The Modal-Epistemic Argument Defended: Reply to Wintein

Abstract

Recently, Stefan Wintein published an article in which he presents four objections to my modal-epistemic argument for the existence of God. His first objection is an alleged counterexample to the argument’s first premise, and the second objection is an alleged counterexample to the argument’s second premise. Wintein’s third objection attempts to show that the modal-epistemic argument is circular. Finally, the fourth objection is a parody objection. In this paper, I show that Wintein’s four objections all fail.

Introduction

Recently, Stefan Wintein¹ published an article in which he responds to my modal-epistemic argument for the existence of God.² He presents four objections to the argument. What I shall argue for is that all four objections fail.

Wintein³ provides the following formulation of the modal-epistemic argument, which accords to the formulation I provide⁴ and which for the sake of convenience I shall use in the remainder of this paper:

P_1″ For every FoCons proposition p that is true in some possible world w there is some possible world w′ (such that p is true in w′ and) such that there is a subject in w′ that knows_R that p.

P₂ It is not the case that there is a possible world in which there is a subject that knows_R that there is no personal first cause.

P₃ The proposition that there is no personal first cause is a FoCons proposition.

C In every possible world there is a personal first cause.

Here, a FoCons proposition is a first-order consistently believable proposition. A first-order proposition only states something about the world itself, such as that Amsterdam is the capital of the Netherlands or that John left Amsterdam. First-order propositions do not state something about the propositional attitudes of a subject.⁵ The propositions ‘John knows that he left Amsterdam’ and ‘Mary does not know that Amsterdam is the capital of the Netherlands’ are thus not first-order propositions. A consistently believable proposition is a proposition for which it holds that believing it is not self-contradictory.⁶ Propositions such as ‘There are no conscious beings’ or ‘Nothing exists’ are thus not consistently believable.⁷

For the modal-epistemic argument, I employ a specific conception of knowledge that requires near-certainty produced in such a way that it is not the result of dogmatism or other epistemically improper conditions. On the conception I employ, for a subject S to know that p is true, S cannot sincerely or genuinely doubt that p is true. That is, S cannot but believe that p is true. Such a degree of certainty is obtained when a proposition is (i) deductively proven, (ii) self-evident, (iii) grounded in indisputable sensory or inner experience (including indisputable memory), or (iv) based on indisputable testimony.⁸ In his article, Wintein refers to this specific conception of knowledge as knowledge_R. I shall use this notation in what follows. The structure of the rest of my paper is straightforward. In the next four sections, I shall successively present and refute each of Wintein’s objections against the modal-epistemic argument. I start with discussing Wintein’s fourth objection in the next section, and shall in subsequent sections discuss respectively his third, second, and first objections. The reason for this is that I want to deal with the least forceful objections first, which can be relatively swiftly done, before I move on to refute the more compelling objections, which will take a bit more time. The sixth and last section concludes the paper.

The Fourth Objection: a Parody Objection to the Modal-Epistemic Core Argument

In his article, Wintein⁹ argues extensively that I am committed to:

(*) A proposition p is knowable_R by some subject if and only if p is knowable_R by a personal first cause.

In the next section, I show that (*) is false. So I am not committed to (*). But for the sake of argument, I assume in the current section that (*) holds. Wintein¹⁰ points out correctly that (*) entails that premises P₁^″ and P₂ are logically equivalent to respectively P₁^* and P₂^*:

P₁^* All possibly true FoCons propositions are knowable_R by a personal first cause.

P₂^* The proposition that there is no personal first cause is not knowable_R by a personal first cause.

In its unabridged form, premise P₁^* states that for any proposition p that is true in some possible world w there is some possible world w′ such that p is true in w′ and such that there is a personal first cause in w′ that knows_R that p.¹¹ Wintein calls the valid argument with premises P₁^*, P₂^*, and P₃ and conclusion C the (modal-epistemic) core argument. Clearly, if we assume (*) the modal-epistemic argument is logically equivalent to the modal-epistemic core argument. But then the modal-epistemic argument is refuted if and only if the modal-epistemic core argument is refuted. More specifically, a successful parody objection to the modal-epistemic core argument extends to a successful parody objection to the modal-epistemic argument. Now, Wintein proposes the following parody objection to the modal-epistemic core argument.

K₁ All possibly true FolKri propositions are knowable_R by Saul Kripke.

K₂ The proposition that Saul Kripke does not exist is not knowable_R by Saul Kripke.

K₃ The proposition that Saul Kripke does not exist is a FolKri proposition.

C_K The proposition that Saul Kripke exists is necessarily true.

Here

[a] FolKri proposition is any proposition that is expressible by a sentence of a first-order language with identity, without any predicate or function symbols and with a single constant symbol k that refers to Saul Kripke. Thus the proposition that everything is self-identical, which can be expressed as “∀x(x = x)” is a true FolKri proposition. The proposition that not everything is identical to Saul Kripke, “¬∀x(x = k),” is another true FolKri proposition.¹²

Now, surely, it would be absurd to accept the conclusion that Saul Kripke exists necessarily, as Wintein rightly points out. We thus cannot accept the parody argument. But then, as Wintein’s parody objection goes, we cannot accept the modal-epistemic core argument either. And thus, if (*) were indeed true, which as mentioned I accept for the sake of argument in this section, we cannot accept the modal-epistemic argument itself. It is untenable and must be rejected.

