Cosmological Argument

| | Comments (21) | TrackBacks (2)

This is the the eighteenth post in my Theories of Knowledge and Reality series. Follow the link for more on the series and for links to other entries as they appear. I've posted an earlier version of this a while ago, but the comments degenerated into a discussion of things completely unrelated to the post. That time, it was a version of my notes on this that hadn't been altered since 2001. I've decided to expand it a bit based on further study of the subject, even though I haven't taught all these issues in the course that this series is based on. I should also say that my presentation depends heavily on William Rowe's work, most importantly the short article he wrote for introductory courses that appears in Reason and Responsibility, ed. Feinberg and Shafer-Landau, with one reference to one other text I have used in that course, Jan Cover and Rudy Garns's Theories of Knowledge and Reality (abbreviated TKR).

The cosmological argument for the existence of God is one of a number of classic arguments sometimes used in conjunction with each other to establish the existence of a being with some of the characteristics generally taken to be true of God. I'm going to look at three such arguments, each contributing something different to the overall picture The cosmological argument in particular occupies a very small role in any overall picture of how some have offered argumentation in support of theism.

The argument relies heavily on the Principle of Sufficient Reason (PSR), which requires an explanation for any actual thing or any fact. William Rowe's criticism of the argument amounts to saying that we have no reason to believe PSR. I will save this discussion for the end. For now, let's take it to be a reasonable principle (if for no other reason than to see what follows).

Given PSR, there are two kinds of things -- dependent things and self-existent things. Dependent beings would be just about anything you come across in ordinary experience and lots more -- rocks, trees, houses, people, stars, electrons, etc. For something to be a self-existent being, its own nature has to provide the explanation for why it exists. God has traditionally been offered as a good candidate for a self-existent being, since traditional theology has always taken God to be necessary and self-existent.

The argument first shows that it cannot be that there are just dependent things, even if they go infinitely back into the past. Thomas Aquinas, Samuel Clarke, G.W. Leibniz, and others who endorse this argument will be perfectly happy to concede (for the sake of argument) that dependent things have always existed, with no beginning, with each explained by earlier dependent things. Aquinas actually doesn't think this could be true, and he gives three very interesting arguments why, but I'm not going to worry about it, because I think the argument can go through even if he's wrong on all those points. So suppose there's an infinite series of dependent things, each explained in terms of a previous dependent thing, with no first dependent thing. Does that explain everything? Some philosophers have thought that it does, since only dependent things exist, and each dependent thing is explained in terms of a previous one.

But this is too fast. Aquinas, Clarke, and Leibniz point out a problem with that conclusion. They can still ask why there are any dependent things at all. Explaining each thing explains each thing but not why there are any at all. The existence of dependent things is still unexplained. This infinite series into the past does not answer that question, which leaves us having to offer another solution -- there must be something that is not dependent, and that thing can provide the answer to the question why there are any dependent things. If PSR is true, then this is in fact the only kind of answer that would do the trick.

Here is a valid argument that reflects this kind of reasoning:

(1) everything is either dependent or self-existent

(2) not everything can be a dependent thing

therefore, (3) there is some self-existent thing

If PSR is true, then (1) is true, because the only other possibility would be for something to be without explanation, and PSR rules that out. Similarly, if PSR is true, (2) is true, since there is no way to explain why there are dependent things by simply adding more dependent things. Some other kind of thing needs to exist, and the only other kind of thing is a self-existent thing. The conclusion logically follows.

A number of objections have been raised against this argument. I will discuss the two that I think are the best in the next post. Some claim that the universe itself is the best candidate for the self-existent thing, and some resist the argument by denying PSR. Comments related to those issues will be ignored. Please save them for the next post in the series. Comments on other elements of this argument are welcome.

2 TrackBacks

Listed below are links to blogs that reference this entry: Cosmological Argument.

TrackBack URL for this entry:

I'm doing a series at my own blog based on the content of an intro to philosophy course that I teach from time to time called Theories of Knowledge and Reality. This is designed to reflect that course, so it's... Read More

Welcome to the 103rd Christian Carnival. With thirty-eight featured blog posts, a variety of topics are available for your perusal. Please take time to scroll down the list and find a few that interest you. Holiness and holy days While Christmastide ... Read More



I'm wondering if you could say something about what you see as the purpose of these sort of arguments/proofs. Do we (christians) have some sort of obligation to convert atheists to theism, or do they serve some other sort of purpose?

