Lewis on Devitt on Ostrich Nominalism

In his “New Work for a Theory of Universals”[1], David Lewis discusses Michael Devitt’s defense of Ostrich Nominalism in his article “‘Ostrich Nominalism’ or ‘Mirage Realism’?”[2], specifically as a response to the One Over Many argument. Devitt had proposed to paraphrase such sentences as

“a and b have the same property, F-ness”

as

“a and b are both F”

which itself can be analyzed as

“a is F”

and

“b is F”.

Lewis thinks that this is not satisfactory. He says:

But Devitt has set himself too easy a problem. If we attend to the modest, untransformed One over Many problem, which is no mirage, we will ask about a different analysandum:

a and b have some common property (are somehow of the same type)

in which it is not said what a and b have in common. This less definite analysandum is not covered by what Devitt has said.[3]

I think there is an obvious paraphrase of Lewis’ example which, though not explicitly covered by what Devitt had said, is in perfect harmony with its spirit. Indeed, I think it’s obvious enough that it's probable someone else has already thought of it, which for a while made me hesitant to make this post. Nevertheless, I’m interested to see if others think the paraphrase works, so I’m posting this anyway, even if I can’t claim originality for it. The paraphrase goes like this (where F, G, H, etc., are all the predicates expressible in the language):

((Fa & Fb) v [(Ga &Gb) v (Ha &Hb)])…

Or, in English:

Either a and b are both F or a and b are both G or a and b are both H…

So my question is: Do you think the paraphrase works? And if not, why not?

---

[1] As reprinted in Properties (Oxford Readings in Philosophy), edited by D. H. Mellor and Alex Oliver. Oxford University Press, 1997.

[2] Reprinted in the same volume as in the above footnote.

[3] See p. 201 of Properties (Oxford Readings in Philosophy).

What do philosophers agree on? A philosophical poll

At first glance, it is remarkably difficult to get philosophers to agree on anything—or so it seems to me. But is it really? I’d like to get your opinion on the subject, via this informal poll. On what philosophical issues do you think a majority of philosophers agree? Just name, or describe—briefly—the relevant issue, and what position you think the majority of philosophers take on it.

There are two things to keep in mind:

First, I don’t mean to restrict this poll to the opinions of contemporary philosophers—if the majority of philosophers throughout history have taken the same position on some issue, I count that agreement as the majority view, even if a majority of contemporary philosophers reject it.

Second, I’d like to include what I will call “Moorean agreements”; things that, based on your experience, a majority of philosophers believe or assume in practice even if some of them claim to reject it while doing philosophy. (I think the belief that people are, in general, morally responsible for their behavior is an example of a Moorean agreement.) Of course, it can be hard to determine what philosophers—or anyone, for that matter—“really believe” about some issue, so when it comes to establishing what the majority view is on some philosophical issue I regard (purported) examples of Moorean agreements as being less conclusive than examples of what philosophers have explicitly claimed to believe. Still, it will be interesting to see how much agreement there is on what the Moorean agreements are!

And that’s it. I anxiously await your contributions. ^_^

A problem for consequentialism

I think one of the main problems with utilitarian theories is that they do not attribute any value to moral agents or moral patients as such. Their experiences might be valuable, as might the satisfaction of their desires, but moral agents and moral patients themselves are of no value at all. (From this point on, I will use the term “moral subjects” to refer to both moral agents and moral patients.) So if, for example, I kill Smith and somehow create a new individual whose well being is the same as Smith’s was, this state of affairs is, intrinsically, neither better nor worse than how things would be if I had left Smith alone. But this seems wrong. To borrow a term from W.D. Ross, it is our prima facie duty not to kill people, and this is so even if killing someone has no impact on the total amount of utility in the universe as a whole. Because utilitarian theories cannot discriminate between actions which result in the same overall amount of utility in the world, they are blind to the fact that one such action can be permissible while another such action is prohibited.

