Three Responses to the Argument from Contingency
August 28, 2013 — 13:53

Author: Kenny Pearce  Category: Existence of God  Tags: , , , ,   Comments: 61

In my view, the cosmological argument from contingency is the most powerful philosophical argument for the existence of God. By a ‘philosophical’ argument, in this context, I mean a way of giving reasons for something that does not depend on detailed empirical investigation, or on idiosyncratic features of a particular individual’s experience or psychology. Thus I do not hold that the argument from contingency is the best reason anyone has for believing in God. I think, for instance, that some people have had religious experiences which provide them with stronger reasons than the argument from contingency could, even making very generous assumptions about their ability to grasp the argument.
This belief of mine, about the relative strengths of the arguments, goes rather against the grain of current discussions. I think the recent philosophy of religion literature overestimates the power of the fine-tuning argument and the first cause argument, and underestimates the power of the argument from contingency.

(A couple of notes: first, I recognize that a great many philosophers have a very dim view of the first cause argument. Nevertheless many of them have a higher opinion of that argument than I do – or so it appears to me. Second, it is not perfectly clear that the fine-tuning argument satisfies my definition of a ‘philosophical’ argument above, since it depends on the discoveries of modern physics. However, it depends only on very general features of physical theory, and not on detailed examination of some particular class of phenomena, so I’m counting it as ‘philosophical.’)
Despite thinking the argument from contingency is a very powerful argument (as philosophical arguments go — I find myself in general agreement with Keith DeRose’s remarks about the limitations of philosophical arguments), I think there are three distinct credible responses to it, and that’s what I want to discuss here. The three credible responses are: (1) accepting the existence of a God-like necessary being; (2) rejecting the Principle of Sufficient Reason; or (3) holding some facts/entities discovered by physics to be necessary. The first two of these options have of course been much discussed, but the last one is less often addressed in work on the argument from contingency. Let me first lay out a version of the argument, and then discuss each of the responses.
A fairly standard version of the argument could go as follows:

  1. Every contingent fact has an explanation (PSR).
  2. It is a contingent fact that some contingent beings exist. Call this fact ‘(E)’.
  3. Therefore, (E) has an explanation (from PSR and 2).
  4. On pain of circularity, no contingent fact or contingent being could explain (E).
  5. If some contingent fact has an explanation, but is not explained by any contingent fact or contingent being, then it must be explained in terms of the free choice of a necessary being.
  6. Therefore, (E) must be explained in terms of the free choice of a necessary being (from 3-5).
  7. If some fact is explained in terms of the free choice of a necessary being, then there exists a necessary being who chooses freely.
  8. Therefore, there exists a necessary being who chooses freely, and this everyone calls ‘God’ (from 6 and 7).

Note first that this version does not require a ‘Big Conjunctive Contingent Fact,’ and hence need not get into any of the technicalities or weirdness of infinitary logic. Now let me quickly dismiss two responses I don’t regard as credible. The first response is the necessitarian denial of (2). I take it hardly anyone wants to take this route. The second is to deny (7) by holding that false propositions can be explanatory. This surprising thesis is ably defended by Nancy Cartwright, and perhaps there are some contexts in which Cartwright’s thesis is correct, but I take it to be quite clear that there is a metaphysical notion of explanation for which (7) holds. Cartwright (or a follower of Cartwright) might perhaps deny that every contingent fact has an explanation of this sort, while holding that every contingent fact has an explanation (or, more likely, many equally good explanations) of her sort. This response I do take to be credible, but, since the PSR, as I understand it, is about metaphysical explanations of the sort that validate (7), I’ll count this kind of response as a denial of the PSR.
Now, one way of responding to this argument is, of course, to accept it. As I have indicated, this is the response I favor, though I should note that I don’t want to commit too strongly to the technical adequacy of a hastily-drawn-up blog post version of the argument. There may be some technical difficulties, for instance, about (4). One might think that the fact that Kenny exists explains why some contingent beings exist in the sense that the former grounds or truth-makes the latter. This, though, is not the kind of explanation we’re going for. (There are many kinds of explanation.) The fundamental intuition is that ‘why is there something (contingent) rather than nothing (contingent)?’ is a perfectly well-formed question, and it makes sense to look for an answer to it, and the free choice of a necessary being is, uniquely, able to provide an answer to this question without rendering the contingent necessary. This seems (intuitively) quite right to me, and none of the critiques I’ve read have convinced me otherwise.
On the other hand, as I have said, there are two other possible responses which I take to be credible. The first of these is the rejection of the PSR. The leading reason given for rejecting the PSR is that it leads to necessitarianism. Those who take this response would hold that premise (5) of our argument is only vacuously true, since nothing necessary can explain anything contingent. There are some difficult issues here in the theory of explanation, but the view that, in correct explanations, the explanans must entail the explanandum is now widely regarded as false, so this general principle can’t be relied on here. But there is some kind of demand for explanation (something the proponent of the argument should be very sensitive to!): if the explanans is true in all worlds, then why is the explanandum false in some? Thus, for instance, if whatever accounts for the existence of contingent beings is also true at those worlds at which there are no continent beings, then why are there no contingent beings at those worlds? The appeal to free choice is supposed to answer this question, and if agent-causal libertarianism is coherent, then it’s pretty easy to see how the answer goes (see Pruss). But the coherence of agent-causal libertarianism is subject to much dispute, and it is unclear whether the problem is solved on alternative accounts of free choice.
This leads into the third response. Even if agent-causal libertarianism is incoherent, there is another kind of non-necessitating explanation which is now much more widely accepted: explanation in terms of indeterministic physical laws. This could provide an alternative method for getting from necessity to contingency, consistent with the PSR.
This kind of approach is suggested by last year’s debate between David Albert and Lawrence Krauss following Albert’s review of Krauss’s book. Albert insisted that Krauss had not explained why there was something rather than nothing because all Krauss had done was explain how, if we started from a state in which all of the quantum fields had the value zero everywhere, we would end up in a state where there is ‘something’ — that is, where the fields have the sorts of local spikes we call ‘particles.’ Albert’s objection was that explaining why the quantum fields are not everywhere zero is not explaining why there are quantum fields at all. Krauss’s response seems to be that the notion of ‘nothing’ with which Albert is working is wrong; that physics has discovered what nothing really is, just as chemistry has discovered what water really is, and, as it turned out, nothing is just all the fields being zero everywhere. It seems to me that a more or less equivalent (but more philosophically perspicuous) way of putting this claim would be to say that it is necessary that the quantum fields exist and obey the laws of physics, but contingent that there are any particles. Krauss can then show how, by indeterministic processes, particles are likely to arise from ‘nothing,’ i.e. from the absence of any particles. But note that this involves the existence of necessary beings: the quantum fields. It also involves making the laws of physics necessary; the laws of physics may or may not count as ‘beings.’
As I said, I take all three of these responses to be credible, by which I mean that the only way to decide between them are by the sorts of broad considerations that motivate theory choice. It seems to me that classical metaphysical theology – the sort of metaphysical system-building centered around a broadly Anselmian notion of God that we find, for instance, in Aquinas and Leibniz – exhibits, in a very high degree, the sorts of virtues we look for in metaphysical theories. It further seems to me that many of the people who take the physics approach do so out of some kind of general anti-metaphysical sentiment, which suggests that maybe the positivists were right: metaphysics, even analytic metaphysics, is a ‘gateway’ to theology. But – and this is my central point – this would have to be shown by a thorough comparison of the theoretical virtues of necessary being theology as compared with the necessity of physics.
(Cross-posted at