I realize that all sort of people have their minds changed by all sorts of arguments, but I'm curious about your position on classical-type apologetics and their usefulness or necessity.

They can still ask why there are any dependent things at all. Explaining each thing explains each thing but not why there are any at all. The existence of dependent things is still unexplained. This infinite series into the past does not answer that question,

I don't find myself really buying this line of reasoning. I have two objections:

1) It seems to me that if you can explain each individual thing, you don't really need to explain the whole class of things. Classes are abstractions, not things in and of themselves, and thus do not require explanations.

2) The argument here seems to be "why are there any dependent things at all?" If you find a way to circumvent my first objection, then this question may end up proving too much. After all, even if you have self-existent things, that still doesn't require that you have any dependent things. Now you still have to come up with some sort of explanation why dependent things exist.

If they are necessitated by the self-existent things, then they are really, in combination with the original self-exstent thing, part of a larger self-existent thing. If they are not necessitated, then why are they there? So you may end up proving too much and end up with everything being self-existent, which is surely not where you want to go.

Wink: So have you changed your mind on the time travel stuff that you posted about here? Because I think this is the same issue. If I have a watch that I go back and time and give myself, thus explaining where I got it, you don't think that's a sufficient explanation even though each stage of its existence is explained. You don't think there's yet an explanation of why there's a watch at all. What's going on here is similar. Each thing is explained by previous things, but there's no explanation why there are any of them? There's no answer to the question why there's anything at all (or at least the question why there's anything besides what exists necessarily). That seems to me to be a pretty reasonable question to ask, and I think there ought to be an explanation for the existence of anything besides what has to exist of necessity.

As for your second worry, I'm not sure I see a problem. The explanation for dependent things would be that the self-existent thing explains their existence in some way. The way I've framed the issue, it's not quite about necessity and contingency. It's about dependence. If B depends for its existence on A, and A is self-existent, then B is still dependent. There's a problem with contingency and necessity that I've dodged by setting up the argument this way, but I think that's a virtue of this argument.

I realize that it's hard to explain creation by a necessary being as a contingent matter if you need an explanation for that contingent act. If there's a sufficient explanation for the contingent act, then it's not contingent. I'm pretty sure that problem doesn't arise for dependence relations the way I've set this up, and the particular kind of explanation I'm requiring with my version of PSR doesn't entail that the dependent thing is necessary or that it's contingent. It's consistent with my whole argument that everything is necessary, but some of the necessary things are dependent on other things for their existence.

Now your last paragraph raises one concern about the concession that they're necessary. It first of all assumes that any two things form a larger thing, which is controversial among philosophers but probably the dominant view today, at least in the circles I've run in. But even if you go this route, I think you would have to say more than that everything is self-existent. Consider the following:

So we have a self-existent thing we'll call Al. Al creates some other stuff that isn't self-existent. Call it Bill. Bill is wholly dependent on Al. On your proposal, Bill and Al together form a third thing that we'll call Chris. Chris is indeed self-existent in a technical sense, because it contains within itself the entire explanation for its existence. But this way of being self-existent is because a part of Chris is self-existent and the rest of Chris is dependent on the part that's self-existent. Bill exists uncontroversially. Bill is just what we call the universe. Establishing a self-existent thing Al that Bill depends on does indeed mean that this third thing Chris is self-existent, but that doesn't mean Bill is self-existent, and that's what it sounds like you're saying when you say that the universe is self-existent. So I think the argument really does show some important metaphysical truth if indeed it's a good argument.

Paul: This is actually what I wanted to avoid, because it monopolized the comments on my previous post. It also came up in my critique of presuppositionalism. I will say something here, but I don't want this to devolve into a discussion about that stuff.

I'll say a few things in summary of my views, but you can read those posts and their comments for more detail.

1. All philosophical thought is intrinsically valuable, because it leads to truth when done well. This is a part of that.

2. Evangelism doesn't always need apologetics, but sometimes apologetics serve one element of evangelism in terms of the necessary intellectual component of Christian belief.

