Home » Library » Modern Library » The Presumption of Atheism Revisited

The Presumption of Atheism Revisited


Antony Flew’s “Theology and Falsification” has become one of the most widely discussed articles on contemporary philosophy of religion. Many anthologies and textbooks included it, and continue to include it today. Twenty years ago it was republished on the occasion of its Golden Jubilee anniversary.[1] Its republication and the responses to it seem to bear out what many have suspected about the fate of the article. It seems to have acquired a reputation as an example of a wrongheaded use of a contemporary empiricist criterion: falsifiability in principle. In the article, it seemed to many that Flew was trying to apply the principle to theological statements such as “God exists” or “God loves us.” To understand why it seemed he was doing so, consider the following passage:

Now to assert that such and such is the case is necessarily equivalent to denying that such and such is not the case. Suppose then that we are in doubt as to what someone who gives vent to an utterance is asserting, or suppose that, more radically, we are sceptical as to whether he is really asserting anything at all, one way of trying to understand (or perhaps it will be to expose) his utterance is to attempt to find what he would regard as counting against, or as being incompatible with, its truth…. [A]nything that would count against the assertion or which would induce the speaker to withdraw it and to admit that he had been mistaken must be part of (or the whole of) the meaning of the negation of the assertion.[2]

It has been repeatedly stated by its most famous proponent, Karl Popper, that this idea was not originally designed to be used as a criterion of cognitive significance, even though some philosophers who learned about it from reading Popper’s works tried to use it in that way. In his 1962 book Conjectures and Refutations: The Growth of Scientific Knowledge (especially in chapter 1) Popper explains forcefully that he meant for it to be used a means of demarcating the class of statements that can belong to science from the class of statements that couldn’t belong to science. He did not mean for it to be used to distinguish cognitively significant from cognitively meaningless statements and argued against doing so. However, it has yet to be clearly demonstrated that being unfalsifiable never can affect any statement’s putative cognitive status. I will return to this point much later in this paper in regard to the status of theological statements. The main part of this paper, if it succeeds in its aims, will make it clear that Flew’s basic idea in “Theology and Falsification,” though flawed, is more important than has often been supposed and still highly relevant to current discussions in the philosophy of religion.

In his 1985 article “The Burden of Proof,” Flew refers specifically to Popper’s philosophy of science and denies that he (Flew) was using falsifiability as a criterion of meaning any more than was Popper. Flew then tries to explain why the charges of misusing the criterion of falsifiability are false. He says:

What I was doing there was developing a challenge to believers … to begin to make clear, not least to themselves, what it was they were asserting…. [T]he particular meanings being attributed to the utterances by particular utterers must depend upon what several answers these several persons took to be appropriate. This contention is itself an immediate consequence of that note’s sole general thesis about meaning: the tautology that any proposition must be equivalent to the negation of its own negation.[3]

Flew states that it is not “necessary in order to be in a position to make individual judgements about meaningfulness and meaninglessness, to cleave to any comprehensive theory of meaning.”[4] He adds:

As for the alleged need to develop and defend a general theory of meaning before attacking any particular questions of significance, this is surely a classic case of putting the cart before the horse. For, were we truly unable to answer questions of this sort without reliance on such a theory, we could never test whether any theory was adequate to the facts which it was supposed to explain.[5]

This part of Flew’s position on the use of the idea of falsifiability is tolerably clear.
Consider an analogous position in regard to making ethical judgments. One need not have a fully developed theory of normative principles in order to make responsible and justified judgments about the rightness or wrongness of particular actions. Rather, our normative principles need to be developed in such a way as to at least capture our most strongly held moral judgments about particular cases of actions or circumstances. Flew says that what he was doing in his 1955 “Theology and Falsification” was pointing out that those who make theological assertions ought to explain whether they are falsifiable or not and, if not, explain how that fact does not have disastrous consequences for the meaningfulness of those assertions. Raising that issue with regard to those specific assertions does not necessarily mean that Flew was then assuming that falsifiability was the criterion of factual meaning or even a necessary condition of factual meaning. Flew could not plausibly deny that his own sympathies with contemporary forms of empiricism would make some such criterion of meaningfulness attractive to him. But that point is not directly relevant to the charge of taking falsifiability as a meaning criterion for granted.

