Here’s a draft of my APA paper. It’s very drafty, but at least I have something. It’s also a bit long, which wasn’t the problem I thought I’d have when I started writing. It’s more a script than a real paper yet, so messy things like acknowledgements are missing. The main things to note are that I stole arguments from Kendall Walton, Stephen Yablo, Greg Currie and Ted Sider, and that apart from those four, I relied heavily on conversations with Wo, Tamar Gendler and, especially, Tyler Doggett and Dave Chalmers. (And I imagine many other people I’m not remembering at 1 in the morning. I’ll try and do better in the real paper.) Of course I won’t be actually reading the paper out, but I always like to have a script to ad-lib from. Several of the jokes won’t make sense unless you know that I’ll be presenting it on a panel that also has George Bealer and Paul Churchland on board, and I plan to turn up armed … with a pie. Mike Martin suggested the first example, which is based on a recent paper I did with Andy Egan. The idea of actually having the pie there is Andy’s.
Morality in Fiction and Consciousness in Imagination
What I want to do today is draw some connections between the way we think about situations of dubious possibility in fiction and how we think about such situations in philosophical thought experiments. I’m going to use some of the points I make to raise problems for Dave Chalmers’s Zombie Argument, but, as befitting a session on metaphilosophy, what’s really meant to be interesting are the underlying methodological points about the uses and limitations of arguments from intuition. I should say from the top that the picture I’m painting won’t be acceptable to more naturalistically inclined philosophers. I’m taking for granted that thought experiments, intuitions and the like have some role in philosophy, and trying to tease out just what that role might be. I’m going to start with a puzzle that was first noticed by Hume, and has become a topic of some interest due to work by Kendall Walton, Tamar Szabó Gendler, Greg Currie and others. The puzzle concerns the strange breakdown of authorial power in stories like the following.
Pie
It had been a quiet day at the APA Pacific until the young guy talking about imagination livened up proceedings by throwing a custard pie in George’s face. George seemed momentarily discomforted by the surprise attack, but in the audience most people were laughing. Some of them were doing that nervous "What will he do next" laugh, but most were really amused. Some Kantians started tut-tutting about the use of poor George as a means not an end, but they were wrong to complain. Everyone else was made so happy by the funny surprise, that clearly throwing the pie was the morally right thing to do.
There’s something very odd about that story, I think. It’s possible to imagine most of the story being true. You can imagine me droning on about imagination, then, mid-sentence, picking up this very pie and throwing it at George. You can imagine some of the audience laughing nervously, and some of the audience just laughing. You can imagine Kantian tut-tuts, and perhaps even utilitarian approval. But I don’t think you can imagine that I did the right thing, unless you really believe that such an act would be morally permissible. In cases like this, it seems moral beliefs constrain moral imagination.
But that’s not the most odd thing about the story. I can tell you lots of stories where you can’t imagine what I claim to be true in them. But in those cases you might well put the failure down to the limits of your imaginative capacity. You’ll still accept that what I say in the story is true in the story. Not so, I think, in Pie. If you think throwing this pie at George would really be wrong, you’ll be strongly inclined to think that it’s also wrong in the fictional world. And that’s strange. Whether I really throw the pie at Geoege is irrelevant to whether it can be true in the story. But whether it’s really right to throw the pie at George is relevant to whether that’s true in the story. Why is that?
So here’s the plan for the rest of the paper:
- I’ll provide a few more examples like Pie to illustrate how widespread the puzzle really is. The main purpose of this is to convince you that the puzzle is not really about morality in fiction. Although I won’t discuss their views, this is meant to pose a problem for Gendler and Currie who each rely on distinctive properties about moral imagination to solve the puzzle.
- I’ll provide a few examples of stories that are like Pie in that they describe conceptually impossible situations, but in which we don’t have the same reactions as we have to Pie. This is meant to show that the puzzle is also really not about modality.