3. Someone can resist any argument by denying a premise (or certain other methods, depending on the argument). That doesn't mean pursuing the arguments is bad. Some people act is if an argument that can be dodged is worthless. Not everyone will find the dodges as plausible as others, and if the dodges can be unmotivated then the argument is in much better shape.

4. I don't think being convinced of theism or even of the propositional content of Christianity will make someone a Christian. All a philosophical argument will do (at most) is bring someone to propositional beliefs (or knowledge) of some component of theism or of Christianity. I don't think a Christian should ever think these arguments could do more than that, as evidentialist apologetics seems to try to do.

5. The classical approach doesn't treat it that way, though. It just sees it as a means (but not the only means) to achieving one of the necessary steps to Christian faith. Christian faith necessarily involves a propositional belief that God exists and that God has certain properties. Classical arguments seek to establish that some being has those properties. That can and does play a role in conversion for some people.

6. Christian apologetics isn't just for the sake of converting people. Some people think of Christian apologetics as more directed toward solidifying the propositional understanding of crucial elements of Christian belief among those who are already Christians.

7. One objection people sometimes give to this understanding of apologetics is that it doesn't rely on the Holy Spirit. I'm talking on the level of efficient causes, what processes lead to other processes. The work of the Holy Spirit is more about final causes, the reasons why any causes might cause certain effects. The argument against classical apologetics usually involves an assumption that God simply cannot work through the means of bringing someone to propositional knowledge or belief by means of convincing them through arguments. I will not place that kind of limit on God.

Jeremy: So have you changed your mind on the time travel stuff that you posted about here? Because I think this is the same issue.

Three responses to this:
1) My original concern with the time-travelling watch example was more about thermodynamics than about explanations (though the comments certainly ended up focussing more on the latter than the former). Though I did highlight the dilemma by asking "Who made the watch?", that was because that question had more visceral impact than a question about spontaneous generation and destruction of matter.

2) I haven't finished changing my mind on this, but I have started changing my mind on this (and I reserve the right to change my mind back). As I come to understand Quantum Mechanics better, this kind of issue becomes less and less of a problem for me. (Though it has not ceased being problematic for me, my remaining issues with this kind of thing may disappear as I understand QM even better. But then again, they may not.) Very non-intuitive things happen with QM and this kind of thing seems to be of that sort. For example, elementary particles and their associated anti-particles can spontaneously appear, then collide, thus disappearing. Essentially, the universe is creating matter/energy ex nihilo, going into "energy debt" to do so, but that is OK as long as it "pays the energy back" by later annihilating the matter. If elementary particles can pop into existence w/o explanation, then there is no particular reason that a watch cannot, though that is almost infinitely more improbable. Plus, Feynman has all sorts of stuff about elementary particles travelling backwards in time (which manifests itself as anti-particles to those travelling forwards in time) which I need to read up on. Those writings are likely to change my mind on this issue even further.

[Note: the benefit to changing my mind on this is that it leaves me with no objections to a consistent theory of time. Plus, it puts me in the same camp as Feynman, which isn't exactly bad company to be in. The downside it that I have to accept all sorts of non-intuitive notions.]

3) These certainly are similar issues, but intuitively they feel significantly different to me. The time-travel example involves a finite loop of time, while the example above is an infinite series. That seems significant. In the above example, each thing has a temporally prior explanation (and, as a bonus, does not run into any thermodynamic difficulties). And you don't need an explanation for the "first thing" because there is no "first thing". Thus, for each thing X, there is always an answer to the question "Who or what made X?" Whereas in the watch example, there is no answer for that question.

So combine the three, and that's why I posit this objection.

The way I've framed the issue, it's not quite about necessity and contingency. It's about dependence.

Ah, I had missed that. I suppose that voids my objection, though it does feel a bit dodgy. I feel like I should be able to re-frame my objection to meet this, but I don't seem to be able to manage it at the moment. Any takers?


First, Merry (belated) Christmas. This question is, I hope, one that can be raised without asking whether the universe is a necessary being, but if not, let me know. It seems that if you think the PSR can shoulder the weight it must for the argument to cover both vertical and horizontal regresses, then there is a good chance that 'reason' is either ambiguous or there are many types of reasons. That's because what counts as a reason in response to a question about horizontal series is quite different from what counts as a reason in response to a question about vertical series (Think of the horizontal series as the temporally ordered series that some think could extend back infinitely).