Consequentialists, more generally, could try to remedy this problem by assigning intrinsic value to moral subjects in themselves, apart from any value their experiences might have, and take this into account in their moral deliberations. After all, consequentialism in general requires that we try to maximize the good, but is silent on which things are good. But I doubt that consequentialists can acknowledge the intrinsic value of moral subjects without giving up on consequentialism or else failing to do justice to what we normally mean when we say that moral subjects are intrinsically valuable. In order to have a genuinely consequentialist theory, consequentialists would have to treat the value of a moral subject as being comparable with other sorts of value, such as happiness, or well being in general. But then we are faced with essentially the same problem we encountered above: Suppose Smith himself is worth 11 units of goodness (utiles), while his well being is worth 10 utiles, so that the total value of Smith’s life is 21 utiles. Why can I not kill Smith, provided that I also create 21 utiles through some other means to make up for the loss? If I do this by creating a new person, they would presumably be worth just as much as Smith was; so as long as their well being is also worth 10 utiles we have a life which is worth exactly the same as Smith’s. All the same, it is still wrong to kill Smith, even if I create a new person to “make up for it”. So a consequentialist theory has again given us the wrong result, even though we augmented it so that it assigns intrinsic value to Smith himself and takes this value into account in determining one’s permissible courses of action. What has gone wrong?

The problem is not that we have assigned Smith too little value—all persons are equally valuable, and so just as valuable as Smith is, and given this we can always concoct a new scenario in which we create enough utiles to make up for murdering Smith. This is so even if one thinks that moral subjects are of infinite value. And that view is problematic on its own: If moral subjects are infinitely valuable, one moral subject is worth just as much as a million. But if forced to chose, one surely ought to save a million rather than one.

If there is any hope for the idea that moral subjects are intrinsically valuable, I think it must lie in something like Kant’s Categorical Imperative, which says that moral subjects ought always to be treated as ends in themselves, and never merely as a means. One of the problems with the above lines of reasoning is that we have taken the term “intrinsic value” at face value and have falsely assimilated the value of persons to the value of sub-personal things like happiness or well being. Moral subjects are not merely valuable, but unique and irreplaceable. Instead of saying that moral subjects are intrinsically valuable, it would be better to say that moral subjects have moral dignity, and that this is something which cannot be measured in the same way that the value of happiness or well being can. Perhaps it cannot be measured at all. What is certain is that we cannot hold that moral subjects may be replaced by other beings whose lives have “equal value”, for a being which has moral dignity is by that very fact one which ought not to be disposed of, even if they are replaced by another being who also has moral dignity. Neither may we use moral subjects as a mere means to improve the general welfare. It might still be true, in general, that we may save the many rather than the few, but it will only be permissible for us to do so if the circumstances which force this choice on us are not of our own making.

Are options the same as alternative possibilities?

When it comes to theorizing about choice, should we distinguish between having multiple options and having alternative possibilities of action?

Suppose a deterministic agent takes a multiple choice test. Consider a specific choice between a, b, c, and d. Suppose further that the agent chooses c. Did the agent have multiple options? It seems so; after all, can't they be clearly seen to be printed on the test? All the agent has to do is circle one! This is, I think, importantly different from a case in which there is only one candidate answer, say a, printed on the test. In neither of these cases is there a possibility of the agent choosing anything other than what they actually choose, because the agent is deterministic. But in spite of that, it seems to me that cases are not on a par; there is something present in the former which is absent in the latter, and this something has implications concerning the agent's freedom (or lack thereof) in choosing as they do.

From this somewhat deflationary perspective, the fact that more than one choice is given for an agent's consideration or that more than one choice influences an agent's course of deliberation is of itself sufficient for their having multiple, genuine options. (Having alternative possibilities of action might also be sufficient for having multiple, genuine options, but on this view it is not necessary.) If that's so, determinism is compatible with an agent's having multiple options, even though they couldn't do otherwise. The question is then whether an agent's having multiple options--along with such background conditions as being rational and knowing the difference between right and wrong--is sufficient for having free will or moral responsibility. I'm interested to see if you think it is sufficient, and if not, what more is required.

A new variant of the liar paradox?