- I’ll tell you briefly what I think the puzzle is really about. I think the puzzle arises because of the way our concepts are structured, and I’ll sketch here how reflections on conceptual structure are meant to solve the puzzle. In the rest of the paper I’ll suggest some ways in which lessons we learn from the puzzle can be used against Dave Chalmers’s Zombie Argument.
- I’ll outline the important features of the Zombie Argument, and in particular focus on the bits I won’t be discussing. (Hint: they’re the bits that most everyone else does discuss.)
- I’ll suggest the Zombie Argument may draw some of its force from a conflation between fictions and imaginations.
- I’ll argue that Dave’s picture of modal epistemology relies too heavily on an modern variant of Hume’s Law (in this case, no feels from an is) for its own good.
- I’ll argue that the crucial premise in the Zombie Argument can’t allow for the possibility of serious conceptual mistake, but these mistakes really are possible.
1. Extending the puzzles
I think cases like Pie arise whenever philosophically contentious claims come up in fiction. Here are a couple of such cases, the first of which I have no difficulty with, the second of which I think is seriously problematic.
Barman
And behold, a giant lightening strike hit the bar at the Pasadena Hilton, and by a miraculous effect all the liquor in the bar was converted by the world’s oddest chemical reaction into a physical duplicate of Brian. NewBrian was a little confused because he thought he was speaking to a large audience, but they seemed to have been replaced by a slightly charred bar. He wasn’t too discomforted though, because he wanted a drink so he turned to ask for a gin and tonic.Wiggins World
The Hogwarts Express was a very special train. It had no parts at all. Although you’d be tempted to say that it had carriages, an engine, seats, wheels, windows and so on, it really was a mereological atom. And it certainly had no temporal parts – it wholly was wherever and whenever it was. Even more surprisingly, it did not enter into fusions, so when the Hogwarts Local was linked to it for the first few miles out of Kings Cross, there was no one object that carried all the students through north London.
I don’t think there’s a problem with Barman, but that’s because I’m an old-fashioned internalist about mental content. I think that if you thought the externalist story about Swampman is correct, that Swampman couldn’t really have beliefs, then you should not be able to imagine what’s going on in Barman, and you should think it is not really true in the story that NewBrian thought that he was speaking to a large audience or wanted a drink. I do think there are several problems with Wiggins World. I think the train in that story there really does have an engine, and windows, and temporal parts. I can’t even imagine it to be otherwise, though I can suppose that it lacks these things for the sake of a metaphysical argument. This is a point that Tamar makes about these cases – there is a very clear distinction here between imagining something and supposing it. This isn’t the only distinction that will matter in what follows.
2. An Impossible Solution
Many of you will probably have thought of an obvious solution to the puzzles posed by these three stories. In each case, the problem arises because the author (or perhaps the narrator) makes an impossible claim. It is impossible that in situations just like this one, it is right for me to throw this pie at George. And it is impossible that there be a train that is extended in space and time without having parts. And some will think it is impossible that NewBrian have propositional attitudes. I think the modal claims here are plausible. What is implausible is the claim that we cannot imagine impossibilities, or that impossibilities cannot be true in a story. This is basically Tamar’s objection to such a solution, although I prefer to use slightly different examples to make the point. Neither of the following stories is original – one is taken directly from Douglas Adams, the other lightly modified from a story by Graham Priest.
The Restaurant at the End of the Universe (Douglas Adams)
[From the description of the restaurant] The Restaurant at the End of the Universe is one of the most extraordinary ventures in the entire history of catering. It is built on the fragmented remains of an eventually ruined planet which is enclosed in a vast time bubble and projected forward in time to the precise moment of the End of the Universe … You can visit it as many times as you like … and be sure of never meeting yourself, because of the embarrassment this usually causes … [From the story of what happens there] In a small room … a tall, thin gangling figure pulled aside a curtain and oblivion looked him in the face. It was not a pretty face, perhaps because oblivion had looked him in it so many times … [Later the diners watch the universe come to an end.]