Suppose, then, we distinguish between PSR-thick and PSR-thin. PSR-thin is just the principle that for every_thing there is, there is a horizontal reason. PSR-thick must go beyond this in some way so that even if we have a horizontal reason for every_thing there is, there is still some further application that commits us to a necessary being. Perhaps it is the reason for everything that is not itself contained within the reasons for every_thing.

I think a challenge you face is formulating a version of PSR-thick that is not immune to counterexample that also finds support in the sort of intuitions that support PSR-thin. I suspect that the sorts of intuitions that motivate people to endorse PSR are those that support only PSR-thin, not PSR-thick.

Oops, that should read, "a challenge you face is formulating a version of PSR-thick that is not ONLY immune to counterexample..."

[The type of counterexample I'm thinking of is one in which there is some event e where we think that the reason for e is e', e' is a horizontal-reason and there is no further vertical reason that is needed].

Clayton, I'm not exactly sure how you're using your terms here. Is a horizontal reason an efficient cause, while a vertical reason is a final cause? Or is a horizontal reason one that naturalism could accept while a vertical reason is one that naturalism wouldn't? Or is it some third distinction? I think those lead to very different results, and I can't wrap my mind around what the objection is supposed to be because I'm having trouble figuring out which one you mean.

Wink, I would say that the Principle of Sufficient Reason as I've formulated it does not require an explanation of the sort that quantum indeterminacy denies will always occur. A quantum indeterminacy just means there's no efficient cause. Maybe this is related to what Clayton is getting at, but the version of PSR that I have in mind just says that there's some explanation, and a divine final cause that explains why something would occur for a purpose, even if it's not efficiently caused, would I think satisfy the principle. This is, I believe, what Augustine and Aquinas would want to say about human free choice anyway. They think it can explain divine sovereignty in the absence of divine causation of human actions.


If we think of cosmological arguments as involving a kind of threat of regress, it seems the regress might move in two directions. Some might argue, using the kind of reasoning from the Kalam argument, that for each thing that happens at one moment, there was some reason at a prior moment why it rather than something else. This series would be horizontal. The cosmological arguments that Clarke and Aquinas offered were supposed to work even if there was an infinite series of prior causes. I don't think this has to do with efficient/final causation or causes naturalists can/can't accept really. Rather, the idea is this. Suppose I say that there has been an infinite series extending backwards through time and the result is this:

There is no event in the history of the world that lacks a reason in the form of a prior event that explains why it rather than something else occurred.

At this point, the 'vertical' question, as I'd call it, arises which is why the series that contains the actual series as members rather than some other possible series? If this question isn't just confused, it must be taken as a question for an explanation that looks for a reason not contained within the series such as a member of the series temporally antecedent to the events in the actual series.

The worry I have is that an application of PSR-thin won't get you from 'e1, e2, e3 ... occurred' to 'There is a reason r distinct from (and not composed of) the reasons for e1, e2, e3 ...'. If there were such a reason, it couldn't be located within the series. If there were no logical guarantee of such a reason, however, there would be no threat left over once the infinite series were posited. But then there's the question as to whether we really accept PSR-thick or whether the version we are convinced of is PSR-thin, a version that requires only that for every_thing that takes place there is a reason but does not require anything further.


I apologize for making you digress. I hadn't read your other stuff on this subject, so I was just curious as to what your thoughts were. I remember reading somewhere (I've forgotten where) some intriguing comments about the classical arguments in Aquinas and the fact that they were not so much directed at atheists (since there weren't a whole lot of them in Europe at that time), as much as being a bolster to the faith of christians going through doubts, similar to your point 6.

Another thought which has intrigued me in apologetics comes from Tim Keller in a lecture on presenting the gospel as part of preaching (delivered at Covenant Seminary and available through their website somewhere). Keller suggests that a regular part of preaching should be adressing proactively the sorts of objections which people tend to raise to christian belief.

Keller calls these "defeater beliefs", meaning that people say "christianity can't be true because of [x]." He goes on in the lecture to identify what he thinks are the most common of these and some ways to address them.

Lastly, to close my little contribution out, if you ever are interested in trying out arguments with actual atheists in a lively forum, check out Their site has quite a bit on debates on God's existence. I had some (very little) fun talking about Jesus' resurrection on their forum some time back.

