A Cost of Contingentism
July 22, 2011 — 11:43

Author: Josh Rausmussen  Category: Uncategorized  Tags: , , ,   Comments: 26

Let Contingentism be the thesis that no concrete thing must exist. Define ‘concrete thing’ as anything that can cause something, or leave it as primitive. (Side note: Contingentism is hotly debated among philosophers of religion. But surely it is a thesis of metaphysics; so why aren’t metaphysicians debating this?)
Arguments against Contingentism typically take the following form:
1. Every fact of type T has an explanation (else: is explicable)
2. If Contingentism is true, then there is a fact of type T that has no explanation (else: is not explicable)
3. Contingentism is not true.
Committed Contingentists usually either end up denying the principle of explanation employed by (1) or withholding judgment. After all, such explanatory principles tend to be very far-reaching.
But here’s another strategy. We count costs. Rather than searching for sound philosophical arguments for/against Contignentism, we identify costs and benefits of Contingentism. That may be a lot easier. And it can help us make progress without having to make converts: for a committed contingentist can, in principle, come to agree that there are certain costs of Contingentism.
I’m going to propose one cost–to get this strategy started. (I do not claim this is the most serious cost, or that there aren’t counter-costs that ultimately outweigh it.)

The cost is that there is a fact that cannot have an explanation if Contingentism is true but can have an explanation if Contingentism is not true. Hence, Contingentism implies a mystery that its denial would resolve. And that’s a cost–be it big or small.
Which fact? I think there are many, actually. Here’s one: the fact C that the c’s exist at all, where ‘the c’s’ plurally (and rigidly) refers to all contingent concrete things. Unless we allow circular explanations, it seems that an explanation of C would be at least in part in terms of something other than one of more of the c’s (no matter how many c’s there are). But Contingentism does not allow for the existence of anything other than the c’s. Thus, Contingentism doesn’t seem to allow C to have an explanation.
On the other hand, if Contingentism is false, then we could suppose that C is explained by a causal chain that is headed by one or more concrete things whose existence is explained by the fact that it/they must exist (for instance).
Now someone might quibble with my example. (There’s the classic objection that the fact that the c’s exist could somehow be explained by the fact that each is explained by another, ad infinitum… (but would that explain why those c’s exist at all?) … I’ve also encountered the idea that contingent things could be explained in some sense by their having a nature that makes them indestructible and beginningless…) But then shift to a different example: let E be the fact that being a contingent concrete thing is exemplified. It seems that no facts about contingent concrete thing(s) could explain E without circularity, and surely E isn’t explained by virtue of being necessary (for surely there could be no contingent concrete things: think principle of recombination).
Here are a some additional candidate examples: (i) the fact that the c’s have the total space-time geometry they do; (ii) the fact that there are n contingent things, where ‘n’ names the number of contingent things there are; (and for fun (iii) the fact that the c’s possibly exist [if Alex is right that a thing possibly exists only if there’s something has the capacity to head a causal chain that leads to its existence]).
One can of course suppose that none of these facts have an explanation. But since we can explain those facts (in straightforward, non-gerrimandered ways) if Contingentism is false, it seems there’s a cost to holding Contingentism.
Of course, admitting necessary concreta could be even more costly. I have only argued that Contingentism has a certain cost, and perhaps even committed contingentists can agree.

  • What is the “straightforward, non-gerrimandered” explanation available to the non-contingentist for the fact that there are n contingent things?

    July 22, 2011 — 14:16
  • Joshua Rasmussen

    Perhaps this is one: there are necessarily existing things that caused there to be n contingent things? Or: there is a causal chain headed by non-c’s that resulted in the existence of n c’s?

    July 22, 2011 — 14:31
  • Take a fundamentally probabilistic process with a probability distribution over the number of things it produces (an atomic decay, say). Whether such a process is initiated by a necessarily existing thing or not seems to me irrelevant to the quality of the explanation we have concerning how many things it produced.
    (Note: I am assuming that our explanatory question is something like, “why are there n rather than n+1 things”, and not “why are there n things rather than no things at all”).

    July 22, 2011 — 14:42
  • I feel like there is something in there I want to disagree with, but I’m not entirely sure what all the C’s, T’s, E’s, etc. mean!