What is still not clear is Flew’s claim that the kind of challenge he was making to those particular utterers is an immediate consequence of the logical principle of double negation. The relation of logical equivalence between any proposition and the negation of its negation is based on the purely truth-functional notion of negation. So, the tautology in question merely states that it is necessarily the case that any conditions in which a proposition is true are the same conditions in which the negation of its negation are true. But this fact does not seem to have anything to do with the meaning of any statement. How does it follow from the fact that p is logically equivalent to not-not-p that anyone who knows the meaning of p must know the meaning of not-p? Was Flew saying that anyone who knows what the conditions are in which p is true must know what the conditions are in which p would be false? Or, was he saying that anyone who knew what it would be like for p to be true must know what it would be like for p to be false? Or, that anyone who understood “p” would be able to state some conditions in which p would be false? Or, to state some conditions in which p could not be true?

Perhaps some of these as well as other questions about Flew’s aims in “Theology and Falsification” might be answered by examination of and reflection on another of Flew’s writings, “The Presumption of Atheism,” chapter 2 of his God, Freedom, and Immortality. In the main part of this paper, I will critically examine some issues that are raised in that later chapter. In “The Presumption of Atheism,” Flew introduced a distinction between what he called two types of atheism: positive atheism and negative atheism. He expresses a preference for negative rather than positive atheism. One goal of “The Presumption of Atheism” is to justify this preference. Flew defines positive atheism as “the assertion that there is no such being as God.” He defines negative atheism as having a meaning that results from the prefixing of the Greek particle “a-” to “theism.” This meaning can be made precise by comparing the meanings of other terms that have the same prefix, e.g., “atypical,” “asymmetrical,” and “amoral.” Thus, a negative atheist is simply one who is not a theist, or someone without a belief in God. Flew is not the only philosopher in the literature of the philosophy of religion attracted to negative atheism. One prominent example of such a philosopher is Michael Martin.[6]

There seem to be strong prima facie objections to defining “atheist” as a negative term. Theodore M. Drange has objected to Martin’s version of negative atheism in the following way:

First, it is not completely clear that the correct translation of the Greek prefix
“a” is “without.” It might also mean “no,” in which case “a-the-ism” could be translated as “no-god-ism,” or “the view that there is no god.”… Second, even if the etymology of the word “atheism” did indicate that it once meant “without belief in God,” that is still not a good guide to current usage…. It seems far more important what people mean by a word today than what it once meant a long time ago.

…[I]t could be said that I (and most people) use the term “atheist” in the sense of “positive atheist.” [Third] [i]t should be noted that all positive atheists are automatically negative atheists, which may sound somewhat peculiar when those expressions are used.[7]

One may add a concern to Drange’s objections. Does Flew help his case is by including “amoral” in his list of examples that he wants us to compare with “atheist”? The term “amoral” is not just logically and semantically opposed to “moral,” it has the sense of “neither moral nor immoral.” If one is guided by that comparison, is a negative atheist, then, someone who is neither a theist nor an atheist? Thinking on this question may incline one to suggest that what Flew really means by a negative atheist is simply what most people mean by “agnostic.” But Flew firmly states that the sort of negative atheism he wishes to defend is not what is ordinarily meant by “agnosticism.” He suggests that what T. H. Huxley originally meant by “agnostic” when Huxley invented the term is close to his notion of a negative atheist, but I will not pursue that lead here. Flew does clear things up somewhat when he writes:

[F]ollowing the present degenerate usage, an agnostic is one who, having entertained the proposition that God exists now claims not to know either that it is or that it is not true. To be an agnostic in this ordinary sense, you have to have conceded that there is, and that you have, a legitimate concept of God; such that, whether this concept does or does not in fact have an application, it theoretically could. But the atheist in my peculiar interpretation, unlike the atheist in the usual sense, has not as yet and as such conceded even this.[8]

Perhaps the comparison of “atheist” with “amoral” is more helpful than it initially seems to be. Unlike the agnostic in Flew’s sense, an amoral person is not someone who has considered the evidence for being moral or not being moral and then found the evidence inconclusive for either alternative. An amoral person is someone who has no grasp of morality in the sense of culturally embedded customs and practices. He/she does not specifically reject morality. One who does that would more properly be called an immoralist. Yet an amoral person is not a moral person either because of lacking any understanding of morality. An amoral person is neither moral nor immoral in the sense just presented. Perhaps Flew is comparing atheism to amorality as we have just explained it. An atheist, in the negative sense, is one who neither affirms nor denies that theism is true or false because he/she does not see that there is even a legitimate concept of God, given the way that the concept has been presented in the tradition known as classical theism. I will return to this suggestion later in this paper.