I always figure these sorts of things only have value as they relate to actual people.

Jeremy - I still think that the infinte regress of dependent things is a powerful argument. While I'm willing to accept (somewhat reluctantly) your assertion that final causes without efficient causes are sufficient explanation, you still haven't really addressed my feeling that infinte regression is different than temporal loop--namely, that each thing has a temporally prior explanation. (See point 3 in my previous comment.)

So, going back to my original argument: with an infinte regress of dependent things, each thing has an explanation. Your response: "Why is there anything at all?" I now have two responses to this:

1) (I've already stated this one.) It seems to me that if you can explain each individual thing, you don't really need to explain the whole class of things. Classes are abstractions, not things in and of themselves, and thus do not require explanations.

2) (This is a new one.) If you can use "Why is there anything at all?" to challenge the class of dependent things, then you can also use it to challenge the class of self-existent things. That question proves too much. Just as every dependent thing in the infinte regress had an explanation, but there was no explanation why there were any of them, just so each self-existent thing has an explanation, but there is no explanation why there are any of them.

You cannot invoke the necessity of self-existent things as an explanation because you have framed the argument so as to avoid issues of necessity and contingency. (And at any rate, if you can show that God necessarily exists, then you don't really need to bother with any of the other proofs for the existence of God.) So you seem to be stuck--why are there any self-existent things? Any answer to that question will force you to accept an infinte regress of dependent things as well.

Clayton: I'm still not 100% sure I'm getting you, but if you're saying what I think you're saying then Wink endorses PSR-thin but not PSR-thick, and I endorse PSR-thick. So I guess you're saying that any version of PSR-thick will be subject to counterexamples? I didn't get any sense of what you meant by that or what sort of counterexamples you had in mind.

Wink: I think one thing that bugs me about an infinite regress of causes is as follows. Suppose A causes B, which causes C. If A wholly causes B, and B wholly causes C, then A is really the cause of C. B only serves as an intermediate cause. So B doesn't really serve as the explanation for C, because it in turn is dependent on A. But with an infinite regress of causes, nothing is a real explanation. Nothing is a real cause of anything, in this sense. This is one of Aquinas' arguments, and I find it a lot more convincing than most philosophers today seem to.

I didn't invoke necessity to explain the self-existent thing's existence. I invoked self-existence. Necessity is about whether it can fail to exist. Self-existence is that its nature explains its existence. I think self-existence implies necessity, but it's not about necessity as opposed to contingency, because dependent things can be necessary as well, as long as the dependence of dependent things is metaphysically necessary given the existence of the self-existent thing. So I want to keep those two things apart.

You say that the "why is there anything at all" question is parallel for self-existent and dependent things. I don't think it is. If no self-existent thing exists, and dependent things go back infinitely, then there is no answer to why there are any of them at all as opposed to there being nothing. If only a self-existent thing exists and no dependent things, there is an answer to where there is anything at all. That answer is in the nature of the self-existent thing. So I don't see how they are parallel in the sense you require.

Paul: Aquinas had multiple audiences. One key audience for him was Muslims. Some of his arguments for theism are arguments for theistic claims that Muslim philosophers at the time tended to deny. His Summa Contra Gentiles was primarily intended to engage in dialogue with Muslims, the one non-Christian group who was most heavily involved in philosophical debate with Christians at the time (though he had Jews in mind as well). It was his altenative to the Crusades, winning people over via argument rather than (as he seems to have seen the Crusades) reclaiming conquered lands for the sake of spreading Christianity.

I tend to stay away from places like If it's anything like it used to be, it tends not to attract the best defenders of atheism. Many there seem to offer any argument they come across, from the more difficult ones for theists (though usually in oversimplified form) to the most ridiculously bad, presenting each one as if it shows demonstrably that theism is logically impossible. I much prefer people like Clayton. Christian apologists have the equivalent problem at many apologetics sites, so this is nothing about either view. It's just that many of the popular sites on either side are, to my mind, largely uninformed about the best philosophical, historical, or theological discussions over the past millenia.

What Tim Keller has in mind is usually called negative apologetics. I think there's a place for positive apologetics as well, but I do think what he's talking about is the most effective place for philosophical argument in Christian apologetics.