    July 22, 2011 — 14:46
  • Joshua Rasmussen

    We could switch to a different fact. But let me see if I understand your proposal. To simplify things, suppose there are just two c’s, A and B. Suppose B exists because of a probabilistic process initiated by A (or would this process be identical to A since there are just two c’s?). That would seem to explain why B exists, but I’m not sure how a probabilistic process would explain why there are two c’s. And the same goes for any number…

    July 22, 2011 — 15:15
  • Mike Almeida

    I had a post related to this here http://movabletype.ektopos.com/cgi-bin/mt-search.cgi?search=hell+worlds&IncludeBlogs=3&limit=20. You can suppose that God is contingent and also such that, if it is possible that God exists, then he does exist. So God is a special kind of contingent object whose possiblity entails his existence. So, God has an explanation, in worlds in which he exists (i.e., his possibility entails that his actuality) and God explains every other concrete thing. I have something on this written up somewhere. You do have to give up S5 in favor of S4, and you do have to assume that there are nihilistic worlds (worlds in which there are no concrete objects at all) and finally you have to assume that nihilistic worlds see each other, but no other world. So God necessarily does not exist in nihilistic worlds, but he does not necessarily exist in -non-nihilistic worlds. In other words, it is a theorem that God is either possible and contingent or impossible. Anyway, for what it’s worth.

    July 22, 2011 — 15:35
  • Justin

    On your side note: Metaphysicians have debated Metaphysical Nihilism (see, for instance, these papers: http://www.jstor.org/pss/20013390 ; http://www.jstor.org/pss/10.2307/3328136?mlt=true). While your Contingentism doesn’t entail Metaphysical Nihilism, they are closely related.

    July 22, 2011 — 15:52
  • If I may question the apparently shared presuppositions of this thread…I have my doubts that the expressions Joshua offers really denote facts in the first place: “that the c’s exist at all, where ‘the c’s’ plurally (and rigidly) refers to all contingent concrete things,” “that the c’s have the total space-time geometry they do,” “that there are n contingent things, where ‘n’ names the number of contingent things there are,” etc. If we replace those expressions with ones denoting facts that are determinate enough to be explained at all, I think we’ll find that they denote facts there’s every reason to think can be explained naturalistically. Details here:
    (I welcome feedback. Email is best, since I don’t have a comments feed from Prosblogion. Cheers.)

    July 22, 2011 — 21:07
  • Joshua Rasmussen

    Thanks for those links. I read the first one and enjoyed it much. It is notable that the author doesn’t discuss any traditional or contemporary arguments for a necessary concrete thing (which would of course be arguments against Metaphysical Nihilism). This may be another reason for arguments for a necessary being to be recognized within metaphysics.

    July 23, 2011 — 10:21
  • Joshua Rasmussen

    Steve, many thanks for the link to your paper. I read it with interest. I’ll email you some questions/comments.

    July 23, 2011 — 14:46
  • Joshua Rasmussen

    Of course, if there is no fact that there are contingent things, then Contingentism wouldn’t be a fact either.

    July 23, 2011 — 14:53
  • “There are concrete, contingent things” is an abstraction from various particular facts, such as that Mars exists, that Tibbles the cat exists, and so on. The abstraction can tempt people to pose a spurious challenge to naturalism: “Natural science may be able to explain why Mars and Tibbles, and so on, exist, but it can’t explain why any such things exist in the first place.” I think this spurious challenge is at the heart of the Cosmological Argument. Many folks seem to hold that “concrete, contingent thing” denotes a category whose non-emptiness needs explaining above and beyond the explanations available for Mars, Tibbles, and so on. I think that’s a mistake.

    July 24, 2011 — 6:54
  • Joshua,
    Let me try again. Suppose we expand your example a little so that the probabilistic process in question might have produced either 1 or 2 things, leaving us with either 2 or 3 things total. In fact, it produced 1 thing leaving us with 2 things total. My claim is that the quality of the explanation we have for why there are 2 things rather than 3 things is independent of whether A exists necessarily.
    This claim is independent of any particular account of probabilistic explanation. If you think it is unclear how we have any explanation at all for why the process produced 1 rather than 2 things, then you should be equally unclear why we have 2 rather than 3 things overall, whether or not A exists necessarily.

    July 24, 2011 — 9:09
  • Joshua Rasmussen

    In response to Steve’s interesting paper, I suggest we switch our attention to the fact that any instances of a “genuine sortal” are contingent. (Surely, if we can ask such things as why is anyone without a pinky, we can ask why any instance of a sortal is contingent.)
    the quality of the explanation we have for why there are 2 things rather than 3 things
    It’s not perfectly clear to me what explanation you have in mind here. But suppose you are right (for the sake of argument). You would agree that Contingentism doesn’t afford an explanation of the fact that there are contingent things, no?