Is this fourth objection convincing? I do not think so. The problem is that the proposed parody argument to the modal-epistemic core argument is in fact not a cogent parody argument at all. The reason for this is that, contrary to premise P₁^*, premise K₁ is obviously false. Let me explain why. Consider the following FolKri proposition: ∃x(x = x and ∀y(y = x)). This proposition reports that the world contains precisely one thing. There is also a FolKri proposition that reports that the world contains precisely two things, namely ∃x∃y(¬(x = y) and ∀z(z = x or z = y)). Similarly, the following FolKri proposition reports that the world consists of precisely three things: ∃x∃y∃z(¬(x = y) and ¬(y = z) and ¬(x = z) and ∀w(w = x or w = y or w = z). These formulas straightforwardly generalize. For each natural number n, there is a FolKri proposition that reports that the world contains precisely n things. Let me refer to such a proposition by World(n). Now, although Saul Kripke ‘may know his logic’ as Wintein¹³ rightly points out, knowing your logic is surely not sufficient to know how many things the world precisely contains—let alone knowing this with near-certainty. But then K₁ is indeed obviously false. There are many sufficiently large natural numbers n such that the FolKri proposition World(n) is possibly true and yet unknowable_R by Saul Kripke. I conclude that Wintein’s parody argument is not a successful parody argument against the modal-epistemic core argument. So, even if I for the sake of argument provisionally accept in this section that (*) holds, Wintein’s fourth objection against the modal-epistemic argument does not go through. It does not result in a successful refutation.

The Third Objection: the Core Argument Is Circular

Wintein’s third objection is also based on his claim that I am committed to (*). He argues¹⁴ that the modal-epistemic core argument already reveals itself—thus even without considering any parodies of it—that it cannot be accepted. But then, as (*) entails that the modal-epistemic argument is untenable if and only if the modal-epistemic core argument is untenable, it follows immediately that the modal-epistemic argument has to be rejected. As he writes:

First, it is trivially true that no proposition that implies that there is no personal first cause is knowable_R by a personal first cause. And so, in contrast to P₁^″, premise P₁^* wears its presupposition – that all [propositions that imply that there is no personal first cause] are necessarily false – on its sleeves. As such, the core argument thus allows us to see more clearly the main flaw of the modal-epistemic argument. Second, given the presupposition that all [propositions that imply that there is no personal first cause] are necessarily false, premise P₂ is effectively a tautology [,which] is clearly revealed by P₂^*. […] The core argument simply reveals that the modal-epistemic argument presupposes its conclusion in a more explicit way.¹⁵

In short, Wintein’s objection is that since the modal-epistemic core argument is clearly circular and thus untenable, it follows that the modal-epistemic argument is untenable as well.

Is this third objection successful? No, surely not. Because (*) is plainly false. A personal first cause may indeed be a very strong knower_R as Wintein points out. A personal first cause may know _R a great many propositions. In his article, Wintein successfully argues that I am committed to the former claim. He does this by providing an overview of various propositions I take to be knowable_R by a personal first cause, such as ‘There is a personal first cause,’ ‘A Spaghetti monster does not exist,’ and so-called Gödel propositions. But from this it does not follow that I am committed to the claim that each and every proposition knowable_R by some subject is knowable_R by a personal first cause.

And indeed, I shall show that this claim is false, i.e., that the left-to-right direction of (*) does not hold. Consider the following counterexample to the left-to-right direction of (*). Knowledge_R is a very strong conception of knowledge. Due to its quite demanding nature, it seems plausibly true that—to borrow from Thomas Nagel’s famous example¹⁶ —only a bat can know_R what it is like to be a bat. What it is like is the subjective character or quality of the experience. A bat could know_R this by indisputable inner experience. To know_R what it is like to be a bat amounts to first-person self-knowledge that is only available with near-certainty to bats or perhaps beings sufficiently similar to bats. A personal first cause is not sufficiently similar to a bat—let alone equal to a bat. Hence, a personal first cause is not capable of knowing_R what it is like to be a bat— neither by indisputable inner experience nor by any of the other sources of knowledge_R I have identified.¹⁷

Furthermore, it seems to me plausibly true that (at least part of) what it is like to be a bat can be expressed by a (complex) proposition q. But then proposition q is a counterexample to the left-to-right direction of (*). For q is possibly true and knowable_R by a bat, but unknowable_R by a personal first cause. So (*) is false. I conclude that Wintein’s third objection fails.

Yet, let all of this pass for the sake of argument. Let us assume for the sake of argument that my counterexample does not hold water. Would it then follow that Wintein’s third objection succeeds? This is not the case. Even then I am still not committed to principle (*). For again, the fact that a personal first cause is indeed a quite powerful knower_R only entails that many, and perhaps even the vast majority, of all knowable_R truths are knowable_R by a personal first cause. But this does nothing to establish that all knowable_R truths are so knowable_R. I can therefore remain agnostic as to whether this is the case or not. So even then the third objection fails.