"I tend to stay away from places like If it's anything like it used to be, it tends not to attract the best defenders of atheism. Many there seem to offer any argument they come across, from the more difficult ones for theists (though usually in oversimplified form) to the most ridiculously bad, presenting each one as if it shows demonstrably that theism is logically impossible."

Well, yes, that is what I found as well. But, they are real people, for whatever that's worth :) Just didn't know if you were aware of the site. The site as a whole has a pretty large set of links to apologetic arguments and debates, so it is one resource for people who want to see that sort of thing.

You may be entirely right about Aquinas. I wouldn't be in a position to know. I just play around at this stuff.

Jeremy - If only a self-existent thing exists and no dependent things, there is an answer to where there is anything at all. That answer is in the nature of the self-existent thing. So I don't see how they are parallel in the sense you require.

Hmmm...let me see if I can clarify my argument. Let us call Set S the set of all things that, if they existed, would be self-existent. Any particular member of Set S may or may not actually exist. Now if any particular member M of Set S actually exists, then M clearly has an explanation: itself. But the question remains, why does that particular member of S exist, and not other members of S? Why do any of the members of S exist since it is conceivable that none (and therefore, potentially nothing) exist?

There are several possible answers, all of which have their own problems:

Resonse 1) All Members of S exist. They must necessarily exist. That is implied by self-existence.
Objection 1a) This seems to confuse two levels of explanation. Why does M exist is clearly answered by M. But that does not mean that M necessarily exists, for that is to confuse the former question with "Why doesn't M not exist?"
Objection 1b) All Members of S cannot exist for some members of S are mutually exclusive. For example, the Gods of Theism, Pantheism, and Panentheism are all members of S, yet each one excludes the other two. Therefore it cannot be the case that being a member of S implies the necessity of existence, or that all members of S exist.

Response 2) All Memebers of S exist, but they do not exist by necessity.
Objection 2) Same as Objection 1b.

Response 3) Only some Members of S exist. The reason some, but not others, exist is found within each existing member of S.
Objection 3) This reduces to response 1 and can be refuted by objection 1b.

Response 4) Only some Members of S exist. The reason some, but not others, exist is found outside of each existing member of S.
Objection 4) This makes the existing members of S dependent on whatever that external explanation is. Thus, those members are not really self-existent.

Response 5) Only some Members of S exist. There is no particular reason why some exist and not others.
Objection 5) This runs afoul of PSR.


The thick/thin distinction is inspired by Hume's 20 grain of sand remark. He says (basically) that once you've given 20 explanations for grains 1, 2, ..., 20, there is no further explanation to be given of the 20. He uses this to criticize the version of the cosmological argument that you get from Clarke that (I think) is supposed to show that even if the universe can be divided into an infinite series without a beginning so that there is no event e without a causal explanation in virtue of a prior event e', there is the further explanation of the series. Hume denies this and I'd put forward his 20 grain of sand example as a threat to PSR-thick. Once you have an explanation as to how each grain got there, there is no further explanation we can 'quantify over' as it were that would pertain to the totality, the series, etc... I think that if you accept PSR-thick, even explanations of events that cite nomologically sufficient priors back through eternity would be essentially incomplete. Denials of PSR-thick just don't seem as jarring as denials of PSR-thin. We would be shocked if an omniscient being could give no explanation as to how this one grain of sand got here but not if it could provide no further explanation as to how the collection got here beyond the one given of the parts. But if the denials of PSR-thick aren't jarring, I'm not sure what props it up.

I'm not sure what evidence there is for PSR-thin, by the way. I take it that people are attracted to versions of PSR on apriorish grounds. Basically, people I know who like it find its denial quite inconceivable. I'm not sure that the same goes for PSR-thick.

Wink, I was distinguishing them because they're distinct in principle and because I think the fundamental issue here is self-existence and not necessity. Your problems are problems for those who think there are members of S that do not exist. I'm not even sure there are members of S that might potentially not exist. I tend to think S is exclusively composed of things that are necessary.

I don't think 1a is a problem, because I don't see any difference between asking why something exists and asking why something doesn't not exist. Those are logically equivalent.

I don't think 1b is a problem, because I don't think those proposed members of S that don't exist are even possible. If the actually self-existent thing is the Christian God, then the Stoic god is impossible (and so on).