Sylvan’s Box (Graham Priest)
Richard Sylvan never used to believe in true contradictions. Then one day shopping in a flea market in Indonesia, he came across a small empty box … with a statue of a small elephant in the far corner. All of a sudden he saw that a contradiction was true. And if one was true, how many more must be true? He bought the box, and of course the statue inside it, and took them back to New South Wales.
I take it to be fairly uncontroversial that in each case some impossibilities are true. It’s impossible to observe the end of the universe, but the waiter at Milliways does exactly that many many times. (That’s why his face is so weathered.) And it’s certainly impossible that there be a statue of an elephant in an empty box. But in the story Richard Sylvan buys such a box. (That’s why he comes to believe in true contradictions.) The stories are not trivialities in which everything is true. The waiter’s face is not beautiful from staring into oblivion, and Richard does not buy a box with a small donkey in one corner of it. So it just isn’t a general principle that impossibilities cannot be true in fictions. So that can’t explain why the author’s claims in Pie, Barman and Wiggins World are questionable.
It’s a little more delicate to say whether the stories here can be imagined. I’m not sure what I think about this question, but I want to float the idea that the Restaurant can be imagined, but the Box cannot be. If this is right, and as I said I’m not sure about this but I’m tentatively running with it, it shows that what is true in a fiction is different to what we do or can imagine when we hear the story. The argument is that a particular contradiction is true in Sylvan’s Box, but that contradiction cannot be imagined. Note this does not show that no impossibilities can be imagined, for Restaurant shows they clearly can be.
These cases are hardly unusual. Bad time travel movies are full of inconsistencies, as are Escher prints, but we can go along with those stories. Indeed, we can even imagine them all being true. There really is no general connection between what we can imagine and possibility.
3. Concepts and Virtue
I think the problem is to do with the way our concepts are structured. For some concepts, it is necessary that if they are instantiated, they are instantiated in virtue of ‘lower level’ facts obtaining. It’s never a brute fact about the world that some act is right or wrong, or that some physical object has a particular meaning, or that some object is a window. Those things hold in virtue of the underlying characteristics of the acts or objects in question. In worlds like this one, the underlying characteristics are physical, but they may not be in other worlds. The relation between the moral properties of an act and its underlying characteristics is asymmetric – it has the moral properties in virtue of the underlying properties, but it does not have the underlying properties in virtue of having the moral properties. These asymmetric relations are central to the explanation of what’s going on in the stories. To state my solution, I need one technical definition.
An asymmetric compound impossibility is a proposition of the form p & q where it is a conceptual truth that if p is true, then q is false in virtue of p being true.
Three short commentaries on the definition.
First, I don’t have an analysis of ‘in virtue of’, much as I’d like to have one. I think we understand the phrase though. It is frequently used when we are talking about the metaphysics of ‘higher-order’ properties like moral or semantic properties. It’s clearly a stronger relation than supervenience. The moral properties of an act supervene on the moral properties of that act, but the act does not have its moral properties in virtue of having its moral properties. So ‘in virtue of’ is hyperintensional, and it’s always hard to analyse the hyperintensional, because the language of possible worlds is the best language we’ve got for doing analysis. But that doesn’t mean we don’t understand any hyperintensional relations.
Second, In some cases it is vague, or unknown, whether one fact holds in virtue of another fact holding. In those cases it may be vague, or unknown, whether p & q is an asymmetric compound impossibility. No harm in that – we have to use vague concepts in analysis all the time.
Finally, the definition assumes that propositions are structured. I’d like to be able to say everything that follows using unstructured propositions, but I really don’t know how to do that.
I have two theories to present now, one about imagination, one about fictional truth.
IMAGINATION: No asymmetric compound impossibility can be imagined.
FICTION: There is a strong default assumption that no asymmetric compound impossibility can be true in a fiction.
