(I was originally going to flesh this out and try and get this published, but have come to the belief that the blogosphere is intellectually and morally preferable to academic journals for discussing these ideas. Also, this is obviously way easier, and I’ve been sitting on the argument for so long I just want to put it out there and see what people think. So I’ve mostly just copied this from the draft, and put some token links in to make it look like a blog post…)
It has been alleged against the epistemic account of vagueness that the determination of the precise meaning of a vague term appears miraculous, severing the connection between meaning and use. Given the infinite possible precisifications of terms like “thin”, “heap” and “tall”, how on earth is one of those concepts picked out by our ordinary expressions? However, an actual refutation of epistemicism has never been forthcoming, and the epistemicist defence (e.g. Williamson 1994) against the above accusation has been to correctly point out that there isn’t any promising account of how the meaning of non-vague terms is related to use, and that it is unfair to demand that one be provided:
“Every known recipe for extracting meaning from use breaks down even cases where vagueness is irrelevant. The inability of the epistemic view of vagueness to provide a successful recipe is an inability it shares with all its rivals. Nor is there any reason to suppose that such a recipe must exist.” [Williamson 1994 p207 (1996 edition)]
Whilst I agree with Tim Williamson (no relation!) that such a recipe may forever be beyond our grasp, we can still ask whether the epistemic account of vagueness is consistent with the necessary conditions of a true account of meaning determination. It could be the case in the physical sciences that a ‘Theory of Everything’ is necessarily impossible to describe, but this would not prohibit us from making less ambitious but nevertheless true statements about physics and seeing whether they are consistent with other propositions. I submit than the following condition would have to be met in the case of meaning determination
(1) The determination of the meaning of a word/expression/concept* makes reference to the actions/assertions/intentions/beliefs* of the speakers of the language
(1) does not imply that meaning is exclusively determined by use, whatever that may mean, but it is impossible to see how we could explain how words change their meaning without some version of it. The ordinary phenomenon of semantic change provides the most compelling intuitive support for there being a role for the speakers of a language in determining the correct meaning of the words they use. Examples abound of words whose meaning has changed over time, with the most plausible explanation of the change being that the speakers of the language gradually took the word to mean something different. This is completely consistent with a realist account of universals, whereby the word “awful” could have once denoted the same universal as “awesome” does now, with the changing actions/assertions/intentions/beliefs of English-speakers eventually changing the denotation.
But who exactly are the speakers of English? The set of English-speakers is a classic example of a vague set; there are borderline cases in terms of who alive today might count as a member of the set (infants/bilinguals/speakers of dialects), and there are borderline cases going back in time (Was Chaucer a speaker of English? How about Shakespeare? Or Samuel Johnson?). On the epistemicist account, it is either True or False as to whether Chaucer spoke my language, even if we cannot ever know whether he did.
However, “English-speaker” is a word in English and, if (1) is correct, its meaning is partially determined by the actions/assertions/intentions/beliefs of the set of English-speakers. The epistemicist cannot account for how “English-speaker” takes on a precise meaning in English, for it requires that an already precise set of individuals be constituted (the set of English speakers), whose actions/assertions/intentions/beliefs serve as an input into the function that is supposed to determine the contents of the very same set. It would be as if a polity attempted to establish the extent of the franchise on the basis of a democratic vote.
Therefore, unless the epistemicist can account for semantic change without positing a (1)-like principle that leads to the problem of the self-constituting set of language-speakers, the epistemicist explanation of vagueness cannot be correct. But since my claim is actually negative (I’m agnostic as to the precise form (1) must take, it merely seems to me that you *can’t* come up with an account of semantic change that won’t result in the circular feature I have described, although I don’t know how to prove it), I would be very interested if anyone can come up with a plausible substitute for (1) that does not have the problem for epistemicism I have described. I can’t think of one, which makes this argument the closest thing to a refutation of epistemicism that I am aware of, and as such is worthy of further inquiry.
Comments/counter-arguments/reading suggestions are most welcome.
*Delete as appropriate, according to your philosophical sensibilities
Suppose you have an intuition, ‘P’*. Should you believe P? Should you believe not-P? For some propositions (e.g. the moral kind) there may not be much to go on other than our intuitions. So for the moral intuitionists in particular, here’s a little puzzle for you.
Suppose intuitions have a probability of being true that is greater than 0.5. Therefore, if you have an intuition that P you should believe it rather than not-P. But if this is true, then the consensus of the masses’ intuition is not only more reliable than yours, but is actually very reliable (Condorcet’s theorem). If intuitions have a probability of greater than 0.5 of being true, you should therefore not trust your intuition, but ask the masses.