Moreover, in addition to the counterexample I suggest above, one could think of other potential counterexamples, such as for example counterfactuals of (libertarian) freedom for human beings or other creaturely beings sufficiently different from a personal first cause. Counterfactuals of (libertarian) freedom are propositions of the form ‘agent S would (libertarianly) freely do action A in circumstances C.’ Agent S (libertarianly) freely does A if and only if S does A and S could have done otherwise. Agent S could have done otherwise just in case it was possible that in exactly the same circumstances agent S does something else than performing action A. The debate as to whether a personal first cause could know what a genuinely libertarian creaturely free agent would do once placed in certain circumstances is certainly not settled yet. The so-called middle knowledge accounts of how this might work are problematic— to say the least.¹⁸ It might therefore very well be the case that it is simply impossible for a personal first cause of reality to know counterfactuals of libertarian freedom of creaturely agents—let alone know_R them. For as earlier mentioned, the notion of knowledge_R is a quite demanding conception of knowledge indeed. It requires a form of certainty that is not easily obtained. Counterfactuals of (libertarian) freedom for creaturely beings constitute a counterexample to (*) in case they are knowable_R by a (libertarian) creaturely free agent but unknowable_R by a personal first cause.

Both counterexamples above are counterexamples to (*). Now, one may respond that in fact Wintein does not need (*) for his third objection. It is sufficient for him to posit the weaker.

(**) A FoCons proposition p is knowable_R by some subject if and only if p is knowable_R by a personal first cause.

Here, the non-trivial left-to-right direction only has it that knowable_R FoCons propositions are knowable_R by a personal first cause. Does positing (**) instead of (*) render Wintein’s third objection successful after all? In what follows, I argue that this is not the case. First, as long as no reasons— let alone sufficiently compelling reasons—are given for endorsing (**), one should remain agnostic as to whether (**) holds as a universal truth. In fact, the rationale I provided earlier for being agnostic about (*) applies— slightly modified — to (**) as well: Even if we take it that a personal first cause knows_R the vast majority of FoCons truths, it does not follow — pace the weight of tradition — that all FoCons truths are so knowable_R. Perhaps Wintein would at most be able to establish the left-to-right direction of (**) as a defeasible or exception permitting rule. But this does not restore the objection, since in that case the left-to-right direction of (**) is not deductively valid. It therefore no longer follows that the modal-epistemic argument is logically equivalent to the modal-epistemic core argument, which is essential for the objection to work.

Second, the aforementioned two counterexamples to the left-to-right direction of (*) can be slightly modified in order to refute the left-to-right direction of (**) as well. Let me explain. A first-order proposition is by definition a proposition that does not report a propositional attitude of a subject.¹⁹ That is to say, it does not report a propositional intentional state of the mind such as ‘S believes that P’ or ‘S knows that P.’ The restriction to first-order propositions blocks a specific class of counterexamples to the first premise of the simplified modal-epistemic argument, namely all counterexamples involving reports of propositional attitudes such as ‘John left Amsterdam and nobody knows it,’ ‘Nobody knows this proposition,’ and ‘There are no known propositions.’²⁰ Note that a first-order proposition may still report (an expressible part of) a mental state of a subject. It may report (an expressible part of) a non-intentional mental state (e.g., itches or moods) or (an expressible part of) a non-propositional intentional mental state (e.g., seeing a cat or imagining being on the moon). For in order to block the aforementioned specific class of counterexamples to the simplified argument’s first premise, it is not needed to restrict the scope of propositions even further to propositions that do not report mental states. It is sufficient to exclude propositions that report mental states towards other propositions. In other words, it is sufficient to exclude higher-order propositions.

Now, take the first counterexample. Proposition q reports (part of) what it is like to be a bat. What it is like to be a bat is not solely a propositional attitude. More specifically, q does not solely report a propositional attitude. Thus, there is a part of q that does not report a propositional attitude. Let us refer to that part of q as q₂. Proposition q₂ reports either non-intentional mental state(s) or non-propositional intentional mental state(s), or a combination thereof. But then q₂ is in fact a first-order proposition. As in the case of the original counterexample to (*), proposition q₂ is knowable_R by indisputable inner experience. So, q₂ is a FoCons proposition. And for similar reasons as before q₂ is knowable_R by a bat, but unknowable_R by a personal first cause. Hence, q₂ is a counterexample to the left-to-right direction of (**) and Wintein’s third objection still fails.

As to the second counterexample, counterfactuals of (libertarian) freedom are in fact FoCons propositions. For propositions of the form ‘Agent S would (libertarianly) freely do action A in circumstances C’ are consistently believable and do not report a propositional attitude of agent S. The second counterexample thus also refutes the left-to-right direction of (**) and Wintein’s third objection still fails. One may respond that it is not sufficiently clear that counterfactuals of (libertarian) freedom are FoCons propositions. Do they not state something about propositional attitudes after all, namely that in circumstances C agent S wants that or desires that ‘Action A is performed by subject S’? This is however not the case. As mentioned earlier, S (libertarianly) freely does A in C if and only if (i) S does A in C and (ii) S could have done otherwise. Agent S could have done otherwise in case it was possible that agent S in C does something else than A. Neither (i) nor (ii) reports a propositional attitude. The proposition in question is thus indeed a FoCons proposition. It only reports which action of a collection of two or more possible actions a (libertarian) free agent would perform when placed in certain circumstances. But then the second counterexample is a counterexample to the left-to-right direction of (**) as well.

