any such theory: it must make it possible for us to talk about, e.g., abstract or non-existent things, i.e. things which are not themselves the cause of our talk about them. (It needn't make epiphenomenalism true, but it had better be expressible!)
Posts
Words have intensions and extensions. The intension of "unicorn" is "a horse with a single horn, attracted to virgins". The extension of "unicorn" is empty. Do you propose that the extension of "consciousness" is empty?
ReplyDeleteCausally tracing back an experience is one way a word can be defined, and it's very helpful when we don't understand the cause of an experience. Compare: "Fire: the true cause of my experience of that bright orange hot thingy" vs. "Fire: the release of phlogiston". The second definition would have an empty extension, if it weren't for the fact that you behave as if "fire" refers to the hot thingy, no matter what kind of verbal definition someone gives. E.g., someone in a crowded theatre shouts "fire", you don't wait around to test the phlogiston theory.
In the same way, consciousness is among other things the cause of philosophers talking about consciousness. I really think you're missing the point of that "among other things" clause. It means I'm not redefining consciousness, just making an obvious-seeming statement about what, among other things, consciousness does. This lets me trace back the cause of my word to find its meaning.
When you start defining "consciousness" intensionally in elaborate ways, you run into the triple problem of ill-formed intensions, possibly empty extensions, and the fact that you behave as if "consciousness" refers to something that makes you talk about an inner listener, no matter how you try to define the word otherwise.
I also don't buy your whole theory of meaning as an extra property that can be spraypainted onto brains. See The Simple Truth.
"The extension of "unicorn" is empty. Do you propose that the extension of "consciousness" is empty?"
ReplyDeleteNo, the example of unicorns is merely to show that we can talk about non-causally related things. One reason for this might be if the idea we have in mind fails to correspond to anything in reality (like unicorns). Another reason might be if the idea we have in mind corresponds to something which is not causally connected to our talk about it (e.g. epiphenomena). This is clearly possible, and so straightforwardly refutes any theory of language which claims otherwise.
"Causally tracing back an experience is one way a word can be defined"
I granted that some words are defined in terms of their causal or functional role. My point is simply that not all are, and phenomenal consciousness is an obvious case of this.
Actually, the sort of 'tracing back' you describe is probably at most a heuristic, and never a part of the ultimate story about meaning and reference-fixing. Consider a friendly example like 'fire'. The "platitudes" or reference-fixing description of fire might invoke (roughly) the fact that it gives off heat, glows orange, leaves behind charcoal, etc. (Obviously "the release of phlogiston" is no platitude of common meaning, but instead a theoretical attempt at explanation.) It turns out the thing in the world which best satisfies these platitudes is a certain chemical/physical process, and so that is what 'fire' refers to.
N.B. I can imagine a scenario in which "the true cause of my experience of that bright orange hot thingy" was mere hallucination, like the God module in the brain producing religious experiences. If it turns out that there's also the real phenomenon out in the world, but I've just never causally interacted with it, still my term 'fire', as I mean to use it, refers to the real phenomenon which satisfies the platitudes. It does not refer to the fire module, or "true cause" of fiery hallucinations in my brain.
So, again, your theory of reference here is plainly false.
N.B. I can imagine a scenario in which "the true cause of my experience of that bright orange hot thingy" was mere hallucination, like the God module in the brain producing religious experiences. If it turns out that there's also the real phenomenon out in the world, but I've just never causally interacted with it
ReplyDeleteFollow the improbability, as Eliezer would say. It would be hideously unlikely for you to hallucinate fire despite literally never having interacted with it - realistically, real fire would have to be way back in the causal chain somewhere, and the word would in fact refer to the cause the way you use it. (I know this is meant to be an argument in principle, but I feel the point is important.)
"I know this is meant to be an argument in principle, but I feel the point is important."
ReplyDeleteI'm not feeling it. A "hideously unlikely" counterexample is a counterexample all the same.
I didn't take Eliezer to be redefining words like Putnam in the BIV scenario at all. From the post you linked to:
ReplyDeleteThe word "consciousness", if it has any meaning at all, refers to that-which-is or that-which-causes or that-which-makes-me-say-I-have inward awareness.
Isn't that the mystery of consciousness? Where does this inward awareness comes from? Anyone can define any term however they want, but Eliezer seems only to be worried about consciousness insofar as it refers to a mysterious problem for scientists and philosophers, not to some mystical nonsense that people spout while using the term 'consciousness.'
Matt - Unfortunately, Eliezer was not speaking English in that quote you offer. He had already redefined 'inner awareness' and 'internal narrative', not to mean the phenomenal feel of your thoughts (or associated auditory qualia), but rather to mean the underlying neural processing that's detectable by brain scans and such.
ReplyDeleteNote that his argument only works if he assumes that 'consciousness' can only possibly refer to that which causes us to say "I have inner awareness". It isn't enough for him to refer to that which is inner awareness (i.e. consciousness), because that would leave open the possibility of epiphenomenalist claims that inner awareness is not itself the cause of our talk about it.
So, again, his argument can only work by redefining the terms. Unfortunately, he does this by stealth (sloppily moving between 'that-which-is', 'that-which-causes', and - what he finally settles on - 'that-which-makes-me-say'), misleading many readers into failing to realize this.