Still, one may share Drange’s misgivings about construing atheism as a negative concept. In order to take the idea of negative atheism seriously, one must supplement Flew’s characterization of it. As it stands, it is an inadequate characterization of any sort of atheism because it makes all sorts of things atheists, with ridiculous results. For example, it makes atheists of newborn babies, amnesiacs, profoundly retarded people, and dogs. Some may be so determined to defend this conception of atheism that, unlike me, they are willing to accept such consequences. Unlike Flew, I will try to construct an enhanced version of negative atheism that does not have such consequences. I suggest that what is needed is to construe negative atheism as a kind of epistemic stance that can be justifiably taken. Part of this construal would, accordingly, be that a negative atheist is not merely someone who lacks a belief in God. A negative atheist is also someone who knows very well that he/she lacks such a belief. But even more than this must be added. Further additions will be suggested in the next few paragraphs.

The main thrust of Flew’s argument for adopting negative atheism is
related to what he thinks is a matter of good methodology in designing the way that a fair debate about the existence of God ought to proceed. Such debates, he argues, ought to be designed by analogy with civil trials in the English common law tradition, in which there is a presumption of innocence for the accused. This presumption entails that the plaintiff bears the burden of proof. Should the arguments of the prosecution fail, the proper verdict will not be “innocent” but, simply, “not guilty.” Flew maintains that in a debate about the existence of God there should be a presumption of atheism. Theist debaters should, therefore, bear the entire burden of proof by presenting arguments for the existence of God. The atheist debater should only be presented with ample opportunity to destroy the arguments of the theist and, thus, turn back attempts to defeat the presumption of atheism.

Part of Flew’s reason for advocating this approach is not difficult to make out. Under a presumption of innocence, the accused is not required to concede anything that might help the prosecution. For example, the accused is not required to concede at the outset, e.g., that he/she could have committed the crime, that he/she knows anything about the crime, that a crime of some sort occurred, etc. Similarly, Flew maintains, no atheist should be required to concede at the outset of a debate that there might be evidence for the existence of God, that theism is at least possibly true, or even that there is a legitimate concept of God. Consequently, the theist in the debate should bear the entire burden of proof in the debate, “first to introduce and defend his proposed concept of God; and, second, to provide sufficient reason for believing that this concept does in fact have an application.”[9]

The existence of God, Flew argues, is an issue that is not in the same class as the issue of the existence of the Loch Ness Monster or an Abominable Snowman. In those cases, there is no controversy over the concepts of the entities in question. Not so in the case of the existence of God. It is conceding more than an atheist should be required to concede to grant (even tacitly) that “even the would-be mainstream theist is operating with a legitimate concept which theoretically could have an application to an actual being.”[10] To allow the legitimacy of only positive atheism in a debate about the existence of God is, therefore, to concede more than the atheist ought to be required to concede because positive atheism presupposes that the concept of God that the theist operates with is a legitimate concept. So, Flew’s position is that the very conditions of a fair debate fully justify taking a position of negative atheism and, by that very fact, placing the whole burden of proof on theism.

There are other philosophers who insist that there is no burden of proof on atheism and that the onus of proof lies entirely on theists. Michael Scriven and Norwood Russell Hanson have taken such a position. However, they do not do so for the same reasons that Flew gives.

Scriven says:

If there is no evidence which points to a particular claim, although some general background considerations make it not too unlikely that something like this should be true (Loch Ness monster, mile record broken twice in 1980) we should say there is some general support for the claim, we shall say it is wholly unfounded if there is no evidence for it in particular and no general considerations in its favor and disprovable if it implies that something would be the case that is definitely is not the case…. [B]oth ‘unfounded’ and ‘disprovable’ correlate with atheism….