The Second Objection: a Counterexample to Premise P ₂

The second objection is an alleged counterexample to premise P₂ of the modal-epistemic argument. Premise P₂ holds that it is not the case that there is a possible world in which there is a subject that knows_R that there is no personal first cause. In short, P₂ asserts that the proposition ‘There is no personal first cause’ is unknowable_R. Here is Wintein’s alleged counterexample:

[C]onsider a possible world in which there are (exactly) two personal uncaused causes, which we call A and B. These uncaused causes are co-creators of the world in the sense that everything that there is, except for A and B, (which are uncaused and so not created) is jointly created by A and B. […] If A and B co-created the world, they can plausibly know_R that they did on the basis of source (iii): their indisputable experience of creating the world. […] It is also plausible that a personal co-creator of the world can know_R that he is such a co-creator on the basis of direct intuition. Finally, […] it is also plausible that a personal co-creator of the world can know_R that he is such a co-creator on the basis of an indisputable inner experience of his own nature.²¹

Wintein maintains that to the extent that it is plausible that there can be a personal first cause that knows_R its own nature (i.e., being a personal first cause) by indisputable experience or direct intuition, it is also plausibly possible that there is a personal co-creator of the world that knows_R his own nature (i.e., being a co-creator of the world) in at least one of those same ways. Now, as Wintein argues, given that I am committed to the former possibility, I should also be committed to the second possibility. So, then, I have to concede that it is possible for there to be a subject (namely A (or B)) that knows_R that he is a co-creator of the world. Subject A knows_R that there is no personal first cause since there being a co-creator of the world implies that there is no personal first cause. This is implied because the mereological sum or fusion of person A and person B does not constitute a personal first cause. After all, sums of persons are not themselves persons.²² But then I have to concede that P₂ is false.

Is the second objection convincing? It fails because there is a significant difference between both cases. A personal first cause is the sole source of all being. Everything that exists originates directly or indirectly from such a person. A personal first cause sits at the root of reality. It is the metaphysical ultimate. All being ultimately comes from it. But then it is plausible that there can be a personal first cause that also knows_R that all being originates from him. A conscious being that is really positioned at the root of all being finds himself in a maximally ideal epistemic position. Without having to ‘reach out’ beyond himself to the world, he can obtain said knowledge by for example indisputable inner experience of his own nature. But then it does not seem to be impossible for there to be a personal being that also knows_R that he is the first cause of reality.²³

The situation of co-creator A is quite different. Person A is not the single source of all being. Since A is a co-creator he is not himself the ultimate root of reality. For person B has not been caused by him. Co-creator B and all beings caused by B have not been caused by A. Person A thus does not occupy the ultimate privileged position that a personal first cause of reality occupies in the order of being. But then how could A know_R that there is no personal first cause? Person A cannot deductively or logically prove that there is no personal first cause. For the proposition ‘There is no personal first cause’ is not a tautology nor a conceptual or analytical truth. Nor can A know_R that there is no such cause by intuition. For it is not self-evident or obviously true that there is no personal first cause of reality. Testimony will not work either. For only a personal first cause might act as such a witness. But in this case there is no personal first cause.²⁴ Could A perhaps rely upon indisputable inner experience of his own nature? No, this will not help either. Since A is a co-creator, his own nature does not encompass the whole of being. He thus has to ‘reach out’ to the world in order to obtain said knowledge. But how could A do this? What could A know_R by means of indisputable outer experience? In any case, it seems impossible for A to experience indisputably that B is uncaused. Person B’s testimony of being uncaused seems also not sufficient to bring A in a state of near-certainty with respect to B being uncaused. For how could A eliminate the possibility that B does not know_R himself or does not tell A the truth? Thus, A does not know_R that B is uncaused.

The only scenario left to consider is the following. Person A experienced that he was not caused. Person A experienced as well that he did not cause B. If the latter experience of A is indisputable, A knows_R that he did not cause B. If the former experience of A is also indisputable, A knows_R that he is uncaused. So, assuming that both experiences are indisputable, A knows_R the following two propositions: (a) A is uncaused, and (b) B has not been caused by A. From propositions (a) and (b), it follows that there is no first cause and thus also no personal first cause. I show how. Suppose for reductio ad absurdum that there is a first cause of reality. Person A is not the first cause because B has not been caused by A. But then A must have been caused. Yet, A is uncaused. We arrive at a contradiction. But then it follows that there is no first cause. One may hold that this is how co-creator A knows_R that there is not a personal first cause of reality. So, does Wintein’s second objection succeed?

No, this is not the case. First, how can co-creator A indisputably observe that he did not cause B? What A may have indisputably observed is the fact that he did not directly cause B. But from this it does not follow that he has not indirectly caused B via, say, a lengthy causal series that originates in A and includes B as a member. How could A rule out by indisputable experience that such a causal chain did not obtain? This seems impossible. But then A does not know_R proposition (b) and Wintein’s objection fails. However, let this pass for the sake of argument. Let us assume that person A knows_R proposition (b). Will this help the objector? No, it will not. The reason for this is that person A does not know_R proposition (a).