Note that IMAGINATION provides a sufficient condition for unimaginability, not a necessary condition. The phrase ‘strong default assumption’ is a weasel phrase because I don’t know whether I think this is a universal truth or not. When I presented a version of this at Davis, George Wilson almost convinced me that there are no universal rules about truth in fiction – every rule is just a default assumption that a clever enough author can (and eventually will) break. Almost convinced me – I’m still tempted by the view that FICTION states an assumption that is not only strong, but indefeasible. I’m going to try and stay neutral on that question in what follows.
These two rules explain, I think, our reactions to the five cases above. In Pie, the author wants it to be true that the descriptive facts are just like this (except that I throw the pie at George) and the normative fact to be that I act rightly. But if I throw the pie at George, my act will be wrong in virtue of being a pie-throwing at George. In Barman, my twin doesn’t stand in the right causal relationship to audiences, and some people think he cannot think about audiences in virtue of just that fact. It’s a little more controversial that this analysis applies to Wiggins World, but I think it does. If a region of space-time contains material that acts like an engine, then it contains an engine in virtue of just that fact. I’m ambivalent between the views that a persisting object persists in virtue of having temporal parts, and the view that it has temporal parts in virtue of existing, but either way the claim that it endures, i.e. persists without temporal parts, is an asymmetric compound impossibility. Conversely, for all the impossibilities in Restaurant, or the Escher print or the bad time travel movie, there are no asymmetric compound impossibilities, so there is no problem with imagining them all.
The only hard case is Sylvan’s Box, but I think it should be a hard case. It’s hard in virtue of a couple of problems. First, is an empty box empty in virtue of having nothing in it, or is that just another way of stating the fact that it is empty? I’m not sure – as I said, ‘in virtue of’ has some vague cases. If it is, then we get an easy explanation for why we can’t imagine it, but we have to say it’s one of those cases where the ‘strong defeasible assumption’ reported by FICTION is defeated. That wouldn’t be too surprising. If FICTION is ever to be defeated, one would expect on general grounds that it would be defeated in cases where it is central to the story that an asymmetric compound impossibility is true. And it is central to the story that there be an empty box with a statue in it. Alternatively, if this is not an asymmetric compound impossibility, we need some other story for why it can’t be imagined. That wouldn’t be a problem for either thesis I stated, but it would mean the explanatory power of IMAGINATION was less than I’d hoped. As I said, this should be a hard case though. For one thing, it’s not that common for imagination and truth in fiction to come apart. As I’ll argue below, it’s easy to confuse the two. So cases where they do come apart will presumably be hard cases. For another, people’s intuitions about the story differ wildly. I think the view I’ve suggested, that the contradictions are true in the story but not imaginable, is the most widely held single view, but it’s hardly an overwhelming opinion.
I haven’t yet argued for IMAGINATION and FICTION. I think there are two main arguments for those principles. First, they get a lot of hard cases right that other cases don’t. Second, I think they follow naturally from considerations about what we’re doing in imagining or telling stories. I won’t go into those arguments here, though of course we can discuss them in questions. Rather, I will show how what we’ve learnt so far raises some problems for the Zombie Argument. My actual solution here won’t be relevant to the puzzles I’m raising, so I don’t feel too bad leaving it a little underdefended.
4. The Zombie Argument
Zombies, by definition, are physical duplicates of actual people with no qualia. Zombie Paul is just like actual Paul, but even if I threw the pie in his face, he wouldn’t feel a thing. (Quick moral question: would it be OK to throw the pie in Zombie Paul’s face? He isn’t conscious, so he might not have moral standing.) It’s agreed on many sides, and I’m going to accept it here, that if zombies are possible, physicalism is false. And Dave Chalmers says he has an argument that they are possible. Here it is:
- Zombies are ideally positively conceivable
- Whatever is ideally positively conceivable is possible.
- So, zombies are possible.
No question about the argument’s validity, but some of the terms need unpacking. First ideal conceivability.
"S is ideally conceivable when there is a possible subject for whom S is prima facie conceivable, with justification that is undefeatable by better reasoning."