  • Steve Maitzen

    Thanks, Kenny, for your post. You wrote: “The fundamental intuition is that ‘why is there something (contingent) rather than nothing (contingent)?’ is a perfectly well-formed question, and it makes sense to look for an answer to it….”
    Can I ask if you regard “contingent” as denoting a kind to which some things belong rather than as merely a predicate that some things satisfy? I think arguments from contingency are committed to the former, and I think the former is problematic. If the distinction I have in mind isn’t clear, let me know, and I can try to clarify.

    August 28, 2013 — 15:15
  • Kenny Pearce

    Steve – I don’t think I am committed to ‘contingent’ denoting a natural kind. If you like you can think of the phrase ‘contingent thing’ as abbreviating the phrase: ‘a thing which is such that the de re proposition which says, of that thing, that it exists is a contingent proposition.’ And, of course, a contingent proposition is one which is possibly true and possibly false.
    Does understanding ‘contingent thing’ in this way undermine the argument in some way I’m not seeing?

    August 28, 2013 — 15:24
  • Steve Maitzen

    My comment could have been clearer: I think premise (4) of your argument depends on the assumption that all contingent beings, as such, form a kind (not a natural kind, but a kind nonetheless). I’m afraid the best I can do to explain why, and why I think it’s a problem, is to point you to this recent paper. See in particular my replies to Objections C and D.

    August 28, 2013 — 16:00
  • Steve Maitzen
    August 28, 2013 — 16:02
  • Very interesting. As I recall, from my days as a philosophy major, many, many years ago, Frederick Copleston (I’m not sure I’ve got that spelled right), S.J., relied heavily on the argument from contingency in a published debate with Bertrand Russell. Needless to day, he didn’t convince, but certainly held his own.

    August 28, 2013 — 16:36
  • Kenny Pearce

    I see. I will have to read your paper when I have time. It is not at all obvious to me that (4) has any such presupposition. Indeed, I am not sure what the presupposition is supposed to be. After all, on standard views, the abundant universals are as abundant as the meaningful predicates, so if by ‘kind’ you mean something weaker than a natural kind (sparse universal) but stronger than having in common a meaningful predicate, then I don’t know what you mean.
    So, as I said, your claim does not seem at all obvious (or clear) to me. But then, interesting papers defend non-obvious theses! 🙂

    August 28, 2013 — 16:37
  • Very interesting. As I recall, from my days as a philosophy major, many, many years ago, Frederick Copleston (I’m not sure I’ve got that spelled right), S.J., relied heavily on the argument from contingency in a published debate with Bertrand Russell. Needless to day, he didn’t convince, but certainly held his own.