Let me explain. The fact that A experienced not being caused is not sufficient for him to know_R that he is in fact uncaused. For A cannot rule out with near-certainty that he has been caused by a cause that he simply cannot experience because it originates from some sector or dimension of reality that escapes him. After all, A does not occupy the root of reality. Therefore, even though he knows_R that he created a realm of beings together with person B, and even if we would assume — contrary to what I have argued above — that A might know_R that B is not directly or indirectly caused by himself, A does not know_R that there is no personal first cause. Indeed, for all person A knows he and B might each still be part of some causal series that terminates in a personal first cause. This state of affairs cannot be sufficiently epistemically eliminated by A. So he does not know_R that there is no personal first cause. Indeed, only a conscious being that solely sits at the ultimate root of all reality could potentially know_R what is ultimately the case with respect to the origin of the whole of reality. A conscious being that does not occupy that ultimate privileged position in the order of being, such as a co-creator, is not able to establish with near-certainty that there is no personal first cause.

I conclude that Wintein’s proposed counterexample to premise P₂ does not succeed. Now, perhaps one could still respond that both co-creators could know_R that they co-created the world— and thus that there is no personal first cause — in case we assume they are ‘unified’ or ‘integrated’ into a single being. Yet, this response fails as well because in that case we do not have two separate co-creators. Instead, we would have a single being that is the first cause of reality and that has multiple (namely two) personal identities. But then there is in fact a personal first cause of reality.

The First Objection: a Counterexample to Premise P ₁ ʺ

The first objection put forward by Wintein seems to me the most promising one. He purports to present a counterexample to premise P₁^″. That is to say, he aims to present a FoCons proposition that on my account of possibility is both possibly true and yet unknowable_R. If such a proposition can be found I have indeed no choice but to abandon premise P₁^″. For this premise has it that all possibly true FoCons propositions are knowable_R. The alleged counterexample put forward by Wintein is the following proposition: ‘The only conscious beings are naturally evolved animals and humans.’²⁵ In what follows, I shall refer to it as proposition Q.

I fully agree with Wintein that Q is unknowable_R.²⁶ So all comes down to the question of whether I am committed to hold that Q is possibly true. Proposition Q entails that there is no conscious being that is the first cause of reality (i.e., no personal first cause). Note that Wintein correctly points out that it thus cannot simply be asserted that Q is possibly true, since that would amount to a direct denial of the conclusion of the modal-epistemic argument — which according to Wintein and me is not a good objection.²⁷ However, Wintein does not simply assert that Q is possibly true. Instead of directly denying the conclusion of the argument, he argues that I am committed to accept that proposition Q is possibly true:

[Proposition Q] is both logically consistent and conceivable which, on a liberal notion of metaphysical possibility, are jointly sufficient for the (metaphysical) possible truth of [Q]. […] Rutten is clearly committed to a liberal notion of metaphysical possibility.²⁸

I am not sure what is precisely meant by a ‘liberal’ notion of metaphysical possibility. In any case I do not hold that all logically consistent and conceivable propositions are possibly true. My view is that our modal intuitions only warrant us to adopt the following defeasible rule: In the absence of evidence to the contrary, we may infer, about any logically consistent and conceivable proposition that it is possibly true. But let this pass. Let us assume that I am committed to logical consistency and conceivability being jointly sufficient for metaphysical possibility. Do I then have to accept that proposition Q is possibly true? No, I do not. For although proposition Q surely is logically consistent, I maintain that — assuming a notion of conceivability strong enough to entail metaphysical possibility — Q may not be conceivable. One could perhaps imagine one or more conscious beings arising from matter — similarly as one could imagine pigs flying through the sky to borrow an example from Chalmers²⁹ —but that may not be sufficient to conclude that it is conceivable that conscious beings arise from matter. For an account of conceivability strong enough to entail possibility may require more than perceptual or visual experience. Entertaining visual images may not be sufficient. For a proposition p to be conceivable in such a way that conceivability entails possibility it may be required to provide a coherent and sufficiently detailed explanatory account enabling us to cognitively understand how or under which circumstances proposition p could be true. In other words, one may need to offer an intelligible scenario, understandable configuration, or comprehensible model that — regardless of whether it actually obtains — would render p intelligible if it (counterfactually) obtains. Such an account of conceivability would require us to (i) grasp the concepts in the proposition and (ii) construe from our available background knowledge a comprehensible ‘mechanism’ that makes the proposition in question true. In other words, we may need to posit ‘explanatory mechanics’ for p. What may be required is a consistent narrative without substantial explanatory gaps that coherently expresses an understandable set of circumstances under which p would be true. In short, it may be needed to show in which way proposition p could be true. And doing so goes beyond entertaining visual imaginary.

Conceivability thus understood may not be a perfect guide to possibility. But then again, how else could we justify modal claims about what is possible? Precisely because conceivability is not a perfect guide to possibility, we should use it only as a default or prima facie rule for possibility. On this specific notion of conceivability, I am not committed to hold that proposition Q as proposed by Wintein is conceivable. In what follows I show why.