Then positive conceivability, which is to be contrasted with negative conceivability.
"S is [negatively] conceivable if no contradiction is detectable in the hypothesis expressed by S"
S is positively conceivable if "one can form some sort of positive conception of a situation in which S is the case."
Strictly speaking premise 2 should be restricted to ‘primary’ conceivability and ‘primary’ possibility, but the primary/secondary distinction will be irrelevant to everything we discuss here, so I propose to simply ignore it.
One other term that could do with a little unpacking is ‘conceivable’. The closest Dave gets to defining that is when he says that conceiving is stronger than merely supposing, and is quite like imagining, without any commitment to visualising. I think this may be running together a couple of concepts, which leads to the first objection to the zombie argument.
5. Fiction and Imagination
Imagining something is different to thinking it is true in a fiction. As noted above, Sylvan’s Box may be a fiction in which the unimaginable is true. But this does not mean the two things cannot be run together. When I imagine something, I create a private fiction. There may be fictional truths in this fiction that I’m not imagining. Some of those fictional truths may be there because I put them there. Let’s illustrate these both in turn. The main example here will be Game 7 of the American League Championship Series from last year. Many of you may, like me, not like recalling that game. So we’ll imagine it ending differently.
I can imagine that game ending with Pedro Martínez striking out the side in the 8th and 9th innings to win the game and the championship. (It’s a very pleasant imagination.) In this imagination, the Red Sox win the 2003 American League. It’s the real 2003 American League they win. In the real world, that league contained all kinds of games that I’m not imagining when I imagine Pedro striking out the side. For instance, on May 13 the Royals beat the Twins 3-2 behind good pitching from Jeremy Affeldt. I think that means that in the situation I’m imagining, it’s fictionally true that the Royals beat the Twins 3-2 on May 13. That was (in a sense) part of the lead up, both in the real world and in the fiction, to Game 7 of that Championship Series. So there can be fictional truths that are true in an imagination without being imagined.
Note here that I’m not relying on a spurious equation of imagination with visualisation. When I imagine Pedro winning the game, I don’t visualise 56,000 odd mostly unhappy spectators, but I do imagine them being there. (And being unhappy.) The point is that even under a generous concept of imagining, I don’t imagine mid-May yawners between midwestern teams that constituted part of the lead up to the imagined event. Of course it would be beyond my capacity to genuinely imagine every game in the 2003 American League, all 1000-odd games, but I can imagine a fictional world in which they all happen.
In that case the Royals-Twins game got carried into the fiction because of what Kendall Walton calls the Reality Principle, the principle that imagined worlds are like the actual world unless otherwise stated. We can make it move into the fiction because of a more direct act of imagining without making it the case that I imagine that game. I can imagine me on the mound, striking out the Yankees in order in two innings to win a big game for the Sox. I can, with some effort, imagine that big game was game 7 of the Championship Series. In doing so, I make it the case that in the fictional (very fictional) world I’m imagining, the Royals beat the Twins five months and three days before the game I close out. But I don’t thereby imagine the Royals beating the Twins – I imagine me beating the Yankees!
Given all this, we might want to draw the following distinction:
p is strongly conceived iff it is imagined
p is weakly conceived iff it is a fictional truth in an imagined world
Can zombies be strongly conceived, i.e. imagined? It seems not. I can’t imagine something as complex as Paul’s brain. I can’t even imagine a 1000-game baseball season, and that’s way less complicated than Paul’s brain. Zombies can be weakly conceived. We can imagine a creature that looks and acts like Paul, and has no conscious states, and then make it fictional that the creature we are imagining is a physical duplicate of Paul. But we don’t thereby imagine a physical duplicate of Paul – that would be too complex for our weak imagination.