    August 28, 2013 — 16:41
  • Thanks Kenny,
    I have a similar interest in Leibnizian style arguments from contingency. My hunch is that these arguments can still be sound if they appeal to a more modest version of the PSR than ones which apply to contingent facts in any possible world in which they are true. What I have in mind is a version of the PSR which stems from the success of abductive inferences in various disciplines such as the natural and social sciences, ethics, metaphysics, and everyday experience. It seems to me that abduction operates on a very basic assumption: that if there are explanatory gains to be had in positing an explanation for a contingent being, then there is prima facie warrant for doing so. I will call this version the Abductive Principle of Sufficient Reason (or APSR)
    The APSR has advantages and drawbacks. On advantage is that it is easier to defend because it is compatible with the possibility that in some possible world, a contingent being simply exists as a brute fact, without explanation. A disadvantage is that reality need not conform to our epistemic demands for explanations of the abductive sort. But until we have good reason to think the APSR does not apply in this case, we seem to be warranted to holding that it does, given how successful it already is in so many other disciplines.
    So if the APSR holds, then if the aggregate of all contingent things is itself contingent (let’s assume), then we have prima facie grounds for positing a necessary being whose existence and causal powers raise the probability that contingent things would exists rather than not (this is just a rough rendition, with missing premises of course!).
    Unfortunately, the problems you mentioned of getting a contingent effect from a necessary cause still apply to this modest version of the contingency argument, though this problem is not insurmountable if an agent’s reasons exist necessarily, but the exercise of power resulting from its reasons only exists in some worlds and not others.
    As for your interpretation of Krauss, that it is necessary that quantum fields exist and obey the laws of physics, and that the laws of physics are necessary – several replies are available here. First, William Rowe defends a thought- experiment such that if God’s existence is possible, then it is also possible that he annihilates quantum fields and the laws governing them in some world, thus making them contingent. Second, we can conceive of counter-to-fact scenarios in which the laws of physics are arranged differently or do not exist at all, so Krauss’ claim is weakened to the extent that conceivability indicates metaphysical possibility. Third, although the current state of cosmology is changing rapidly, there are some persuasive indications (stemming from some lectures I recently watched by Alexander Vilenkin: that the multiverse has an absolute beginning, as well as the quantum vacuum from which it supposedly originated. Now even if these “persuasive indications” are false, they aren’t necessarily false. It is reasonable to believe that the quantum vacuum could have had a beginning, even if, as a matter of fact, it didn’t. Consequently, it is a contingent being which, by APSR, calls for explanation.
    I’ve written a lot here, but I share these comments because a more modest version of the Argument from Contingency may bypass some of the problems you highlight with stronger versions. What do you think?
    Thomas Rauchenstein

    August 28, 2013 — 20:52
  • Kenny Pearce

    It is far from clear that we can conceive of different laws of physics, in the sense of conceivability on which it is a guide to possibility. In fact, I can’t even really conceive of the actual laws of physics – that is, I don’t understand them or see how the phenomena follow from them. Then again, I’m not a physicist.
    I can’t imagine what a world would be like with different laws. What I can do is imagine what it would be like for some event to occur which is prohibited by the actual laws of physics. Even then, though, the world I’m imagining will be similar enough to the actual world, that it might be better described as imagining a violation of the laws than imagining different laws.
    I feel a lot of pull in the direction of the contingency of physics, and some of my reasons for that are the ones you give. This is among the reasons why I think accepting the existence of God is the better course. But I don’t think the whole thing is all that obvious.

    August 28, 2013 — 21:04
  • Eric Steinhart

    I agree with this argument except for one point in (5) “If some contingent fact has an explanation, but is not explained by any contingent fact or contingent being, then it must be explained in terms of the free choice of a necessary being”.
    Where did the “free choice” part come from?
    What’s the argument for that?
    Why not just “explained in terms of a necessary being”? The necessary being could be the Plotinian One, or some Tillichian ground of being, or the Wiccan ultimate deity, or the Form of the Good, or even some axiarchic proposition.
    Or does the “free choice” part have to be slipped in for the sake of theological correctness, since the output of the argument (the Christian God) is already known in advance?

    August 28, 2013 — 21:21
  • Kenny Pearce

    The free choice is supposed to be needed in order to get contingency out of a necessary being. For Plotinus (I think), the world emanates from the One necessarily, so that would be a version of the necessitarian response I mentioned.

    August 28, 2013 — 21:26
  • Eric Steinhart

    It still doesn’t belong in the argument, especially if it’s just a way to reply to an objection to premise (5). My point is that, as stated, the argument looks biased towards the Christian God right from the start.
    Run the argument without the “free choice” phrase; then say that there’s a problem based on “getting contingency out of necessity”; one solution is by positing an agent with free choice; another is some sort of necessitarian response.
    Running the argument that way allows it to function just as an argument involving necessity and contingency; it permits the theological debate to come afterwards, and doesn’t try to shut it down in advance. At the very least, that’s more honest.
    The larger point is this: PoR contains some very deeply interesting arguments, which have been mis-understood as arguments for God. There are interesting arguments for a necessary being, a first cause, a cosmic designer, and a maximally perfect being. But it takes a lot more work to identify all of those outputs with each other, and to say that the result is God.

    August 29, 2013 — 7:09
  • Kenny Pearce

    Eric – On reflection, I agree that it would be more perspicuous to state the argument as an argument for the conclusion that some necessary being must explain contingent beings, and then ask the question how this could be so, at which point it is appropriate to introduce free choice as a possibility. But as far as I can tell, this is merely expository and not substantive. Indeed, I discuss in the post the possibility that considerations along the same lines as this argument could be used to argue that some of the entities of physics are necessary.

    August 29, 2013 — 12:44
  • Eric Steinhart

    Kenny, yes, you do consider the options well, and I’m not accusing you of being unfair. What I’d like you think about is this:
    On logic alone, (5) should read: “If some contingent fact has an explanation, but is not explained by any contingent fact or contingent being, then it must be explained in terms of some non-contingent being”. And it’s arguable that logic alone lets you gloss non-contingent as “necessary”.
    Put that way, the argument is something that both theists and non-theists can appreciate. We can all meet at this argument, as at a kind of common metaphysical sanctuary. Obviously, we’ll develop things differently from there, given our different commitments. But when an argument is stated in a neutral way, it can be more appealing to everybody, and we can work together.
    However, when (5) includes things like “the free choice of”, it immediately poisons the conversation. The non-theist has to fight it, and it looks like the non-theist hasn’t even been given a fair chance. It just makes for needless conflict. For any non-theist, who is constantly feeling bullied by theism, it’s far from “merely expository and not substantive”. If PoR is about truth-seeking, and not merely Christian apologetics, then it is substantive indeed (especially for people in PoR who are not Christians).
    It’s far more honest, and charitable, to put the shared antecedent logic first, and then to attach specific sectarian interpretations (and then it’s also entirely fine to attach those sectarian interpretations, if they are plainly displayed).
    – Eric

    August 29, 2013 — 13:10
  • Kenny Pearce

    But (5) is a premise. I don’t know what it means to say “On logic alone, (5) should read…” since I’m not claiming that (5) follows logically from anything I said earlier. As far as I can tell ‘logic alone’ doesn’t have anything to say about how (5) should read.