In the neuroscientific literature and in the philosophy of mind literature, one finds numerous theories of consciousness. The problem with all of these theories though is that they do not render proposition Q conceivable in the sense I have described above. None of these theories provides a narrative that makes it understandable and intelligible how matter could generate consciousness. They either do not even raise the question because they only address what David Chalmers³⁰ has famously called ‘easy problems’ (a systematic inquiry into the functional properties of conscious processes and the way these processes correlate with brain processes) or they settle what Chalmers calls ‘the hard problem’ (where does subjective inner experience come from?) by mere stipulation. As to the former: solving easy problems does indeed not provide us with an account of how consciousness could arise from mere matter. For theories, solving easy problems presupposes the existence of consciousness and merely aims to investigate its functional characteristics and neural correlates. Examples include the investigation of perception, memory, information integration, behavioral control, and attention. As to the latter, most if not all theories explain the coming into being of consciousness from mere matter by referring to ‘emergence’ or equivalents thereof. In short, these theories proclaim that simple material configurations evolved over time into complex material configurations and beyond some complexity threshold certain kinds of material structures gave rise to consciousness. But such accounts are not sufficient to render proposition Q conceivable. Speaking of ‘emergence’ is just a way of stipulating that subjective inner experience arises beyond a certain threshold. It does not provide an intelligible mechanism. There remains a huge explanatory gap between material structures and the coming into being of a wholly distinct kind: subjective inner experience. Material structures are from an ontological point of view nothing more than spatiotemporal arrangements of material particles. And however functionally complex those material structures might become over time through natural evolution, they remain in the end precisely what they are: solely material structures.

The fact is that we do not even have the beginning of a compelling model, mechanism, or account that renders Q conceivable in the way I have described above. And therefore it is unclear that Q is conceivable. But then I am not committed to the claim that Q is possibly true. I can leave the matter open and remain agnostic as to whether Q is possibly true or not. Thus, unless and until Q is rendered conceivable by some understandable narrative that explicates how Q could be true, Q does not constitute a counterexample to premise P₁^″. Therefore, Wintein’s first objection fails as well.

One may rebut that my response raises a profound difficulty for the modal-epistemic argument. For is it not on the above notion of conceivability also inconceivable how matter could arise from consciousness? If so, I would not be allowed to posit a possible world in which a personal first cause (being a conscious being) creates a (partly) material universe. This however goes too quick. Even if it would be inconceivable how matter could arise from consciousness, a problem for the argument does not occur. For I do not appeal to a possible world in which matter arises from consciousness in order to defend the argument’s premises or to refute objections against the argument. There are two types of possible worlds I appeal to. First, I appeal to possible worlds that do not refer to God but contain (i) planets made from iron, (ii) hermetically sealed non-transparent cubs, (iii) intelligent extra-terrestrial civilizations whose physics, technology, or visual apparatus is highly advanced — or whose members can observe an extensive geographical area on our planet with the same immediacy and resulting degree of certainty as we can observe our hands, (iv) stars that for some physical reason always emerge in pairs, or (v) observers that are able to observe particles without having to rely on light waves.³¹ None of these possible worlds refers to matter arising from consciousness. And there are no good reasons for why the possible worlds of the first type I appeal to would be inconceivable. Plausibly, these possible worlds are conceivable.

Second, I appeal to possible worlds that explicitly refer to God: a world where God exists and reveals his existence to someone by direct revelation; a world where God exists and knows he is God; a world where God exists, in which God is good and in which God knows his moral nature by indisputable experience in the form of introspection; a world where God exists and decides not to create anything; a world where God exists and decides to create exactly one causally inert external (material or immaterial) object; a world where God exists and knows all logical and mathematical truths by direct intuition, including all Gödel propositions.³² None of these worlds refers to God creating a material universe. Hence, similar to the first type of possible worlds, for the worlds of the second type I do not assume the possibility of matter arising from consciousness. Moreover, similar to the worlds of the first type, there are no good reasons for why the possible worlds of the second type would be inconceivable. There are no wholly unexplained ontological gaps. Given a possible world where God exists, it is not inconceivable that God could reveal himself to another immaterial person, or be good and know his moral nature, or know he is God, or create nothing, or create a causally inert external immaterial object, or know all logical and mathematical truths. Moreover, appealing to a possible world where God exists for the argument is not question begging or conceptually contradictory or otherwise problematic.³³ So, plausibly, the worlds of the second type I appeal to are conceivable as well. Moreover, strictly speaking I do not need to appeal to a possible world where God exists and reveals himself or where God exists and creates an external object. I can respectively appeal instead to a possible world where God knows he is God or where God decides to create nothing.³⁴ So even if appealing to both worlds is for whatever reason problematic, I can do without them.

To summarize: at no point do I appeal to the possibility of matter arising from consciousness to defend the premises of the argument or to derive the conclusion from them. It is only after the conclusion of the modal-epistemic argument has been established —that is, established without relying upon the possibility of matter arising from consciousness —that we can derive that it is possible for matter to arise from consciousness. Here is how. If there is a personal first cause in every possible world (as the conclusion of the modal-epistemic argument has it), the actual world contains a personal first cause. The actual world clearly also contains a material universe. But then the personal first cause—being a conscious being—is the (in)direct cause of the material universe. So matter can arise from consciousness. In any case, the earlier mentioned alleged profound difficulty for the argument does not arise— regardless of whether I am able or not able to provide a compelling narrative of how matter could arise from consciousness.