Is Hume using this example to criticize Clarke? The way Rowe presents it, Clarke's argument is in response to the kind of point Hume is making, and Hume just failed to see that point. I can't see how Hume's point could be a response to Clarke anyway, because Rowe seems right to me. Hume just ignores what seems to me to be a perfectly valid distinction between explaining each actual thing and explaining why it isn't true that no such thing exists.

Denials of PSR-thick do seem a little jarring to me. Wink's time travel example seems to me as if it's just got to be metaphysically impossible, and PSR-thick is the only reason I can think of why it would be.

I don't think a priori understanding is supposed to be the only reason we should believe PSR-thin. It's not evidence in the logical positivist sense, but it's a good argument to say that we should believe PSR because it makes sense of why we should believe in scientific laws. The whole notion of science makes no sense if we don't think there are explanations for things, and Occam's Razor warns against thinking we should seek explanations most of the time but not in cases like this, unless we can produce a good reason not to believe PSR applies in this case. That's why I think there's an argument for believing PSR, though I wouldn't call it evidence. Whatever reason we have for believing science is worth doing gives us reason for thinking that our belief in PSR should be at strong as our belief in scientific laws.

Jeremy - I don't think 1a is a problem, because I don't see any difference between asking why something exists and asking why something doesn't not exist. Those are logically equivalent.

I guess to clarify what I mean by that (so that they aren't logically equivalent: once a self-existent thing T exists, then the explanation for its existence is there--itself. But consider the state prior to the existence of T (whether temporally prior, or logically prior). In that state when T does not yet exist, there is yet no reason for T to exist (since the reason, T, doesn't exist). So in that state, T would have no reason to exist. So why not no T?

This is parallel to the infinite regress of dependent things. Once they exists, the reason for all of them is clear. But consider the state when none exist. Why then should they come into existence?

I don't think 1b is a problem, because I don't think those proposed members of S that don't exist are even possible. If the actually self-existent thing is the Christian God, then the Stoic god is impossible (and so on).

But the Stoic God is only impossible because the Christian God actually exists, not because the Stoic God is self-contradictory. You can't appeal to the existence of the Christian God to disprove the Stoic God when you are trying to prove the existence of the Christian God. That is begging the question.

You'll note that my definition of S, combined with your not being allowed to beg the question, should dodge your entire complaint. S is composed of all things that, if they existed, would be self-existent. The stoic God, or the Pantheist God, or the Panentheist God, if any of them existed, would be self-existent. There is nothing in any of them that is inherently impossible--only the actual existence of the Christian God makes them impossible.

Your rebuttal in this light only serves to reinforce my claim that members of S must necessarily exist because several of them are mutually exclusive.

It's only question-begging if I need the assumption to establish the conclusion of the argument, but I don't. The conclusion of the argument is consistent with the Stoic god, the Christian God, or any other self-existent thing. I did later on indicate why I don't like to think of the universe itself as the self-existent thing, but the argument itself assumes no such thing. It just concludes that something is self-existent.

What we were talking about here, though, was my particular account of what that explanation might be. It was the Christian God as the self-existent thing that ultimately explains all dependent things. You objected to seeing God that way on the grounds that other potentially self-existent things might not exist, so you wondered why God would have to exist. This was my response to why, in my account, this wouldn't have to be the case. None of this was needed for the original argument, though. This was just exploring why in my positive, unsupported, account it's still a coherent picture.

I'm having trouble seeing how your account of a self-existent thing in the first part makes sense. There's nothing logically prior to the existence of a self-existent thing. If there were, it would be a dependent thing. So how can it make sense to ask about what would be true in the logical progression of things "before" it exists? It does make sense to ask this of a necessary thing that isn't self-existent but is dependent, but I can't see how this is even a coherent description of a self-existent thing.

In other words, I think we've got the relation between self-existence and necessity backwards. You're talking as if something can be self-existent while not being necessary. I want to say that something can be necessary but not self-existent, but I'm not sure how something can be self-existent but not necessary. You've given purported cases, but I think that just shows that only one of them (at most) can really be possible. Maybe they're all logically possible, but I can't see how they could all be metaphysically possible, since they're mutually incompatible but all supposedly explained from within themselves.