    August 29, 2013 — 13:17
  • Eric Steinhart

    (1) x’s have explanation;
    (2) If x’s are not explained by Fs then they are explained by non-Fs.
    (3) Therefore: x’s are explained by non-Fs.
    I take it that (2) is purely analytic; it is true by logic alone.

    August 29, 2013 — 14:06
  • Kenny Pearce

    That seems right. So a version of the argument could be written which would replace (5) (in my version, a substantive premise) with a tautology, and this version would come to a weaker (less God-like) conclusion.

    August 29, 2013 — 14:42
  • Eric Steinhart

    Exactly! And this new argument would just be for a necessary being, which somehow provides the explanation for the contingent beings. And it’s a far more interesting and appealing argument now. Now there’s real philosophical work to be done: what is the nature of this necessary being?

    August 29, 2013 — 15:27
  • Mark Rogers

    What is the nature of this necessary being you ask? Well, assuming this being ever was and still is in existence the best I can do is perhaps note an attribute of this being. God does display hiddeness. You can not by your own efforts find him but some people believe God has found them. So the question becomes what is this foundedness based upon. Can we then place ourselves in a position that will increase the likelihood that God will find us? Let’s think about what Cleanthes in DNR III has to say.
    Consider, anatomize the eye; survey its structure and contrivance, and tell me, from your own feeling, if the idea of a contriver does not immediately flow in upon you with a force like that of sensation (DNR III).
    Cleanthes seems to think that upon reflecting on the order of nature a special something may happen. I notice that the idea of a contriver does not spring from within but rather it flows in upon Cleanthes. It is also associated with a force like that of sensation. So to have a belief in a designer would seem to imply not just a sense that a designer is possible, but in this instance that the idea of the designer would spontaneously flow into to you, in addition it comes with a force like that of a sensation. That sensation which points to the veridicality of an intersection with deity, along with the spontaneity of the idea of the contriver may lead to belief. So while some may say the reflection on nature is an external trigger of belief, without the external affirmation from deity I think this belief would be shaky indeed. That God, if God exists, displays hiddenness seems undeniable as an attribute. It also implies that God, if God exists, is coming and going from a place by our own free will we cannot go. Any intersection with this neccesary being is contingent on a first action by this neccesary being.

    August 31, 2013 — 9:00
  • Joshua Rasmussen

    Interesting post, Kenny. I have two comments.
    1.Perhaps one way one may completely avoid Steve’s worry is to work instead with the “plural” contingent fact C that there are [tenseless] the actual contingent concrete things (else: the fact that there are the contingent concrete things and the causal connections between them). I can’t see how that fact could be explained without the pain of an exactly circular explanation (where a fact f explains f).
    (Note that C isn’t a big conjunctive fact. So there is no prospect of explaining C by its conjuncts. Besides, we may ask what explain the plurality of conjuncts…)
    2. It seems to me that one can motivate the premise that E has an explanation without PSR or even a restricted version of PSR. Consider an argument along these lines:
    A. *Some* contingent facts have an explanation.
    B. E is not relevantly different from them. (This premise gets into tricky issues, but I think it’s defensible.)
    C. If E is not relevantly different from facts that have an explanation, then E has an explanation.
    D. Therefore, E has an explanation.
    Anyway, it’s another option worth considering, since there is no “deny psr” response.

    September 1, 2013 — 1:06
  • Eric Steinhart

    Joshua – That looks like a promising strategy for avoiding worries about PSR. – Eric

    September 1, 2013 — 9:10
  • Josh:
    Regarding your argument, how about this argument against B. It’s easy to argue that E isn’t intrinsically different from the contingent facts that have an explanation. But we are in a different epistemic position with regard to E: unlike in the case of the other contingent facts, we know that E has no explanation in terms of contingent facts. This is an epistemic difference between E and the other facts, and that’s enough to produce a defeater to the inference to an explanation of E.
    Let O be the first organism on earth.
    A. Some organisms on earth have an ancestor.
    B. O is not relevantly different from them.
    C. If O is not relevantly different from organisms that have an ancestor, then O has an ancestor.
    D. Therefore, O has an ancestor.
    E. Therefore, there was life outside of earth and earth life descended from it.
    Now, E may be true, but the above argument gives us only weak reason to think E to be true. The right response is that while O may not be relevantly intrinsically different from the organisms we know to have an ancestor (one might cavil: O is presumably very simple), O is epistemically different: we know O has no earthly ancestor, while we do not know this for other organisms.

    September 1, 2013 — 9:15
  • Tyron

    Hi Kenny,
    I hope such self-promotion isn’t very rude–but I think you might find the very recent volume I edited of interest:
    That’s the volume Steve’s paper is in. And there’s another cool paper there on whether the question is well-formed by Kris McDaniel. Since I see you’re at USC–Shieva Kleinschmidt and Jake Ross have some stuff in it on the PSR. Here’s the full TOC:

    September 1, 2013 — 14:42
  • Kenny Pearce

    Tyron – Thanks for pointing this out. (No, self-promotion isn’t rude when it’s relevant!) I read the draft of Shieva’s paper recently but I hadn’t heard anything about the book in which it was included. This looks like a very interesting collection, and it’s got some big names in it!

    September 2, 2013 — 11:50
  • Dianelos Georgoudis

    That a radioactive nucleus should decay at time t1 (and not, say, at time t2) is a contingent fact but, at least according to current scientific understanding, has no explanation. Doesn’t this defeat the PSR?