One may rebut that in that case the many correlations between brain states and mental states as empirically discovered by the neurosciences provide ample evidence that consciousness can arise from matter (as Q asserts) — even though we are not able to provide a compelling narrative of how this could in fact be true. This rebuttal is however ineffective because those correlations are epistemically compatible with both emergence accounts and mind–body interaction dualism and thus by and in themselves those correlations do not justify the claim that consciousness can arise from matter. After all, on mind–body dualism, matter does not produce consciousness.

But let this all pass. Let me for the sake of argument assume that I need to provide a reasonable model of how matter could arise from consciousness. In other words, how could a personal first cause —being a conscious being — create a material universe? Could God have created ex nihilo? Is creation ex nihilo possible? Theists, one may say, need a proper account of how this could happen. Or else theism becomes inconceivable. Well, let me bite the bullet and propose one.

How could God have created a material universe out of nothing? Is not that metaphysically impossible? Now, surely, ex nihilo nihil fit. From nothing nothing comes. This metaphysical principle seems to me indisputable. Yet, creatio ex nihilo does not violate it. For God creating a material universe from nothing is not the same as a material universe coming out of nothing. After all, God is the originating cause of the material universe and God clearly is something rather than nothing. More specifically, the potency to create a material universe can be understood to reside in God’s power. Nevertheless, one might respond that it is also metaphysically impossible for an agent (even as powerful as God) to create something out of nothing. But is that indeed the case? It might very well be metaphysically impossible for any agent, even God, to create concrete substances, such as planets, trees, animals, and humans, directly out of nothing. But that is not the way creatio ex nihilo needs to be understood. Here is an alternative picture of creatio ex nihilio that is not only in accord with modern science, but does in fact seem metaphysically possible (or at least less impossible than the first scenario). God created initially no concrete objects. In the beginning, God formed abstract objects from his thoughts. Afterwards, God reified or hypostatized these thoughts as separately existing abstract objects. These abstracta were subsequently transformed by God into non-substantive concrete particulars (e.g., one or more singularities, space-times having radius zero or virtual particles having mass and charge zero). From these non-substantive concrete particulars, God brought substantive concrete simple objects into being. These substantive concreta evolved over a long period of time into the world of concrete composite substances as we know it. This multi-pass picture of creatio ex nihilo might in fact be the only metaphysically possible way to create a world of concrete substances out of nothing. If so, it explains why God created our world via a lengthy process of evolution from some non-substantive singularity instead of bringing it directly into being out of nothing.

But again, let this all pass. Let me for the sake of argument suppose that the above model is not sufficiently intelligible to render creatio ex nihilo conceivable and thus (defeasibly or prima facie) metaphysically possible. Would the modal-epistemic argument then finally be defeated? No. For we could replace the first premise P₁^″ of the model-epistemic argument by the following weaker premise:

P₁^‴ For every FoCons proposition p that is true in the actual world w there is some possible world w′ (such that p is true in w′ and) such that there is a subject in w′ that knows_R that p.

This weaker premise only requires that actually true FoCons propositions are knowable_R. It is thus no longer required that all possibly true FoCons propositions are knowable_R. Together with the second premise P₂, it logically follows that there is a personal first cause in the actual world.

Moreover, all evidence I provide³⁵ for premise P₁^″ carries over to the weaker premise P₁^‴. Besides, the basic concepts of possibility, truth, and knowability are ontologically so fundamental that we should reasonably expect any central relationship between them to be maximally uniform and natural. Now, plausibly, the vast majority of truths in the actual world is knowable_R. Hence, it is reasonable to hold that all such truths are in fact knowable_R rather than there being some ad hoc exceptions. But this is exactly what the weaker premise P₁^‴ expresses.

So we still have a model-epistemic argument for God’s actual existence —even though it now no longer follows that God exists in all possible worlds. Whether this is so remains open. Wintein’s counterexample Q is not effective against this version of the modal-epistemic argument. For in order to refute this version of the modal-epistemic argument, it would obviously be question begging to assert that proposition Q is actually true instead of being merely possibly true.

On the basis of above considerations, I conclude that Wintein’s first objection fails. But since this is his strongest objection, I end this section with three additional remarks.

First, to successfully block Wintein’s purported counterexample Q (‘The only conscious beings are naturally evolved organisms’), I do not have to claim that Q is not possibly true. It is sufficient to remain agnostic as to whether Q is possibly true or not. Such agnosticism is justified even if I assume a notion of conceivability strong enough to entail possibility. For due to the explanatory gap as described above, on such a notion of conceivability one can withhold judgement as to whether Q is conceivable or not. Note that I thus also do not have to claim that Q is inconceivable.

Second, one may respond that on the assumed account of conceivability it is inconceivable that there are conscious beings other than naturally evolved organisms. So that there being a personal first cause becomes inconceivable. That might be so, but that does not affect the modal-epistemic argument. The argument only needs the weaker claim that a personal first cause is possible.³⁶ One is not committed to hold that it is also conceivable by humans in the strong sense as described above. For there is no good reason to accept the reverse principle that possibility entails such a strong sense of human conceivability. What is claimed is that possibly there is a conscious being that is the first cause of reality. This claim is not unreasonable.³⁷ For one thing, one does not have to assume that the necessarily existing omni-God of classical theism is possible, which would surely be far more controversial. Nor does adding this claim as a premise render the modal-epistemic argument circular or otherwise epistemically insignificant.³⁸ After all, the argument, if successful, takes us from a personal first cause’s possible existence to its actual or even necessary existence. Wintein, to whose objections I reply, seems to agree. For notably none of his objections attacks the assumption that a personal first cause is at least possible. Furthermore, for a proposition to be conceivable in the assumed sense, an intelligible explanatory mechanism that makes it true needs to be provided. But on such a strong account of conceivability it might actually, regardless of whether a personal first cause is possible, not be unreasonable to hold that a personal first cause is inconceivable. For how could such a mechanism be provided in the case of an ultimate first (uncaused) cause?