    July 25, 2011 — 9:29
  • Joshua,
    I’m going to stick to the original point of my comment for now, since I still seem to be having a hard time making it clear.
    Let’s build a little more on your example. Consider two possible worlds W1 and W2 in which the only concrete existents are A and B, and B is produced by a fundamentally probabilistic process initiated by A that might also have produced another concrete existent C. Now suppose that the only difference between the two worlds is that in W1, A exists necessarily, while in W2 A exists contingently. And consider the explanatory question, asked of each world, “why are there 2 things rather than 3 things?”.
    So far I have simply claimed that there is no difference between these two worlds capable of making a difference to the quality of the answers to this question. As I said earlier, this claim doesn’t require me to say what the explanation actually is, in either world. Your initial claim, if I understood you correctly, entails that there is an explanatory difference between these two worlds. If you think so, the burden is on you to identify it.

    July 25, 2011 — 11:26
  • Joshua Rasmussen

    Thanks, Brad, for pressing me on this.
    I think I can see how to give an explanation in W1 (of why there are n c’s): A includes the necessarily existent concreta and heads a causal chain from which 1 c is produced.
    I’m just not presently sure how the explanation would go in W2. I’m very open to suggestions. (I anticipate you or someone will persuade me.)

    July 25, 2011 — 13:40
  • If I understand you correctly, the explanation you propose for W1 is:
    A exists necessarily and caused one thing.
    Here is an explanation I claim is at least as good:
    A exists and caused one thing.
    This explanation is also available for W2.

    July 25, 2011 — 14:15
  • Joshua Rasmussen

    A exists and caused one thing.
    Hmm. I would have thought this doesn’t work on the grounds that it doesn’t even make the explanandum likely (for surely A could exist and cause one thing even if there were say 55 things).
    As far as I see, the same problem only arises for “A exists necessarily and caused one thing” if there can be causal chains that don’t terminate in A (you did stipulate A is the only necessary thing in W). But I’m not sure there could be such causal chains, and anyway, if there could be, then presumably their reality would be without explanation. What do you think?

    July 25, 2011 — 15:10
  • I don’t see anything here that discriminates between the cases. Any complete explanation for the total number of concrete things presumably requires, in addition to the particular explanations for the individual concrete things, a proposition to the effect that those are all the things there are. For presumably for most worlds, there is some distinct world that includes the (counterparts of the) existents of that world as a subset. We’ve both been leaving this implicit. You’ve now made it explicit, but it seems to me independent of whether or not some of the existing things are necessary.
    In our example, I see zero reason to think there couldn’t be worlds with other existents that don’t trace back to A. Hence, whether or not these other existents could be explained were they to exist, the problem you identify for the explanations afflicts both W1 and W2 above.
    PS. When I said “possible worlds” in my earlier setup of the example, I should have said “epistemically possible worlds”.

    July 25, 2011 — 15:48
  • Joshua Rasmussen

    I see zero reason to think there couldn’t be worlds with other existents that don’t trace back to A.
    1. Well, there may be the cost of an unexplained causal chain. 🙂
    2. But ok, suppose (in the epistemically possible situation in which there is one necessary concrete thing A) we have no reason to think there couldn’t be causal chains that don’t terminate in A if A is necessarily existent. Do we have reason to think there could be? If not, then perhaps we can try this explanatory hypothesis: “A is necessarily existent, A caused one contingent thing, and necessarily, A heads every causal chain.”
    The parity principle (for contingent A) would presumably be this: “A caused one contingent thing, and necessarily A heads every causal chain.” But don’t we have good reasons to doubt this is possible–because, for instance, an intrinsic duplicate of A might have headed the chain instead?
    3. I have this feeling that in general we cannot explain why there are n instances of a kind k merely by reference to instances of k–or at least that any explanation wholly in terms of k’s would be less satisfying than an explanation that runs outside the k’s. But these are deep waters, and I appreciate your parity strategy. Your having different sentiments about this makes me more cautious.

    July 25, 2011 — 20:31
  • Andrew

    “Here is an explanation I claim is at least as good:
    A exists and caused one thing.
    This explanation is also available for W2.”
    Suppose that there are just two things: A and B (the thing caused to exist by A). If we assume that A is a contingent thing and we’re looking for an explanation of the fact that there are 2 contingent things, then I can’t see how the fact that A exists and A caused B is a good explanation. Wouldn’t we need an explanation that explains the existence of both A and B?

