I’ve been reading over Chalmers’ conceivability/possibility arguments against materialist theories of mind. For those that don’t remember, his argument is:
1. We can conceive of zombies.
2. Conceivability implies logical possibility.
3. Therefore, zombies are logically possible.
4. If zombies are logically possible, then physicalism is false.
5. Therefore, physicalism is false.
In trying to give a more charitable reading of his argument, I’d like to find instances where such arguments have been successful in uncontroversial cases. We expect some in math or logic, but even there it isn’t clear that conceivability arguments aren’t just treated as hunches or invitations to explore something further. In my experience in mathematics, anyway, that’s how it has been, they certainly aren’t taken as conclusion-establishing game-stoppers.
We also have the related issue of thought experiments, which certainly have been useful for clarifying conceptual points. Perhaps some thought experiments are instances of a conceivability argument succeeding? There are many instances of conceivability arguments falling on their face, the Don Quixotes of hindsight (e.g., arguments for aether, vitalism, geocentrism).
Such failures are easy to find. What I’m after is conceivability’s Knight in Shining Armor, a single clear historical precedent that suggests it has done real work in the history of western thought.
This may be helpful if you haven’t seen it:
https://consc.net/papers/conceivability.html (Does Conceivability Entail Possibility)
“It is very implausible that conceivability entails physical or natural possibility”, but I don’t know if I understand how it he makes it work for metaphysical possibility.
I’ve seen that paper, one reason I posted is b/c I was hoping it would
have a shining example of success of this strategy in uncontroversial
cases but I didn’t remember seeing any. It’s possible that I missed it.
Again, a single historical precedent showing where this has been a
useful argument strategy for some real problem. Is there one I missed in that paper?
I remember material on the Goldbach Conjecture, but that isn’t such an
example of such a success story where something was actually settled by
these arguments, but more the ‘invitation to proof’ I discussed in my
post.
There must be something!
Schrödinger’s cat? The Turing machine (not the Turing test)?
I thought of Schrodinger’s cat, but it didn’t really establish anything. It’s a useful tool for testing and expressing one’s intuitions, but doesn’t seem to be a conceivability argument, at least not one that gave a knockout punch to macroindeterminacy. Most thought experiments in physics that I am aware of are not conceivability arguments. E.g., EPR, the rotating bucket, shining a flashlight on a train, Maxwell’s demon, twin paradox, etc..
Not sure at all what you are getting at with Turing Machines…
Gettier cases?
Hmmm. Not sure. First reaction is that it’s a thought experiment that is not only (logically) conceivable but likely nomically conceivable and likely nomically possible. Similar to Twin Earth, again seems less a conceivability argument than a good old fashioned thought experiment (as if they are mutually exclusive).
Which brings up the question what exactly do I mean by a ‘conceivability argument’? One necessary condition is that it is any argument that relies on premise 2 (conceivability implies possibility)? That’s one of the obvious weak links in the argument, so one I would want a good case study for, a historical poster child, in a less controversial field.
That may not be sufficient, though, so I’d be curious about other hallmarks that the philosophers here would say are necessary to count as a ‘conceivability argument.’
One reason I think premise 2 isn’t sufficient is because, colloquially, there are arguments to the effect ‘I cannot conceive of X; therefore, not X’. Which I would also call conceivability arguments. That is not Chalmers’ conceivability argument, at least as stated in the original post. And many people act as if that is his argument (the Huxely style ‘I can’t imagine how brains produce this, therefore brains don’t produce it’ (though admittedly you do pretty much find this intuition expressed as a cardinal assumption of the entire project, so it’s not as if the ‘hard problem intuition’, which is required to push throgh premise 1, is immune to this sort of criticism)).
As far as I can see, the example you are looking for is Hume’s so-called argument against induction (or at least one of the standard interpretations of it). All samples of copper tested so far have been observed to be good electrical conductors. However, it is conceivable that some as-of-yet-untested sample of copper is an electrical insulator. Therefore, it is possible that it is not the case that all samples of copper are good electrical conductors. This is not a completely uncontroversial case (nomic necessitarians like me think it cuts no ice) but it seems to me to be clearly a case of a conceivability argument most analytic philosophers currently accept. Personally, I would venture to claim that (this reading of) Hume’s argument is to be blamed the grip this conceivability-entails-possibility business has on analytic philosophers.