Suppose instead intuitions have a probability of being true that is less than 0.5. Therefore, if you intuit P, not-P is more likely. But not only should you should not trust your intuition, you can be pretty certain of the right answer by asking the masses and then believing the opposite!
Obviously I have posed the question in a very simplistic way, and there’s questions of statistical independence, the different reliability of different kinds of intuitions** etc. But it does seem that whatever the probability of your intuitions being true are, in a lot of cases it could still make sense to ask around and then either believe what the masses say, or the opposite (depending on your view of the prior probability of the particular intuition being correct).
*This is philosopher-notation for an arbitrary proposition. P could be anything – ‘Murder is always wrong’, ‘I am the same person I was yesterday’ etc.
**There even may be classes of intuitions that are necessarily reliable on the level of the masses, because they are intuitions about proper conventions… but let’s not go into that here
Suppose you are accosted by an omniscient highwayman – and instead of demanding your money or your life, he is after something altogether more tricky: empirical truth. Tell him something empirically true and you escape unharmed. What should you say*?
Consider, for a moment, externalist definitions of knowledge – for example that a belief is the production of a reliable mechanism for producing truth. Would this knowledge, if I had it, be of any use to me here? It would not, as it is not a condition of externalist knowledge that I can internally discern what is true and what is not true. My perceptual beliefs may be reliable, and if they are reliable then I know them, according to the externalist. But that knowledge is of no comfort to me in this situation, as it doesn’t help me make a choice whether to assert one proposition or the other. P may be the production of a reliable mechanism and Q may not be, but if I have no idea whether P or Q are produced by reliable mechanisms, I have no reason consciously available to me to assert P or Q to the highwayman.
If someone demands of me that I speak truth, in order to be able to make choices based on reason in order to meet that demand, it must be the case that any reason for believing something be available to me. Those are the only kinds of reasons that are any good to me if I’m to actively try and stay alive in this situation. I need to consider my internal justifications for believing any particular proposition to be true.
Now, suppose I’m considering whether proposition P would be a good candidate to assert to the highwayman. I look at my reasons for P. Say I believe P because of [R → P] and R. But why then not assert either R or [R → P]? Surely P cannot be on a better internal epistemic footing than both of those two propositions put together? For if my belief that P is truly justified only by [R → P] and R, then P would have the same evidential status as [[R → P] & R]. Since you cannot increase the probability of a proposition being true through conjunction, I should be better off asserting either R or [R → P].
Suppose then I choose to assert R. Why do I believe R? If I give reasons, then I am better off asserting the reasons for R than R itself. I can never get to a point where I can rationally assert one proposition rather than any other. This is true even if I gave some coherentist-style justification for a proposition, as the issue with coherentism is there is absolutely zero reason to believe than a coherent set of propositions remotely resembles the truth.
Now, fortunately we are not usually in the position of having to assert truth to an omniscient psychopath. But consider the following not-implausible epistemic norm:
A) You should assert only what is justified
As our highwayman example shows, if we go about actually trying to meet such a test based on what is consciously available to us, we’ll never get to a point of justification. External justification is of no help to us when consciously trying to meet epistemic norms. And that, in my view, is the real challenge of the sceptic. There is nothing intrinsically wrong with a belief having the property of being produced by a reliable mechanism, or of its being caused by the truth of its content. It’s a genuine property that any number of our beliefs could have, but since the justification is external it doesn’t help us in the deliberate practice of choosing what we should believe or assert. If we were to instead invoke a stronger epistemic norm:
B) You should only assert what you
know can discern to be true
And combine this with the ‘ought-implies-can’ principle, the fact that we cannot give any internal reasons for believing one proposition rather than another from our own internal perspective leads to the unhappy conclusion that
C) One ought not to assert anything.
*Some clever-clogs is going to figure out that the optimal strategy in this situation would be to say ‘My life is being threatened’. If it’s true then you pass the test, if it’s not then your life isn’t threatened anyway. So let’s also assume than our omniscient highwayman also has an irrational dislike for clever-clogs
If I murder someone, I can be imprisoned by the government. However, it is only the government that is allowed to imprison me – if you do it, that is called false imprisonment and is itself a criminal offence. Why is it that governments are allowed to lock people up for the good of society and individual citizens aren’t?
Given that using overwhelming force to get people to do what you want them to do (e.g. remain in an 8×10 room for the next 20-25 years) is generally speaking not acceptable for individuals or corporations to do, one might think there should be a reason why governments are allowed to. Let’s review a few candidates:
1) You expressly consented to the government having such a right.