Recalling that to get even a little evidential support for the existence of a Being with supernatural powers will require that the little be of very high quality (little does not mean “dubious”), we see that the failure of all the arguments, that is, of all the evidence, will make even agnosticism in the wide sense an indefensible exaggeration of the evidential support.[11]

Scriven, however, has no doubts about the legitimacy of the concept of God.
In essence, he says that, if we can argue successfully (as we have) that the evidence supports claims of the existence of beings with less than human power, such as various prehistoric lifeforms, then there is no reason why one could not argue, from some body of evidence, that there are beings with powers greater than those of human beings. Scriven says, “If such a Being exists, we might or might not be able to go on to argue that it is infinitely powerful, etc.”[12]

So, Scriven says that we should reject claims that God, by his nature “transcends merely human categories of thought.”[13]. Scriven’s strategy in building a case for atheism is to define a being he calls “Basic God”—who is inferior to the God of classical theism—and then to show that the evidence for the existence of even Basic God is drastically insufficient. If atheism in regard to the existence of Basic God is justified, then atheism in regard to the God of classical theism is, he claims, certainly justified.

Hanson’s position is quite similar to Scriven’s. Hanson finds it altogether peculiar and puzzling that so many people seem willing to admit that there are no rational grounds for theism but then, in apparent triumph, issue a challenge to atheists to prove that God does not exist. In one sense of “proof,” Hanson states, it is trivially obvious that there can be no proof of the nonexistence of God. But in this same sense of “proof,” there can be no proof of the nonexistence of fairies, goblins, or unicorns, either. This sense of “proof” is deductive proof: i.e., arguments with true premises that contradict the negations of their conclusions. But this sense of “proof” is not even the one that is usually meant by the term “proof.” Here Hanson reminds us that:

[P]roof of the non-existence of X usually derives from the fact that there is no good reason for supposing that X does exist…. [T]he lack of any formal deductive proof that God does not exist provides no reason whatever for supposing that God does exist…. [A] proof that God does not exist turns on no more than the demonstration that there is no good reason for supposing that he does exist…. [I]f atheism is erroneous, it is so in the manner of claims like “Boston is south of Miami” and “Eagles outweigh elephants”.[14]

At one point, Hanson seems to come close to stating doubts about the concept of God that recall the doubts at which Flew hints:

[T]he reader must be able to cite some actual happening, some genuine experience, some de facto description of events which would be relevant to the conclusion that God exists. Because if no such factual description, however rich and subtle, and theologically sophisticated, could even be germane to further confirming the claim “God exists” for a believer, then it could never be reasonable for him to believe in God’s existence.[15]

Hanson goes on to insist that some factual data must be relevant to confirming the claim that God exists if “God exists” is to have any semantical content.

But Hanson clearly thinks that some factual data would be relevant to confirming the claim that God exists, and he even gives some (vivid) examples of events which, were they to occur, would go a long way toward confirming this claim for him. So, Hanson does not have any doubts about the actual cognitive meaningfulness of the claim. Given the lack of factual evidence for the claim, Hanson observes that the only other way of circumstantially strengthening the case for theism would be for something like the traditional arguments for the existence of God to succeed. Of course, Hanson thinks that they do not succeed and that they have been so thoroughly discredited (even by many theologians) that atheism is the only reasonable alternative left. Given the fact that there are also no empirical facts that provide good reasons for theism, Hanson says, one is by that very fact justified in denying the existence of God just as, on similar grounds, we are justified in denying the existence of unicorns. Hanson argues that the sort of agnosticism that allows that theism could be confirmed, but that the evidence for it happens to be insufficient, must avoid making the mistake of then going on to deny that atheism can be confirmed by any possible body of evidence. It is confirmed by the very fact that theism has little or no general or particular rational support. His argument for the justification of atheism thus shows how he puts the entire burden of proof on theism.

We may call positions like those of Scriven and Hanson unburdened positive atheism. It seems to me that the main difficulty for their versions of atheistic argument strategy is explaining why positive atheism should not have to bear any burden of proof of its own. It is one thing to reject demands for deductive proofs of atheism. It is quite another to maintain, as Scriven and Hanson do, that atheists have fully justified their position just by demonstrating the lack of evidential support for theism.