I should argue that, even if we buy into this way of reconstructing Hume’s argument, Hume would seem to have intended ‘possible’ as what today we would call ‘epistemically possible’ or ‘logically possible’ not ‘metaphysically possible’.
Interesting case.
Corollary question: are there instances of arguments for the logical possibility of X that did not rely on conceivability at some level?
Let’s take a noncontroversial claim: it is logically possible that stones of different weights could fall at different speeds in a vacuum. How do I know this? Is it because I can conceive it? At some level, of course, because conceiving/believing/thinking that X is a necessary condition of knowing X.
But being able to think something is the case is obviously not sufficient to establish its logical possibility. I can conceive that 0.99999 (repeating 9’s) is not equal to 1.0, but we know that the two numbers are actually equal. Such cases are why Chalmers has to say that the sense of conceivability is ‘ideal positive conceivability’. Such mathematical cases are instances of mere prima facie conceivability, which is more subject to error.
So three questions: is ideal positive conceivability really a good guide (can we know when we are in such a state); second, is Hume’s argument an argument that relied upon ideal positive conceivability (did Hume explicitly or implicitly use a conceivability argument); third, what did Hume’s argument actually establish? Is it a good poster child of conceivability argument success?
Selenium is an electric conductor in one arrangement of atoms (allotrope) but not in another. Therefore, it seems conceivable that an as yet undiscovered allotrope of copper may not conduct electricity (not a physicist so don’t quote me). Upon hypothetical discovery of such an allotrope, is that the sense in which this conceivability argument would be considered a success?
Interesting example. Problem is that it seems another case of an invitation to look around more, not conclusive. The conceivability argument isn’t what showed us about the allotropes. That took a lot of extra work.
How about Putnam’s perfect-actor argument? Here is a blurb about it from Daniel Stoljar’s paper “Actors and Zombies”:
“For the conceivability argument we are concerned with is in important respects analogous to arguments that are used and accepted throughout philosophy, and in philosophy of mind in particular. For example, consider a very different argument of Putnam’s (1965): the perfect actor objection to (philosophical) behaviorism. Perfect actors are people that behave actually and potentially exactly like ordinary people but have quite different phenomenal states. It seems conceivable, and so possible, that there are such people. And, if this is possible, behaviorism is false, for behaviorism entails that behavioral truths entail the psychological truths. It is standard practice in philosophy of mind to assume that this sort of argument is successful—a standard practice I assume is perfectly legitimate. But it is bad form to use a method of argument against theories you don’t like, and then turn hypercritical when the same method is deployed against theories you do.”
I am fascinated by his claim that this is an argument strategy that is standard practice. Part of me is sympathetic to this: it’s a huge part of what philosophy does after all. Go through different scenarios to get a feel for our intuitions, enlarge our conceptual sphere, and such. To do philosophy is to think, to conceive, to push our thought into new corners it has never been in before. This is done with thought experiments, and coming up with seemingly deviant cases to help clean up the conceptual landscape.
On the other hand, this is also philosophy’s weakness, explains its lack of strong “results”, because every substantive debate ends up a dialectic (e.g., realism vs antirealism) with no clean resolution.
So in some sense I see Chalmers’ argument as a test of old school analytical philosophy. But I am already biased that analytical philosophy has largely been a failure, not contributing much of substance to our knowledge of the world, based on distinctions that Quine and others have taken down that tried to insulate it from evidence and science.
Now the perfect actor argument might work against analytical behaviorism, just as Chalmers’ argument could arguably work against analytical naturalism (our everyday talk about phenomenal states can be translated, or is synonymous with, talk about neuronal states or whatever). But who the hell holds to these old-school analytical views of things? These days its all about accepting the meanings are divergent (this goes back even to Smart) and going from there.
So in a sense, these semantic arguments might work against other semantic arguments (that’s why I think Frankish’s argument is so great–turn Chalmers’ argument on its head, and the two annihilate each other). But as game-stopping substantive arguments about how the world is structured, I am extremely skeptical.
But regardless, perhaps something like Putnam’s argument against analytical behaviorism is a good case study. I’m not sure I like that one in particular (the behaviorist would quibble that the actor has a different history than the normal subject, so of course they can account for that easily within their framework, and he has not produced the right supervenience base for claims about mind—in their theory it must include the history of reinforcement of the individual!).
But something like this seems along interesting lines, as a possible success story for conceivability arguments. Something within analytic philosophy seems most promising.