A few weeks ago I thought of the following variant of the liar paradox. I won't pretend that this post advances our understanding of the liar in any way, I'm just wondering if anyone else has thought of it before.

Consider the following statement:

(1) If there are elephants, then (1) is false.

If there were no elephants this statement would not be problematic. But there are. So let's assume that (1) is true. Given that there are elephants and that (1) is true, it follows by Modus Ponens that the consequent of (1) is true, and hence that (1) is false. So if (1) is true, and there are elephants, then (1) is false.

Let's now assume that (1) is false, granting once more that there are elephants. If (1) is false, the falsity of the consequent of (1), and hence (1) itself, cannot follow from the fact that there elephants. On this supposition (1) and its consequent are false, to be sure, but their falsity must not be entailed by the fact that there are elephants. But if the fact that there are elephants does not entail the falsity of the consequent of (1), and hence (1) itself, it must be possible for there to be elephants while they are both true. But we have already seen that if (1) is true, and there are elephants, then (1) is false. Thus it is not possible for there to be elephants while (1) and the consequent of (1) are both true. So the fact that there are elephants does entail the falsity of (1) and its consequent. But then what (1) says is true: If there are elephants, it must be false! Hence if (1) is false, it is also true.

Consequently, if there are elephants and (1) is true then (1) is false, and if there are elephants and (1) is false then (1) is true. Unless we're prepared to deny the existence of elephants, we have a paradox.

Does moral realism entail moral verificationism?

First, I would like to restrict this discussion to duties or responsibilities, as opposed to good or bad states of affairs, for it seems plausible to me that states of affairs may be good or bad even if no one has a duty to bring them about or prevent their occurrence. For the purposes of this post, then, I will understand moral realism as the thesis that people really do have duties to each other.

Consider the following principle:

(1) Necessarily, if someone has a duty to do something, it is possible for them to find out or discover that they have a duty to do it.

(I say ‘find out or discover’ instead of ‘know’ to rule out externalist analyses of knowledge. This is because I think one can know one has a duty in an externalist sense even if one is unable to “tell from the inside” whether one has it. In my view this sort of knowledge of one’s duties is as good as none, for one cannot act on the basis of it.)

I think (1) is highly plausible. To see why I think so, consider what things would be like if (1) were false: If (1) were false, it would be possible for you to have a duty to do something and nevertheless not have even the slightest suspicion that this was so. And if, furthermore, you did not fulfill that duty, you would be morally responsible for violating it. Of course you would be completely ignorant of this—and so would everyone else, otherwise it would be possible for them to show you that you were in fact responsible (assuming the same evidence is available to all). No matter what you or anyone else did, you could never come any closer to discovering your duties than making a lucky guess.

I find this result intolerable, and I suspect many would agree. If the intuition supporting my view is correct, one’s duties must, in principle, be discoverable by one. From this it follows that moral skepticism—the idea that one could have duties that one could not discover one had—is incoherent; there cannot be any such thing as an undiscoverable duty. The only available options, then, are a form of moral verificationism and a form of moral nihilism: Either one has duties which are such that it is possible for one to find out that one has them, or one has no duties at all. So moral realism, understood as the thesis that people really do have duties to each other, entails that moral verificationism is true.

God, Explanation and Conditional Decrees

I think one of the deepest questions philosophy can ask is why nature exhibits the order and regularity it does, and one of the most fundamental divisions among philosophers is between those who think it capable of explanation and those who do not. In western philosophy, one of the most common explanations proposed in answer to this question has been that our world is the product of some kind of design or creative intention, usually on the part of some spiritual beings, such as the gods or God. In this post I want to (somewhat briefly) explore the theistic version of this answer.