    September 3, 2013 — 4:53
  • Joshua Rasmussen

    You bring up precisely the “tricky issues” I had in mind. The way I see it, one’s estimation of the *prior* probabilities are relevant here. If one has reason to think non-earthly ancestors are unlikely, a priori, then *having no earthly ancestor* is epistemically relevant. Same goes for the prospect of a non-contingent cause. But I’m inclined to doubt that the prior probability of a non-contingent cause is low. (This involves responding to arguments to the contrary.)
    (Also, per your cavil, the first earthly organism may be simpler in a way that is relevantly different. Suppose we somehow *knew* the first organism was a rabbit…)
    BTW: Tyron’s book is *very* good.

    September 3, 2013 — 16:27
  • Mark Rogers

    Hey Dianelos!
    Noting your example of atoms I do not think Dr. Rasmussen will be arguing that all contingent facts have an explanation. Rather I think he will fall back and claim every concrete contingent thing has a cause for it’s existence. Then He will note the natural propensity of rabbits, which we are all aware of, for creating non-contingent finite causal chains. At which point he may pull out axiom 5 of S5 modal logic to say If A is possible, then it is necessary that it is possible. Who knows from there?

    September 4, 2013 — 8:46
  • residentoftartarus

    The short answer is no because there may be explanations for the relevant events not suggested by QM. To wit, QM doesn’t tell us that such explanations could not possibly exist.

    September 4, 2013 — 12:14
  • Kenny Pearce

    The answer to Danielos’s question is in fact that the explanans of a successful explanation need not necessitate (either logically or causally) the explanandum. We don’t need hidden variable theories here (though perhaps divine omnidetermination would give an explanation of a more robust sort). Again, see Pruss’s book.

    September 4, 2013 — 12:21
  • residentoftartarus

    In my opinion, the crucial point that is often missed is that modern physics does not so much tell us about the intrinsic character of physical events as it does their mathematical structure. This is why we know a lot more about the mathematical structure of gravity than we do how action at a distance works vis-a-vis gravity (a similar statement can also be made for something as vanilla as inertia). Hence also the debate as to how the mathematical regularities of QM are to be “interpreted.” Once this point is fully appreciated it becomes clear that no theory of modern physics could ever properly function as a defeater to something like PSR.
    In sum, there may be an explanation for such quantum events in the strongest sense and perhaps they can be discovered as part of an extension to QM or maybe not. In any case, QM proves nothing vis-a-vis the plausibility of something like PSR.

    September 4, 2013 — 13:51
  • Kenny Pearce

    Well, there are some serious philosophical difficulties about what the kind of knowledge you are talking about could amount to. What would it mean to understand how gravity or inertia work (in a stronger sense than present-day physicsts actually do)? I’m not really sure.
    I’m sympathetic to the view, defended by Berkeley, Leibniz, Schopenhauer, and, more recently, John Foster that the only kind of ‘intrinsic nature’ of anything that we can understand is what we grasp in introspection. This provides a strong argument for some form of idealism. Depending on one’s philosophy of science, and how scientific explanations match up with metaphysics, this may be consistent with a mental efficient cause (e.g., God) that necessitates the outcome of physically indeterministic events.
    But the key point is this: if ‘physical’ means ‘having to do with the science of physics,’ then we can be quite confident that physical indeterminism is here to stay. Among physicists, Boehmian mechanics is basically dead. Everettian mechanics is deterministic in a certain sense, and might be able to satisfy a stronger version of the PSR than collapse interpretations can, but it’s not deterministic the same way classical physics is.
    Once we go outside physics and start talking about some other sort of explanation, we have to get into a lot of messy details of metaphysics and philosophy of science.

    September 4, 2013 — 17:20
  • Dianelos Georgoudis

    It is true that strictly speaking physics reveals mathematical patterns within physical phenomena, and thus tells us nothing about causality. On the other hand such mathematical relationships can often be interpreted as implying causal relations. Take for example Newton’s F=m*a. When we see a ball accelerating it is natural to think that there is a force causing that acceleration. Or if we want to cause a ball to accelerate we figure we must apply a force to it.
    QM is different in two ways. First, there is nothing in it that can be interpreted as causing an actual change of the physical state. So, for example, there is nothing in QM that can be interpreted as causing an unstable nucleus to decay at time t1 and not at time t2. Second, the mathematical structure of QM is such that it positively suggests the *absence* of such causes. Von Neumann in his Grundlagen mathematically proved that (under some self-evident metaphysical assumptions) any scientific theory that uses physical parameters that can be interpreted as being causes (i.e. any “hidden-variable” theory) would contradict QM’s predictions. But QM is the most universally and most precisely confirmed theory of all time, which strongly decreases the probability that such a hidden-variable theory exists, and therefore the probability that there are grounds for believing in the existence of such hidden physical causes.
    Now a defeater of a belief is not a proof that this belief is false. Rather it is reason such that it warrants doubting the truth of this belief. And doubt it sufficiently that one does not accept it as a premise in an argument. Kenny is using the PSR as a premise to his argument from contingency, and I claim QM is a defeater of the PSR in this sense.
    A defeater-defeater could be a theory according to which it is not some hidden physical parameter, but God’s hidden will that causes the nucleus to decay at time t1 – and thus explains why it did decay at time t1. But until we have a good theistic argument for such an absolutist view of general providence, I think the PSR cannot be used as a premise in a theistic argument. And even then it can’t be used in an argument that is supposed to make sense to an agnostic – which is the type of the argument from contingency.

    September 5, 2013 — 7:16
  • residentoftartarus

    Alright, I see that it’s time to play my trump card.
    Conway’s recently published Free Will Theorem (FWT) says that if human experimenters are capable of making “free” choices (i.e., choices not determined by the past history of the physical universe) then the outcome of a certain type of quantum event is similarly “free.” Moreover, Conway also showed that this freedom cannot be attributed to any kind of randomness.
    Now, if we have good reasons for thinking that human experimenters are capable of making “free” choices in the relevant sense, as I think we do, then Conway’s FWT entails that we have good reasons for thinking that the very same freedom also exists at the quantum level and is non-stochastic! Hence, the idea that the quantum events discussed here might be explained in terms of non-physical causes is quite reasonable. Indeed, I would say that such an account in terms of non-physical causes is more likely than not at this point.