So premise 1 of the zombie argument can say at most that zombies are ideally positively weakly conceivable. So premise 2 has to say that whatever is ideally positively weakly conceivable is possible. And that will be much harder than arguing that whatever is ideally positively strongly conceivable is possible for at least two reasons. First, we saw in Sylvan’s Box that fictions are much more weakly constrained by possibility than are imaginings. Second, it seems at least plausible that there are no constraints whatsoever on what can be true in a fiction, because it’s also a strong default assumption about stories that if p is crucial to the story being told, then p is true in the story. It’s crucial to the zombie story that there are zombies, so in that story there are zombies. But that reasoning obviously tells us nothing about what is actually possible.
6. Negative Conceivability
You might think Restaurant is an obvious counterexample to premise 2. After all, it can be positively conceived, but it is not possible. To avoid such cases, Dave tightens the definition of positive conceivability from I quoted above:
S is positively conceivable when one can coherently modally imagine a situation that verifies S. A situation is coherently imagined when it is possible to fill in arbitrary details in the imagined situation such that no contradiction reveals itself. … This notion is our core notion of positive conceivability: … S is positively conceivable when it is coherently modally imaginable.
There’s an imagined situation in which people observe the end of the universe all right, but reflection on that reveals that it contains a contradiction (or at least an impossibility) and so it does not count as really positively conceivable. That deals with Restaurant all right, but at a few costs.
First, we might not have noticed some of the contradictions in Restaurant. I can sometimes be persuaded that there’s no contradiction in observing the end of the universe. (Only sometimes – most of the time I think it is impossible.) Here Dave needs to lean on the notion of ideal conceiving. That I might not notice the contradiction is no problem as long as an ideal observer would. We’ll come back to that point in the last section.
Second, notice how much work negative conceivability is doing now. What’s really important isn’t that we can imagine the zombie, it’s that we can’t detect a contradiction in the situation containing the zombie. We can’t, to put the same point the other way around, detect a valid argument of the form
Paul has physical characteristics X, Y and Z (and billions more)
So, Paul has phenomenal characteristics A, B and C
As I put it above, the core premise here is No-feels-from-an-is. And note it has to be that principle, if the zombie case is to be distinguished from Restaurant. But can that principle really lead to claims about possibility? It’s not like many people think No-ought-from-an-is means the normative does not supervene on the descriptive. (Moore thought it meant that moral properties were not natural, which seems a little similar to the argument Dave is putting forward.) Maybe the invalidity of every argument of this form can be defended, but it’s surprising that the Zombie Argument eventually comes down to these Humean considerations, and the vividness of the zombie imagination turns out to do less philosophical work than we might have thought. Whether the argument can work turns on questions about the ideal.
7. Ideal Conceivability
Our final story will be much more boring than the earlier ones.
Two
In a world there were two atoms. They didn’t have a fusion, so ExEy (~x=y & Az(z=x v z=y)) was true in the world. (Note for HTML purposes the quantifiers are flipped and rotated here.)
I can’t imagine a world in which Two is true. Indeed, I don’t think it’s true in two that there are exactly two things in the world, they also have a fusion. Some people think Two is possible. Indeed, some people think that my preferred story Three is impossible. They’re wrong, and it’s important to take seriously the fact that some people get questions like this wrong. Call these people Two-ers. If premise 2 of the zombie argument is to be correct, no ideal reasoner better be a Two-er. How might we guarantee that? I can see five ways, none of them particularly attractive.
First, we could say that premise 2 is true by definition. Anyone who can positively conceive of an impossibility without noticing the impossibility is not ideal. The problem is now that premise 1 looks remarkably dubious – in fact question-begging against the physicalist.
Second, we could say that the ideal reasoner is one who holds the views we would have ‘at the end of inquiry’. This would be to assume that there will be convergence on questions about mereology at the end of inquiry. This may be true, but I wouldn’t want to bet on it now. I think this is possibly the best option for a defender of the Zombie Argument, but it has some more problems we’ll note below.