Third, one may argue that even if we exclude for good dialectical reasons³⁹ all counterexamples to the first premise that involve a blunt denial of the modal-epistemic argument’s conclusion (e.g., ‘There is no personal first cause’), the first premise is still obviously false if naturalism is necessarily true (allegedly because if the only possible conscious beings are naturally evolved organisms, we simply lack sufficiently powerful possible knowers to ensure that all possible truths are knowable), so that the first premise still begs the question against necessarily true naturalism. I would respond that in order to obtain an actual real refutation of the first premise and so to effectively force the proponent of the argument to abandon the first premise, the objector seems here to need to presume the necessary truth of naturalism and thus the falsehood of the conclusion of the modal-epistemic argument. If so, this would be dialectically inappropriate. Wintein and I agree that in the present context, of course, the falsehood of the conclusion of the modal-epistemic argument cannot be presupposed.⁴⁰ Various reasons have been offered in support of the first premise that are each independent of whether naturalism is necessarily (or actually or possibly) true.⁴¹ Also a wide range of alleged counterexamples —such as Gödel propositions or propositions about the elementary particles of quantum physics — have been refuted.⁴² These refutations are also independent of whether naturalism is necessarily (or actually or possibly) true. For it can be shown that of those alleged counterexamples, the ones that only involve naturalistic propositions are (or can be) refuted by appealing only to possible (and in some cases extremely cognitively or physically advanced) naturally evolved knowers. This all warrants at least prima facie the adoption of the first premise. Thus, if one wants to cogently refute the first premise, then instead of proposing counterexamples that involve a blunt denial of the argument’s conclusion or merely purporting to point out that the first premise does not sit well with naturalism, one has to show that those supporting independent reasons are bad, or one has to gather independent evidence of a positive sort that there really are exceptions to the first premise. The latter is precisely what Wintein aims to do in his paper⁴³ and earlier in this section I argue that his attempt is not successful. But let these considerations pass. More importantly, I would argue that as long as blunt denials of the modal-epistemic argument’s conclusion are dialectically appropriately excluded as counterexamples to the first premise, it might in fact not follow that the first premise is obviously false if naturalism is necessarily true. After all, for each possibly true naturalistic proposition, it might very well be the case that there is always a possible world in which that proposition is true, in which there are natural laws that happen to be favorably suited for knowing it, and in which there are naturally evolved conscious beings that are intellectually and scientifically advanced and knowledgeable enough— and conceptually or geographically close enough to the vicinity of its relevant truth-makers — to actually know it. But let this pass as well. For even more importantly, as earlier mentioned, the modal-epistemic argument deploys in fact the premise that a personal first cause is at least possible, so that naturalism being necessarily true is excluded by the premises of the modal-epistemic argument. And as stated as well, adding the mere possibility of a personal first cause as a premise is not unreasonable and does not render the modal-epistemic argument epistemically irrelevant. As to the closely related objection that the first premise begs the question against actually true or possibly true (but not necessarily true) naturalism, I would respond that here again the objector seems to have to bluntly deny the argument’s conclusion — which is as said dialectically inappropriate— in order to force the proponent of the argument to actually abandon the first premise. But let this pass. More importantly, if we again dialectically appropriately exclude all counterexamples that involve a blunt denial of the conclusion of the modal-epistemic argument, it becomes clear that in this case the first premise is not obviously false at all. For now, one can appeal to the possible existence of a personal first cause and such a being is a quite powerful knower indeed. Such a knower can be invoked to refute many alleged counterexamples to the first premise. I thus conclude that the closely related objection fails as well.

To conclude this section, let me briefly say two further things about the argument. First, even if the first premise P₁″ turns out to be untenable after all, one may, as mentioned, replace it by the weaker P₁‴. In that case, we still get the striking conclusion that if a personal first cause possibly exists, it also actually exists. Since necessary existence is not part of the definition of a personal first cause, this does not reduce the argument to an ontological argument of sorts. Second, I surely do not claim that the modal-epistemic argument is on its own sufficient to establish God’s existence. If successful, it merely increases the overall likelihood of God’s existence. The argument can then be included in a cumulative case for theism and this case would then need to be compared with a cumulative case against theism. But this is not part of the scope of this article.

Closing Remarks

In this paper, I presented and discussed Wintein’s four objections against my modal-epistemic argument for God’s existence. Two of his objections involve an alleged counterexample to respectively the first and second premises of the unabridged version of the argument. Wintein’s other two objections are respectively a parody objection to what he calls the core argument (which extends to a parody objection to the modal-epistemic argument) and the claim that the core argument already clearly reveals itself that the modal-epistemic argument is circular (i.e., presupposes its conclusion). I argued that all four objections fail. The modal-epistemic argument thus still stands.