    September 5, 2013 — 9:21
  • Kenny Pearce

    ‘Theorems’ like this always rely on extremely suspicious definitions.

    September 5, 2013 — 11:05
  • Nathaniel Hagthorpe

    Danielos, but isn’t it the case, that you are still speaking about causal, necessitating explanations, thus ignoring the short answer Kenny gave (“The answer to Danielos’s question is in fact that the explanans of a successful explanation need not necessitate (either logically or causally) the explanandum.”)?

    September 5, 2013 — 11:35
  • residentoftartarus

    This will be my last comment on this subject.
    The problem I have with your suggestion that the relevant explanation is not necessarily a sufficient one is that it fundamentally breaks from the spirit of the contingency argument (see below). In my opinion, the defender of the contingency argument is better served by saying that there probably is some explanation for the relevant quantum events, which is then not reflected in the current state of quantum theory for whatever reason.
    The way cosmological arguments work is that they reason to God by applying some causal/explanatory principle to a fact about the world. As you know, the stronger the causal/explanatory principle the more mundane this fact about the world can be. Hence, cosmological arguments might be thought of as existing along a continuum, with the KCA on one end and the more scholastic versions on the other end. On this taxonomy, the different contingency arguments would be somewhere in the middle, though closer to the more scholastic arguments than the KCA in my opinion.
    My point in saying all this is that a cosmological argument that doesn’t have sufficiency baked into its causal/explanatory cake is working with a weak enough principle that I wonder if it shouldn’t just be a KCA style argument. After all, I have a much easier time thinking about things that cause other things to exist in a non-necessary way than I do explanations that account for the relevant explanandum in a non-necessary way.

    September 5, 2013 — 14:01
  • Dianelos Georgoudis

    I don’t understand what “the explanans of a successful explanation need not necessitate the explanandum” means. I started by pointing out that given modern science the PSR is not at all uncontroversial, and now find myself thinking that the very meaning of the PSR is not at all clear.
    Now perhaps all would become clear and certain to me if I first read Alex’s book (I only skimmed through the chapter on quantum mechanics). But, it seems to me, an argument which is supposed to be the most powerful philosophical argument for theism should not be based on a premise the acceptance of which depends on one’s agreement with a separate book-length argument.
    Let’s revisit my original doubt. Assuming physical realism, the following question suggests itself: What causes a particular radioactive nucleus to decay at time t1 (and not as some other time)? It seems to me there are only 3 possible answers:
    1. Some physical cause necessitated its decay at t1.
    2. Some non-physical cause necessitated its decay at t1.
    3. No cause necessitated its decay at t1.
    In cases #1 and #2 there is an evident explanation. In case #3 there isn’t an explanation. Case #3 may allow for an explanation at a higher level, for example an explanation of why it is the case that there is nothing necessitating the decay of radioactive nuclei at some particular time. But if nothing necessitated this particular nucleus’ decay at t1 then certainly nothing explains why it did decay at t1.
    Now let’s change gears and consider the question not from the agnostic but from the theist’s point of view. Traditionally theistic philosophers are inclined to assume that the world must be through and through rational or intelligible, and thus assume that the PSR or something like it must be true. But this attitude of expecting perfection not only in the whole but also in the parts strikes me as misleading. For it moves people (theists and atheists alike) to ask questions such as: What explains, or what’s the reason for this particular instance of natural evil? Or: What explains, or what’s the reason for this particular instance of bad biological design? These I hold are all misleading questions, for theism does not entail that there must be a good reason for each single bit, but only a reason for the whole of creation to be like it is, including perhaps many bits for which no good reason exists. Thus I find that on theism the PSR is probably false.

    September 6, 2013 — 6:56
  • Mark Rogers

    The PSR does face challenges. Dr. Steve Maitzen in his paper  QUESTIONING THE QUESTION, which was linked to at the beginning of this thread, at footnote 11 offers up another fact with possibly no explanation. So I am not surprised that you would need a book length argument to support a powerful philosophical argument for theism. I would think it would be a tedious endeavor as well to fully explain why according to orthodox quantum mechanics there is no reason for a radioactive atom to decay at T1 rather than T2.
    Certainly conventional quantum mechanics, if correct, will not allow a contrastive  explanation for why a radioactive atom decays at T1 rather than T2. That being said consider the following:
    1. When radioactive atoms release their extra energy, in an attempt to reach a more stable arrangement of their protons and neutrons, they are said to decay.
    2. Radioactive decay represents the transformation of an unstable nuclide into a stable nuclide.
    3. The unstable nuclei in a radioactive sample do not all decay simultaneously. Instead the decay of a given nucleus is an entirely random event.
    4. The actual moment of disintegration of a particular atom can be anywhere from the very moment the nuclide’s life comes into existence to infinity.
    Therefore the  reason for a radioactive atom decaying at T1 rather than T2 is that the individual atoms in a radioactive sample are not stable so they are prone to decay.

    September 7, 2013 — 5:57
  • Eric Steinhart

    Roger, you wrote: “Certainly conventional quantum mechanics, if correct, will not allow a contrastive explanation for why a radioactive atom decays at T1 rather than T2.”
    Why not say that the reason it decays at T1 rather than T2 is because the decay happens in world W1 rather than in world W2. On the many-worlds interpretation, that doesn’t seem too problematic.
    Random events just define the address of a world in the space of possible worlds.