In terms of giving an argument for the legitimacy of actual states, this is a non-starter. No such consent has been given, except maybe by those who acquire citizenship by choice.
2) You tacitly consented to the government having such a right.
It isn’t necessary for consent to expressly say ‘I consent’. Sex is an obvious example; non-consensual sex is illegal, but there are (mercifully) a number of uncodified ways of signifying consent. Even if we assume that such tacit consent would legitimize government coercion, there do not exist the kind of uncodified ways of tacitly signalling consent to government coercion as in the case of sex. Proposed acts of tacit consent to such coercion – voting, not voluntarily leaving the territory – do not have the necessary link between the proposed act and the understanding on the part of the actor of what it is they are consenting to by performing that act in order for the analogy with express consent to go through.
3) The principle of fair play
Another non-starter. The principle of fair play essentially this: if a community engages in a mutually beneficial enterprise that requires co-operation, then if you benefit from the enterprise you shouldn’t be allowed to free-ride on the cost of producing that benefit. It is a non-starter because it fails to explain why the enterprise of government generates a strong enough claim to allow coercion when other co-operative enterprises do not. It may account for a moral obligation to co-operate, but simply demonstrating the existence of a moral obligation is insufficient for legitimizing coercion as an instrument of ensuring the fulfilment of that obligation.
4) Natural Duty of Justice
The gist of this line of thought is that we have an obligation to support just institutions. If the state is just, then we should do what it says. But again, this only speaks to a moral obligation – that I ought to obey the rules of a just state does not speak to the right of that state to coerce me into following those rules.
5) Hypothetical Consent
Medical procedures, in the ideal world, are consented to. However, sometimes we are not able to consent to things that we would – I cannot consent to surgery after being involved in a motorbike accident where I am left unconscious, but the doctors are quite permitted to perform the surgery because I would grant the permission if given the opportunity. The hypothetical consent tradition in political philosophy says it’s fine that people don’t actually consent to government coercion, it’s enough that they would if asked. Why would they do so? Because it is way better that we have a government than not. However, hypothetical consent (when it comes to government) proves far too much. If it legitimizes all government coercion where having such a coercive authority is better than having none, then if anarchy is really really bad – as those in the hypothetical consent tradition are inclined to think it is, cf Thomas Hobbes – then just plain ol’ bad tyrannical government passes the hypothetical consent test. Furthermore, it might be better to allow coercion by both the government AND (say) me. If having government is better than anarchy, then probably so is having government and a few select individuals locking people up in their basements.
We can generalize the considerations I have briefly and flippantly adduced by saying that any theory of why it is OK for the government and no one else to engage in certain coercive practices should meet two constraints: it should cover at least the vast majority of people who can be subjected to such coercion, and most crucially it should explain why we reserve this right exclusively to the government. The second constraint has (to my knowledge) been ignored by all the literature on the subject, despite its being such an obvious and important feature of the modern conception of the state.
Now, I am not by any means a political anarchist. Indeed, I think these considerations point somewhat in the other direction, politically speaking. For example, suppose you are a hard-line libertarian who thinks it is the job of the government to enforce contracts entered into by consenting, mentally capable adults and nothing else. The government in such a contract is, in effect, an unnamed third party who the contracting parties agree to have settle a dispute and execute that settlement – by force, if necessary. But why, on this hard-line libertarian position, can’t consenting adults name some other third party? Why can’t the parties say in the event of a dispute or breach of contract that Mr T be brought in to settle it*? Even the most extreme of libertarians are incapable of accounting for why only government is allowed to use coercive force.
The lesson I take from this is that politics, like ethics more broadly speaking, involves us trading off against each other all sorts of things we find valuable. For example, there is a long and tedious debate amongst political philosophers as to what ‘freedom’ means. Does being free mean not having people interfere with what you want to do, or having the ability to do the things you want to do? The answer is: both matter, and both taken to the extreme lead to highly undesirable outcomes. If we want to feed everybody, then we may have to take stuff from one person and give it to another – stuff they might not give up otherwise (this is also known as ‘taxation’). That a homeless person might be ‘free’ to eat but unable is a simple demonstration that many things other than freedom from coercion are valuable. Deciding what a government or anyone should and should not be allowed to do is essentially trading off the value of not being coerced against whatever it is that can be achieved by the coercion.