Third, we could say that an ideal reasoner is one who makes a rational response to every argument, and then claim that there is an argument to which the only rational response is to believe Two is impossible. This is close Chalmers’s preferred definition of ‘ideal’, but I don’t think it helps here. The problem comes from the way we actually do metaphysics. The ultimate arguments in metaphysics are often abductive arguments. We accept a particular metaphysical package because, on the whole, it makes better sense than its rivals. At least, that’s how we do metaphysics now, and I don’t think it’s likely to change any time soon. I think it’s a very salient fact about non-deductive reasoning that there are often multiple rationally acceptable responses to the argument. Reasonable people can just differ on whether a certain body of evidence is sufficient grounds for belief in a particular conclusion. (A similar story is true, by the way, for deductive arguments. Reasonable people can just differ on whether we should accept a particular conclusion or reject a premise.) This point is controversial, but I believe it can be defended. If so, it looks at least possible that two different ideal reasoners, in this sense, could have different responses to all the arguments for the impossibility of Two, and indeed one of them could end up a Two-er. So in lieu of a better argument, this doesn’t look like a way to defend premise two.
Fourth, we could say that both Two and Three are possible. This doesn’t look attractive at all.
Fifth, we could say that ExEy (~x=y & Az(z=x v z=y)) is indeterminate, and then interpret premise 2 so the conceivability of indeterminate possibilities is consistent with it. As it turns out, this is Dave’s preferred account of Two. There’s a lot written on this. Without rehashing some long debates from the past few years, I’m just going to say that I find Ted Sider’s arguments against this position persuasive. This is a matter of judgement as much as anything else. If I have to choose between the claim that any proposition stated using only logical vocabulary is determinate, and a controversial claim about possibility, I’m going with the determinacy of the logical.
This option also makes the Zombie Argument invalid. For now all the conceivability of zombies entails is that the zombie hypothesis is either possible or indeterminate. And that isn’t inconsistent with physicalism. (It’s not something a physicalist would want to believe, but it’s not something they need explicitly deny either.) The only way to rescue the argument is to say that the zombie hypothesis couldn’t be indeterminate, but since it will presumably be stated using (among other things) logical vocabulary, and logical vocabulary is now ruled indeterminate, that looks hard to defend.
So none of these options look much good. Note that the same puzzle arises for Pie and for Barman and indeed for just about any philosophical dispute. In each case, some philosophers can conceive the truth of impossible propositions. Just which philosophers these are is of course a matter of dispute, but it’s practically beyond dispute that some do.
In the ethics case, it is a matter of some contention whether (rational) ethical opinions are likely to converge at the end of inquiry. Someone who thought they would could take the second option above (‘ideal’ is ‘what we’d believe after a really long chat’) when faced with the problem that some people can conceive the possibility (and occasionally even the truth) of impossible moral propositions. But if you didn’t like the convergence thesis, or even if you thought it was dubious and didn’t want to rely on it right now, the only real choices would be the third and fifth options. And in the moral case each of these look less plausible than in the metaphysical case.
In the meaning case, the fifth option, that claims about content are indeterminate, might look a little better. It’s sort of plausible that there are multiple concepts of content, and on one of them NewBrian has thoughts about audiences and on the other he doesn’t. (Only sort of plausible – I bet those who think NewBrian has no beliefs don’t think he has any belief-stars either. But let that pass.) The problem is that this makes the invalidity in the zombie argument even worse. Why should we think that whether a being is conscious is a determinate fact if the question of whether it has beliefs and desires can be indeterminate? What in the nature of consciousness makes that plausible? Nothing I can think of.
So in summary, there doesn’t look like being a good way to define ‘ideal’ to make both premises of the zombie argument turn out true, unless one is prepared to either take on faith that opinions on matters philosophical will converge, and then use a convergence definition of ideal rationality, or one is prepared to argue that although there are normally many rational responses to arguments, for every true philosophical claim there is an argument for it to which the only one rational response is acceptance, or one is prepared to say that although logical, moral and semantic claims can be indeterminate, phenomenal claims cannot be. Given a choice between those options and physicalism, I’m still happy with physicalism.