    September 7, 2013 — 13:08
  • Nathaniel Hagthorpe

    Danielos, you wrote: “I don’t understand what “the explanans of a successful explanation need not necessitate the explanandum” means. I started by pointing out that given modern science the PSR is not at all uncontroversial, and now find myself thinking that the very meaning of the PSR is not at all clear.”
    If i understood Kenny correctly, a non-necessary explanans would be something like “Atom A decayed at timepoint T because it had a 58% chance of decaying at timepoint T, given Theory N”. Though it would be nice if Kenny could spell it out a bit more and correct me if i am wrong.

    September 7, 2013 — 17:21
  • Mark Rogers

    Dr. Steinhart,
    It is not clear to me that that is a contrastive argument.

    September 8, 2013 — 6:48
  • Eric Steinhart

    I’m not saying it is a contrastive explanation (or argument). Just that quantum randomness (like any other randomness in a universe) doesn’t entail the absence of a sufficient reason.

    September 8, 2013 — 8:24
  • Steve Maitzen

    Since I was mentioned recently, maybe I should emphasize that although I do think the orthodox interpretation of QM rules out any explanation of why a given particle decayed at T1 rather than at T2, I don’t regard that as a virtue of orthodox QM. I think of it as a defect, as I say in the cited paper. Yet the many-worlds interpretation invoked by Prof. Steinhart seems ontologically extravagant: countless nearly identical copies of the whole universe just to accommodate the decay of a particle that could have failed to decay (or the non-decay of a particle that could have decayed)? I think both interpretations are anomalies signalling the need for a better theory rather than consequences we should accept in order to save the theory.

    September 8, 2013 — 8:40
  • Eric Steinhart

    Steve: “countless nearly identical copies of the whole universe just to accommodate the decay of a particle that could have failed to decay (or the non-decay of a particle that could have decayed)”.
    No, the multiple worlds interpretation wasn’t developed “just to accomodate” that.

    September 8, 2013 — 11:09
  • Mark Rogers

    Dr. Steinhart,
    Nice one! Sorry I misunderstood what you did not say in the earlier comment. So you are saying that not only does the random decay of a radioactive atom at T1 rather than T2 not pose a challenge to the PSR, but I should move beyond my weaker reasoning because quantum mechanics provides a strong reason for that (all?) randomness. Quantum mechanics displays a radioactive atom decaying at T1 rather than T2  for a reason which entails neither a sufficient reason nor an absence of a sufficient reason.

    September 8, 2013 — 17:23
  • Dianelos Georgoudis

    Eric Steinhart,
    “Why not say that the reason it decays at T1 rather than T2 is because the decay happens in world W1 rather than in world W2. On the many-worlds interpretation, that doesn’t seem too problematic.”
    Consider any contingent fact X we know of, the question Q=“What’s the reason for X?”, and the answer A=“The reason for X is that we live in some world within Everett’s multiverse where X obtains”. We immediately realize that A is not a proper answer. After all the same A gives the reason for *all* contingent facts. It’s like answering Q with “The reason for X is that X has obtained.”

    September 8, 2013 — 18:17
  • Steven:
    “… although I do think the orthodox interpretation of QM rules out any explanation of why a given particle decayed at T1 rather than at T2”
    It depends on what one takes to be the conditions on contrastive explanation.
    Suppose the following is sufficient for p to explains q rather than r:
    (a) p explains q and ~r, and
    (b) were r rather than q true, p wouldn’t explain r.
    Suppose that T2 is earlier than T1 (the other case is actually easier). Then here is an Aristotelian causal powers explanation p: the particle had the proximate power not to decay at T2 and it had the proximate power to decay at T1. Note that this explains why it decayed at T1, why it didn’t decay at T2, and if it did decay at T2, p wouldn’t have explained that.
    A proximate power is like Aristotelian proximate matter: it has all that is needed for actuation.
    Or one can give a quantum explanation. This works most neatly if T1 and T2 are short intervals (I am not sure sharp times are actually going to be present in the physics). Let p = the particle had probability at least delta1 of decaying during T1 and probability at least 1-delta2 of not decaying during T2. This explains why it decayed during T1 and why it didn’t decay during T2, and if it had decayed during T2 instead, p wouldn’t have explained that.

    September 11, 2013 — 13:34
  • Steve Maitzen

    Alex: You remove any red card from an ordinary 52-card deck and shuffle the remaining deck thoroughly. q = The top card in the deck is black. r = The top card in the deck is red. p = 26/51 cards in the deck are black. Questions: If q, does p explain (q and ~r)? If q, does p explain (q rather than r)?

    September 11, 2013 — 19:05
  • Dianelos Georgoudis

    Following your conventions:
    T2 is less than T1 (for simplicity’s sake we can assume that T1 and T2 are non-overlapping short intervals).
    q=particle decays at T1
    r=particle decays at T2
    Aristotelian p = the particle has the proximate power to q and has the proximate power to ~r.
    Now at this point the issue of the exact meaning of “proximate power” arises, since from “having all that is needed for actuation” it is not clear whether actuation necessarily follows or not. If the latter than clearly p does not explain why q is true, but only why q is possibly true. But if the former, then p simply reasserts rather than explains q. Either way p does not explain q and the contrastive explanation as described does not obtain.
    Quantum p = (Pr(q)>=delta1) and (Pr(~r)>=1-delta2)
    It seems to me this p only explains why placing a particular bet on (q and ~r) is the best winning strategy, but again it does not explain why q or why ~r.

    September 12, 2013 — 2:49
  • Steve:
    In your example, p may explain (q and ~r), but will not explain q rather than r, because the counterfactual condition is not met: were r rather than q true, p would explain (r and ~q).
    To get a contrastive explanation in my sense, you can make p be: at least 26/52 cards are black.
    No, the existence of the power does not entail its exercise. But it explains it, by Hume’s principle that the cause explains the effect.