What this means is that there can't be a set S of things that are self-existent, not all of which exist. There is a set N of things that are necessary, not all of which are self-existent. But I don't think your objection works with that set.

Hmmm...we seem to be losing the point of my objection, so let me restate in simpler and more concrete terms:

Consider the God of Theism (GoT) and the God of Panentheism (GoP). We can make 3 observations about GoT and GoP:

o1) If GoT exists, GoT is self-existent.
o2) If GoP exists, GoP is self-existent.
o3) GoT and GoP are mutually exclusive.

Now consider the following states:

s1) Neither GoT or GoP exist.
s2) GoT exists and GoP does not.
s3) GoP exists and GoT does not.

So my question is, what is the explanation for why one particular state is true, and not a different particular state? Why s2 and not s1 or (in particular) s3?

Response1: GoT is necessary.
Objection1: Whatever argument that GoT is necessary can also be made for GoP, so this does not help to explain why s2 should be true while s3 is false. Furthermore, GoT and GoP cannot both be necessary for they are mutually exclusive.

Response2: s2 is true because of something within GoT.
Objection2: whatever that something is within GoT, it can also be asserted for GoP, so this does not help explain why s2 should be true while s3 is not.

Response3: s2 is true because of something outside of GoT.
Objection3: If this is the case, then GoT is dependent, not self-existent.

Response4: There is no explanation why s2 is true and s3 is not. It just is.
Objection4: This violates PSR.

Response5: States like s2 and s3 do need require explanations for they are not things per se.
Objection5: This is my defense of the infinte regression of dependent things. If you can use it here, then I can use it there.

I think that this list of responses is exhaustive, and all have serious problems. Am I wrong?

I think what I said in my previous comment applies to Response1 (and perhaps to Response 2, though I'm less sure of that). Where you're having trouble here is that you think I'm going to give some argument that GoT is necessary in order to show that the cosmological argument shows that GoT exists. I've done and intend to do nothing of the sort. I don't think the cosmological argument (at least my version, anyway) allows us to choose between GoT and GoP. It doesn't tell us what the self-existent thing is like. It just tells us that there is one.

(This is leaving aside the issue of whether GoT and GoP are the same being, with one descriptor just being a false descriptor that nevertheless succeeds in referring. The issue here is what makes God be like GoT rather than GoP. Even if the descriptor would refer but with false implications about the referent, you still have something to say in asking why God is one way rather than the other, because PSR seems to require an answer to that question. I'll frame the question the way you did it rather than my preferred way, but I don't think it matters.)

What I'm doing in bringing up GoT rather than GoP is offering my preferred account of what the self-existent thing is like. You're asking why GoT rather than GoP. If my view is correct, then there must be an answer to that. But that doesn't mean I need to have an argument for why I should believe that GoT is necessary. It may not be something I should be able to prove. I don't happen to think I can prove it. I happen to think it's true. It's my preferred view. It's just not something I think I've philosophically established.

Give that, I think you have to reword your objection to Response1 so that it's focusing on metaphysical issues rather than epistemological issues. But once you do that, I think it loses its force. The only way I can conceive of the rewording would be to say that whatever explanation you can offer why GoT is necessary would also apply to GoP, but that just doesn't seem right to me. The explanation for why GoT is necessary would be within GoT. You might think you can have a logical parallel within GoP, but that's ok. The kind of necessity we're dealing with here isn't logical necessity. It may be a logical possibility that GoP exists, but it can't be a metaphysical possibility if GoT exists. The explanation for that lies within the nature of GoT. What makes you think either could be true and thus necessary is that they're both logically possible. But what makes GoT necessary on my undefended view (which we're debating here only regarding its coherence and not on whether it's true) is not a matter of logical necessity but metaphysical necessity. At least that's one way to make the position coherent while maintaining PSR and the cosmological argument.

Leave a comment


    The Parablemen are: , , and .



Books I'm Reading

Fiction I've Finished Recently

Non-Fiction I've Finished Recently

Books I've Been Referring To

I've Been Listening To

Games I've Been Playing

Other Stuff


    thinking blogger
    thinking blogger

    Dr. Seuss Pro

    Search or read the Bible

    Example: John 1 or love one another (ESV)

  • Link Policy
Powered by Movable Type 5.04