If there is indeed a God Who is roughly that of traditional theism—that is, an omniscient, omnipotent, omnibenevolent (etc.) being—it would seem that we have, on the face of it, a good explanation for the existence of an ordered, law-governed, and relatively life-friendly universe such as we find ourselves in. As an omnibenevolent being, God would plausibly have the desire to create other, in some respects similar beings with whom God could share God’s love. And as an omniscient and omnipotent being, God would certainly have the knowledge of how to create them and the power to do so. But we must be careful here. For what sort of explanation could one offer for God’s infinitely many creative decisions? As Hume pointed out long ago in his Dialogues Concerning Natural Religion, it would seem no good to explain the order and regularity of the created world by appealing to a mirror image of that world in the Divine Mind, when this image is itself without explanation. Perhaps the answer is that God’s nature dictates that every world which God could create has some minimal amount of order. (I think Duns Scotus believed something like this, but unfortunately I have not yet found a reference.) If so, we would have an explanation for the fact that the universe is ordered without having an explanation for the particular order it has.

We can flesh out the above solution to this problem in terms of the notion of a conditional decree, my term for something which is nicely spelled out by Alexander Pruss in his post “Creation, Aseity, and Providence” over at the Prosblogion. Basically, conditional decrees are of the form “If p then q, and if not-p then r”, where God both permits p to be true and permits not-p to be true, leaving the outcome to chance. If this is so, God could accommodate libertarian free will by willing conditionals of the form “If Scott freely chooses A then x will follow, and if Scott freely chooses B then y will follow, and if Scott freely chooses C then z will follow…” In this way God can plan “for all possible contingencies”, as Pruss puts it, so that even though certain matters are left to chance—such as Scott’s choice between A, B, and C—something which God wants will occur in any case.

If this account of God’s creative plan is correct, we might suppose that God makes the same decrees, both conditional and unconditional, in all possible worlds where God exists. Thus God could decree unconditionally that some contingent things exist, that some of them are to be conscious and intelligent (etc.), while leaving their exact nature unspecified. God could then use conditional decrees to cover all the matters that are left unsettled by God’s unconditional decrees. On this supposition we do not have to explain how it comes about that God wills one thing when God might have willed something different, while at the same time holding that the existence of God is compatible with a wide variety of different possible worlds. We also wouldn’t have to worry about God’s decrees depending on God’s foreknowledge of what will actually happen. Thus God’s will—that is, the totality of God’s decrees—supervenes on God’s nature, for God’s nature does not vary from world to world. God’s will constrains reality without determining it in every detail, though nothing occurs without God’s permission. Thomas Aquinas once said that it does not follow from the fact that certain things change that God’s will changes, only that God wills that things should change (Summa Theologiae (A.k.a. Summa Theologica), Prima Pars, Question 19, Article 7). In much the same spirit, we can say that from the fact that certain things happen differently in different possible worlds, it does not follow that God’s will differs in different possible worlds, only that God wills that certain things should happen differently in different ones.

Rethinking the limits of dialetheism

In a previous post I tried to undercut one of the main motivations for believing in dialetheism by giving the following argument:

Consider for a moment (2):

(2) This statement has the same truth value as “0 = 1”.

Assume (2) is false. If so, it must have a different truth value than “0 = 1”, for what (2) says is that they have the same value. Since “0 = 1” is false, (2), if it has a different value, must be true. But if (2) is true, it has the same truth value as “0 = 1”, for that they have the same truth value is precisely what (2) says. Now if (2) is true, and it has the same truth value as “0 = 1”, then “0 = 1” must also be true, and hence we can conclude that 0 = 1!

We cannot give (2) a dialetheic treatment—holding that it is both true and false— for we can substitute any falsehood we like for “0 = 1” and use the paradox to show it must be true as well as false. We would then end up with trivialism—the view everything is both true and false! Since (2) cannot be solved by dialetheic means, it must have a different, consistent solution.

I now think this argument doesn’t work. A dialetheist can simply say that (2) is both true and false while “0 = 1” is false only, because it is the statements’ conjoint falsity which accounts for the truth of (2), and not their conjoint truth. After all, it would appear that any statement of the form “p has the same truth value as q”, where q is both true and false, is itself both true and false, but surely this does not entail that p is both true and false. For example, one can easily generate statements of this form by substituting a Liar statement for q and an arbitrary statement for p, but as long as one rejects explosion—the principle that contradictions entail everything—this gives us no reason at all to think that p must also be both true and false. It may have taken me a year, but at least I caught my own mistake. :-P