    September 12, 2013 — 9:10
  • Steve Maitzen

    Alex: You wrote, “p may explain (q and ~r).” In your view, does it? I ask because I’m not sure that either of your conditions on contrastive explanation is satisfied in my example, and I’m therefore not sure that either condition is satisfied by the Aristotelian or QM explanations you gave. Nor do I see how “At least 26/52 cards are black” (which is compatible with “Exactly 26/52 cards are black”) explains why the top card is black rather than red, or black rather than not.

    September 12, 2013 — 9:51
  • Dianelos Georgoudis

    If the existence of proximate power does not entail its exercise, then it does not explain its exercise. It seems to me it only explains the possibility of its exercise.
    I understand you are saying:
    1. The particle has the proximate power to decay at T1.
    2. The particle did decay at T1.
    3. Therefore we have an explanation of why the particle decayed at T1.
    Consider the analogous case:
    1. The glass has the proximate power to remain whole (not break) when it falls to the ground.
    2. The glass did remain whole when it fell to the ground.
    3. Therefore we have an explanation of why the glass remained whole when it fell to the ground.
    But we don’t, do we? It seems to me the right #3 would be that we only have an explanation of why it was possible for the glass to remain whole when it fell to the ground. More is needed to explain that it did in fact remain whole. And by analogy in the previous case we only have an explanation of why it is possible for the particle to decay at T1.
    You mention Hume’s principle that the cause explains the effect. The idea I take it is that if we believe that C causes E we also think we have an explanation of E, namely that it was caused by C. But I wonder if that always holds. Suppose we believe that our brain causes our consciousness. Would we then think we have an explanation for our consciousness? It does seem we can explain something, for example answer the question “Why is there this particular consciousness?” with “Because there is that particular brain causing it”. But we don’t explain the fact of their being consciousness, for we can imagine the brain being there without causing the effect of consciousness.
    And coming back to the case of the particle, the anti-PSR hypothesis is that nothing causes it to decay at T1. Rather the nature of the particle is such that it may spontaneously and with no reason or cause at all (but with a particular probability) decay at T1. I have no trouble at all conceptualizing such a state of affairs. I realize it leaves something intrinsically unintelligible hanging in there in reality – but as I explained above as a theist I don’t feel bothered by this. And I figure the naturalist will be even less bothered.
    On the contrary, it seems to me that the PSR makes theodicy much harder, if not impossible. What’s more it renders the world a much less interesting place. In a word in which the PSR holds I have trouble making sense of free will, or of creativity. And doesn’t the PSR lead to infinite regressions? I see problems galore. I suppose I should read your book, but I don’t find a kindle edition of it.

    September 14, 2013 — 0:34
  • Dianelos:
    The Kindle edition is here.
    It seems to me that you’re putting much too stringent a requirement on explanation when you say that the possibility of something else having happened is incompatible with explanation. On this strong account of explanation, nobody has ever given a good scientific explanation.

    September 16, 2013 — 20:02
  • Anynom1

    Question not pertaining to the syllogism of the argument. While one could appeal to God’s nature as necessary, isn’t God’s decision to create the universe a contingent fact?

    September 21, 2013 — 22:57
  • Kenny Pearce

    Anynom1: Yes.

    September 21, 2013 — 23:30
  • Edward Moad

    Thomas Rauchenstein: above, you said “The APSR has advantages and drawbacks. One advantage is that it is easier to defend because it is compatible with the possibility that in some possible world, a contingent being simply exists as a brute fact, without explanation.”
    This brought to my mind the question, how to distinguish a contingent being that exists as a brute fact, from a necessary being?
    One way I thought to phrase this question was: Consider two possible worlds, one in which a contingent being exists as a brute fact, and the other, identical to it in every way, except that the same being exists necessarily rather than contingently. What is the difference between these worlds?
    Of course an obvious reply, I suppose, would be that to be necessary just is to exist in all possible worlds.
    But that leads me to ask: is there a difference between a necessary being, and a being that exists, as a brute fact in all possible worlds?

    September 24, 2013 — 11:07
  • Kenny Pearce

    Edward: A being who existed as a brute fact in all possible worlds would exist as a matter of brute necessity, whereas one that existed as a brute fact in only some possible worlds would exist (at each of those worlds) as a matter of brute contingency. Of course, that’s just definitional.

    September 24, 2013 — 13:54
  • Ryan smith

    Maybe I’m missing something, but 4 is false. As written, it’s not circular to explain E with a contingent fact. It’s only circular given the premises if you think facts are beings. It may lead to an infinite regress to explain a contingent fact with another contingent fact if things don’t eventually get necessary, but that’s exactly the kind of thing this argument was postulated as avoiding.

    November 19, 2013 — 17:38
  • Ryan – If there are contingent facts that don’t depend on the existence of contingent beings, then yes, I suppose (4) might turn out to be false.

    November 19, 2013 — 19:43
  • Ryan Smith

    I wonder if a law of physics would qualify as a contingent fact that doesn’t depend on a contingent being.

    November 20, 2013 — 16:51
  • That is a view one could have about laws of physics, if one were a non-reductivist. However, attempts to reduce the laws of physics typically appeal to contingent beings, generating the circularity in question. (Reducing the laws of physics to divine volitions has also been attempted, e.g. by Malebranche.) Second, even if one is a non-reductivist, one might think that the laws of physics are contingent beings rather than irreducible contingent facts. After all, non-reductivists often want to talk about laws of physics making things happen in some robust, anti-Humean sense, and most philosophers attribute that kind of causal efficacy only to substances.

    November 20, 2013 — 16:58
  • Leave a Reply to Kenny Pearce Cancel reply

    Your email address will not be published. Required fields are marked *