    July 26, 2011 — 14:16
  • Andrew: I’ve been assuming that the explanatory question is why there are two rather than three things, not why there are two rather than zero things (see my second comment in this thread).
    Joshua: I take it you are joking in (1) (presumably explanatory considerations tell us which possibilities to believe, not which possibilities there are). I don’t really understand your point in (2). For what it’s worth, I do think it is reasonable to believe that for any concrete existent, it is metaphysically possible that there exist causal chains not headed by (counterparts of) that existent, regardless of whether that existent is necessary or contingent. I can’t see any metaphysical principle worthy of believing that would help deny this, but maybe you have one up your sleeve. Regarding (3), I agree that those are deep waters, which it seems Steve’s paper attempts to address. But the point is moot if we don’t have an explanation at hand of the type you desire, and I haven’t seen it yet.

    July 27, 2011 — 12:35
  • Joshua Rasmussen

    Brad, good considerations. Here’s some thoughts:
    1. It seems to me that explanatory considerations could, at least in principle, count against there being certain possibilities: for example, if there aren’t unexplained causal chains, then a simple explanation of that fact might be that there can’t be (and there may a cost in thinking there are unexplained causal chains).
    2. The explanation I gave in (2) can, I think, break parity even if one has no reason to think that necessarily, A heads every chain, as long as one also has no reason to think deny that (because one does have a reason to deny the analogous claim where A is contingent). Does that seem right to you (even if you don’t meet the “as long as” condition)?
    3. Not sure it makes a difference, but I had in mind the question “why are there n contingent things?” (or “why are there 2 things?”; or “why is there 1 contingent thing?”), not “why are there n contingent things rather than n+1”? The intuition Andrew expresses appears to be an instance of the feeling I expressed in (3). If you lack the intuition, that’s cool; I respect that. 🙂
    (But to make sure the intuition I expressed is clear, here’s an example using a kind other than Contingent Instance. Suppose there are 2 emeralds. My feeling is that we can’t explain why there are 2 emeralds [rather than 3] by reference to nothing but emeralds (e.g., that emerald A indeterminsically caused one emerald rather than two and nothing else caused any emeralds doesn’t suffice). On the other hand, we could explain why there are 2 emeralds if part of the explanation were in terms of things other than emeralds. If you don’t share that feeling, fair enough–we’ve still made good progress.)

    July 28, 2011 — 10:54
  • Steve:
    That’s a neat paper.
    The crucial assumption in your paper is that certain kinds of fundamental metaphysics are in principle impossible. I reject that assumption. 🙂 It seems perfectly reasonable to say that the questions you ask about how many objects lie on your hand have perfectly well-defined answers that simply depend on the correct physics and metaphysics. It seems that the difficulty is merely epistemic: for instance, we may not know whether among the things that exist there are cross-sections (I myself am pretty sure there are no cross-sections, unless by accident there happens to be a simple there which “counts as” a cross-section).
    More seriously, I do think that you need to worry more about the reformulation of the question in terms of sorts once you grant, as you do, that it is possible to quantify over sorts. You have to deny one of the following:
    1. One can quantify over sorts.
    2. “IsInstanced(s)” is a wff.
    3. Standard sentence formation rules hold.
    (For if you accept 1-3, then “(Es)(IsInstanced(s))” will make sense.)
    Each of these is costly to deny. (I happen to deny 3, and I am friendly to denying 1, but I do this for other reasons.) For instance if you deny 1, then a lot of stuff in the paper becomes unsayable. If you deny 2, then you have to either deny that there are meaningful sentences like “Dog is instanced” or you have to deny existential generalization (which would take us from “Dog is instanced” to “Some sort is instanced”). Likewise, you then can’t say seemingly perfectly sensible claims like “New scientific theories sometimes posit a sort S whose instances enter into causal explanations.” 🙂
    If you deny 3, then more needs to be said about how to do that.

    August 1, 2011 — 20:58
  • Alex:
    Thanks for the compliment and the comments. I’ll add them to the growing list I need to address when I follow up the paper, as I hope to have the chance to do. I don’t see why I need to deny any of 1-3, but I’ll think further about it. Maybe I need to distinguish between a question’s being sayable and its being conceptually unconfused. “What if the universe and every physical thing in it doubled in size?” is sayable but, in my view, conceptually confused and so in that sense meaningless. I think likewise about many cosmological questions as they’re usually intended.

    August 5, 2011 — 6:41
  • Steve:
    I think that for any sentence s that is true, the question “Why s?” is not conceptually confused. If the PSR is false, the question may not have an answer, but not having an answer is not the same as its being conceptually confused.
    Now, if the sentence “(Es)(IsInstanced(s))” makes sense, it is surely true. After all, we have the perfectly good FOL argument:
    1. IsInstanced(Horse)
    2. (Es)(IsInstanced(s)) (by Existential Generalization).
    And if it is true, then the question of why it is true should make sense, no?

    August 5, 2011 — 16:28