* I pity the fool who would agree to such an arrangement
Suppose that determinism is true, and that I just put my hand down on my desk. As a compatibilist, I claim that this is a free but determined act. I was able to act otherwise, for instance to raise my hand. But there is a true historical proposition H about the intrinsic state of the world long ago, and a true proposition L specifying the laws of nature, such that H and L jointly determine what I did, and jointly contradict the proposition that I raised my hand. If I had raised my hand, then at least one of three things would have been true: contradictions would have been true, H would not have been true, or L would not have been true. So if I claim that I am able to raise my hand, I am committed to the claim that I have one of three incredible abilities: the ability to make contradictions true, the ability to change the past, or the ability to break (or change) the laws. It’s absurd to suppose that I have any of these abilities. Therefore, by reductio, I could not have raised my hand. – David Lewis
So goes the classic ‘consequence argument’ for incompatibilism, the thesis that free will is inconsistent with a deterministic universe. Against the consequence argument, Lewis goes on to point out that there are two different kinds of counterfactuals concerning the possibility of my raising my hand:
A: If I had raised my hand, either the state of the past or the laws of nature would have been different
B: If I had raised my hand, my act would have caused the past or the laws of nature to change
B is plainly absurd; one could not say that I ‘could’ raise by hand on the basis of B, and compatibilists do not say so. A, however, is perfectly all right on the determinist thesis, and incompatibilist ought to readily admit so. Now, with A there comes also an ability which I can be said to have, which I shall call A1:
A1: I am able to do something, such that if I did it, the state of the past or the laws of nature would have been different
I agree with the compatibilist that this represents a genuine ability I have. There is nothing contradictory in claiming anyone has such an ability, so I see no reason to think that it doesn’t represent a genuine property. The compatibilist claims that this argument for incompatibilism fails because it equivocates between the ability A1 which I do have, and the corresponding ability B1:
B1: I am able to do something, such that if I did it, the state of the past or the laws of nature would have been caused by my doing it
Against ability B1, the argument goes through*. But that, so says the compatibilist, is not what is at issue for we have abilities of the A1-type and those are good enough for making us free. Those who know me more than a little will be aware of my general disdain for arguing over definitions – if you want to call compatibilist freedom ‘free will’, then you are very welcome to make that stipulation. But the incompatibilist has available what I believe to be a very strong response, which is to say that the abilities such as A1 are morally irrelevant. For example, I have an A1-type ability to flap my arms and fly. I am able to flap my arms and fly, such that if I did it, the state of the past or the laws of nature would have been different. But it would be ridiculous to be accused of failing to act if I neglect to save someone from falling out of a building on the grounds that I ‘could’ have flapped my arms and flown to save them. So if we then extend this point to classic cases of, say, a child drowning in a pond who I fail to jump in and save because I didn’t want to get my clothes wet, the incompatibilist will make the analogous point which is my A1-type ability to jump in the pond and save the child is not a morally relevant one. It is of exactly the same kind as my ‘ability’ to flap my arms and fly, and I fail to see the morally relevant different between them except that in the save-the-child case we very much believe that I am morally responsible and so we just work backwards to get the right answer. But this is no way to do philosophy, for surely the onus is on the compatibilist to show why we should make a distinction.
I therefore conclude that while the consequence argument doesn’t necessarily succeed against it’s intended target – free will – it does make a powerful case for the incompatibility of moral responsibility with a deterministic universe. Indeed, I have often wondered whether it would be a big step forward in the debate over free will if we realized that what we are really arguing about is moral responsibility.
In philosophy, when we are trying to understand a concept or figuring out whether it actually applies as much as we initially believe it does, we can often get confused by two different ways of proceeding. One way is to look at the ordinary cases of the things we call ‘freedom’, ‘knowledge’ etc. and then see what is in common with those cases. Another way is to look at the kinds of reasons we give for or against the concept applying to a particular case. The problem is, these two ways of looking at it tend to come into conflict. In the case of freedom/moral responsibility – and, I would argue, in the case of knowledge – we have a conflict between our ordinary applications of the concept and the kinds of reasons we have for giving and (especially) denying that the concept applies. Sceptics about knowledge and moral responsibility say that if we consistently applied the kinds of reasons we give for denying that someone has knowledge (e.g. they haven’t ruled out all the possibilities) or is morally responsible (e.g. I can’t change my genes, so whatever my genes cause is something I am not responsible for) then we will end up denying that the anyone knows or is responsible for very much at all. It seems as if we either have to give up a whole lot of common sense thinking about who knows or is responsible for what, or we have to arbitrarily limit our application of the reasons we give for the concept not applying.
It strikes me that philosophy cannot tell you how to decide between those two, only that we must if we want to be consistent in our thinking.
*This very brief exposition of the consequence argument is indebted to Kadri Vihvelin’s excellent “Arguments for Incompatibilism” entry at the Stanford Encyclopedia of Philosophy