Selenium is a non-metal and a semi-conductor. Copper’s conductivity is due to the “looseness” of its outer electrons. I don’t see how there could be a sample of copper that is an electrical insulator (although I see how one could take themselves to be conceiving of such a scenario).
Sorry, I’m writing this quickly so I hope it will make sense.
I think that your fist question is really two questions (is IPCing that p good evidence for believing that possibly, p? and can we know we are IPCing that p?) but I can see why you are putting them together. If I recall Chalmers’ definition of IPC correctly I would say that IPCing that p may well be good evidence for believing that possibly p if we only could know we are but we have no way for us to know that we are IPCing that p.
As for the second question, I’m no Hume scholar but it seems to me that there are at least two versions of Humes argument and the argument is very complex but does not employ an appeal to conceivability (let alone IPC). I think it’s the way Hume’s argument(s) has/have been interpreted/reconstructed/adapted by contemporary analytic philosophers that employs some sort of conceivability claim. (Those same “Humean” philosophers think something along those lines establishes the contingency of the laws of nature.)
As for your third question, I think that the philosophers think that Hume’s argument is an ideal poster child of the success of conceivability arguments.
the Turing machine was a conceivability argument or thought experiment when Turing wrote about it, but now is embodied in common tech. Except for the infinite memory part.
If you want examples in philosophy of mind I don’t know any though.
I want examples in any discipline.
Still not sure how the Turing machine when introduced was a conceivability argument. It is a mathematical model of a computer, right? Is a finite state automaton a conceivability argument? I don’t see it yet.
Oddly enough, the way you state the conceivability argument reminded me of Anselm’s ontolological argument, namely:
1) I can conceive of the greatest being
2) A being that exists in reality is greater than one that exists in reality alone
3) Therefore, the greatest being must exist.
4) We know that being to be God
If the similarity holds, it doesn’t necessarily help out conceivability arguments, though.
frankly, I stop reading at (1), I was taught from an early age (maybe not coincidentally when positivism still ruled the world!), the conceivability – intuition – is entirely worthless and cannot possibly anchor any argument.
I conceived of lunch, and sure enough, an hour later I had lunch – is not an argument for necessity or existence. I conceived of riding a unicorn home after lunch, and that didn’t happen. The only response to that is, “oh, then you didn’t really conceive it, because only true things are conceivable”, but that simply asserts the conclusion.
As for Dr. Chalmers, all I can say is, whatever dude. I could say it in detail in answer to his abuse of the concept, but he is hardly alone in appealing to intuition, especially. I just can’t really understand how anybody can take it seriously in this day and age. I cannot conceive of it. Therefore, perhaps it does not exist, hmm?
I don’t think it is all that awful to appeal to conceivability in all cases, especially where terms are relatively simple, well-defined, and uncontentious. E.g., if this, then that follows, therefore this theory has problems. Basic counterfactual reasoning seems to depend on it.
It doesn’t seem to me like there could be any powerful conceivability arguments. I don’t see arguments of that type as anything more than “according to my intuitions, X seems plausible”.
I don’t see counterfactual reasoning as any more dependent on conceivability than forward reasoning.
So what about that then, is “Socrates was a man” a conceivability issue? What *does* differentiate “I conceive of lunch” from “I conceive of unicorns”?
I suggest that the answer is something like Quine’s “web of belief”, “our statements about the external world face the tribunal of sense experience not individually, but only as a corporate body”. What that says is the claim of conceivable(X) carries no more weight than p(X).
I have come across this conversation a few weeks late. For what it’s worth, though, I concur with Gabriele that Hume’s critique of induction is a conceivability argument. Hume would have accepted a version of premise 2 (that conceivability implies logical possibility). According to Hume, logic was just about the relations of ideas. If there was nothing in the idea of the brain and the idea of non-consciousness that made them incompatible, then a non-conscious brain must be logically possible. Yet this would not yield any metaphysical conclusions. Hume importantly does think that induction is metaphysically possible, it just can’t be a matter of logic. So perhaps zombies are impossible; but if so, it can’t be merely a matter of logic.
Hi PD thanks for commenting. Interesting. So would he also say that if there is nothing in the idea of consciousness or the brain that makes them incompatible, that a conscious brain must be logically possible.
What about the cogito? I cannot conceive that I do not exist, therefore I do exist. It is the inverse of what you wanted ((~C->~P) rather than (C->P)). But it is certainly historical and influential, and students seem to find it compelling.