Even if one accepts the claim that demonstrating the lack of evidential support for theism provides some justification for atheism, there is still some point to the demand that the positive atheist also provide other, distinct reasons for atheism given that this burden has been placed on theists. In part, one must have an appropriate reply to those who will raise some sort of demand for basic fairness. Why, it will be asked, should theists have to do all of the work in the debate? Such a demand would reveal a certain misunderstanding. It should go without saying that any atheist debater will have plenty of work to do just in making effective replies to the arguments for theism.

Still, one may argue for a burden of proof for positive atheism as follows: Arguments for a claim ought to be distinct from successful criticisms of arguments for a claim’s contradictory. Many atheists have in fact produced arguments for positive atheism that are not specifically destructions of arguments for theism. For example, some atheists have argued that a naturalistic metaphysics is inconsistent with theistic metaphysics, and then presented arguments in defense of the superiority of a naturalistic metaphysics. Examples of defenses of naturalistic metaphysics include arguments for purely physicalistic theories of mind, as well as arguments that naturalism is more likely to be true than antinaturalistic metaphysics given the fact of the evolution of all life forms by natural selection. Clearly, unburdened positive atheists should explain why they need not bear any such burden of proof. After all, there are positive atheists who have taken on just such a burden. Flew seems to be opposed to accepting this sort of position and part of his reason for his opposition is specifically related to insuring basic fairness in a debate about the existence of God. In order to understand his opposition fully, we need to take some of his other explanations into account.

Flew says that there are clear differences between “presumption and “assumption.” Arguing a case under a presumption of innocence does not mean assuming at the outset that the accused is innocent. To see why it does not have this meaning, consider the fact that a trial that rendered a verdict of “guilty” would show that an assumption that the accused is innocent was a mistake. The same verdict would not in any way show that a presumption of innocence was a mistake. Flew argues that, in the same way, the presumption of atheism “is neutral as between all parties to the main dispute, in as much as to accept it as determining a procedural framework is not to make an substantive assumptions.”[16] However, he warns the reader that debating under a presumption of atheism for the reasons he presents “puts the whole enterprise of theism in a fresh perspective that makes it even more difficult and precarious than it did before.”[17]

Flew does clearly explain why theists bear the whole burden of proof in a debate about the existence of God. Given his negative version of atheism, the atheist has no belief to defend. The negative atheist is simply one who lacks belief in God. The atheist makes no knowledge-claim with respect to the issue of God’s existence at all. He does not deny the existence of God. Hence, he cannot be required to prove that there is no God, even when he is engaged in the debate. It is the theist who makes a knowledge claim insofar as a theist participates in a debate about the existence of God, The theist is not merely someone who professes belief in the existence of God. In a debate, this belief must be defended and justified by some sort of positive argument. Merely lacking a belief need never be justified unless it can be shown that one ought to have that belief. All of us lack an enormous number of beliefs but we are certainly not required to justify each and every one of these doxastic absences. It has to be shown, in each case why one ought to have this or that belief. It cannot be one’s epistemic obligation to justify not having a belief that p merely because one does not have the belief that p. This point, Flew explains, follows from the usual distinction between true beliefs and knowledge. It is the theist who not only has the belief but insists that he knows the proposition he believes. Thus only he must give reasons for it.

In a later contribution also titled “The Presumption of Atheism,” Flew makes a helpful restatement of the position he presented in the earlier chapter 2 of God, Freedom, and Immortality titled “The Presumption of Atheism.” In particular, he makes the analogy between the presumption of innocence in English common law and the presumption of atheism much clearer than it was in the earlier version. He distinguishes three elements in the analogy:

  1. “[I]n both of these contentions about the burden of proof, the word ‘proof’ is being used in the ordinary, wide sense in which proofs embrace any and every variety of sufficient reason.”[18]
  2. Both presumptions are defeasible: “they are, consequently, not to be identified with assumptions. The presumption of innocence indicates where the court should start and how it must proceed.”[19]
  3. “Because these are not assumptions but contentions about the burden of proof, that people succeed in proving what they were challenged to prove does not even begin to show that the burden was wrongly placed upon their shoulders.”[20]

For the sake of clarity, it is worth quoting Flew on this final point at length:

To adopt a presumption about the burden of proof is to adopt a policy. And policies have to be assessed by reference to the objectives and the priorities of those for whom they are being proposed. Thus the policy of presuming innocence is rational for all those for whom it is more important that no innocent person should ever be convicted than that no guilty person should ever go free…. [T]he objective by reference to which the policy of accepting the presumption of atheism has to be justified is the attainment of validation of knowledge about the existence and activities of God, if such knowledge is indeed attainable. The inquiries pursued under this procedure are directed either toward acquiring such knowledge or showing who, if anyone, is already possessed of it…. The presumption requires us to begin by examining any proposed or presupposed conception of God as if we were meeting it for the first time; considering, that is to say, whether it is coherently applicable and, if so, inquiring what evidencing reasons would be necessary and sufficient to establish that it does in fact have application.[21]

In order to support his claim that a presumption of atheism is neutral as between all parties, Flew he gives an impressive example of a theist who he says readily accepted a presumption of atheism in the very same spirit of fairness that Flew wants to insure. The theist in question is Thomas Aquinas. Flew calls attention the fact that an integral part of Aquinas’ procedure in presenting his five supposed proofs for the existence of God is to first pose objections to the thesis that God exists, objections that Aquinas himself formulates. The proofs are then presented as replies to these objections.

The first of these objections reads:

It seems that God does not exist. For if of two contrary things one were to exist without limit the other would be totally eliminated. But what is meant by the word ‘God’ is something good without limit. So, if God were to have existed, no evil would have been encountered. But evil is encountered in the world. Therefore, God does not exist. (Summa Theologica I, Q2, art 3)

This argument for atheism is Aquinas’ way of presenting the traditional problem of evil. In the same article Flew also calls attention to the second of the objections that Aquinas formulates:

Furthermore, what can be accounted for by fewer principles, is not the product of more. But it seems that everything that can be observed in the world can be accounted for by other principles, on the assumption of the non-existence of God. Thus natural effects are explained by natural causes, while contrived effects are referred to human reason and will. So there is no need to postulate the existence of God. (Summa Theologica I, Q2, art 3)

Notice that, in both objections, Aquinas presents arguments. In the first objection, he presents an argument of the modus tollens type that purports to be a direct refutation of theism. The other argument, presented in the second objection, is not one that purports to be a direct disproof of theism. It is an argument that theism is not the only reasonable explanation of what may be observed in the world, and that another explanation of the observable facts, one available to naturalistic atheists, may appear to be sufficient to account for the same facts. Furthermore, the argument directly implies, since theism requires assuming more principles than atheism does, it may appear that explanations available to atheists are more reasonable than those available to theists. Aquinas’ replies to these two objections are of two sorts. The reply to the first objection proceeds by trying to show that one of the premises (if God were to have existed, no evil would have been encountered in the world) is false. The reply to the second involves an attempt to show that not all natural effects can be explained by natural causes and thus that explanations consistent with atheism are insufficient to account for the observable facts.

Flew seems to think that one of his strongest points, given Aquinas’ procedure, is that no one could accuse Aquinas of beginning with an assumption that God does not exist. All that one could reasonably interpret Aquinas to be doing is adopting a presumption of atheism as part of a methodology of argument, even though he obviously is aiming at making the strongest possible case for theism.

But Aquinas’ strategy seems to be the following: Begin by presenting the strongest known arguments for atheism without actually endorsing those arguments. Then refute those arguments. Last, argue for theism. Is this the same procedure as the one Flew means by adopting a presumption of atheism? There is some resemblance in that Aquinas certainly does not endorse the arguments stated in the objections. We might say that what Aquinas does in the objections is to display specimens of atheological arguments in order to make it clear what sort of task faces a theist who wants to show that theism is more reasonable than atheism. Moreover, it is clear that Aquinas accepts a burden of proof for theism. Finally, it does seem that Aquinas wants to show that theism can be decisively defended even when the most powerful kinds of atheological arguments are given the privilege of being stated at the outset. All of these facts support the interpretation that Aquinas adopts a presumption of atheism as a methodological procedure. It is plausible to interpret Aquinas as trying to make his case for theism even more powerful than it would be if he simply presented his supposed proofs. He may well have thought that the case for theism is more powerful when it is first shown that atheism fails even when it is given the rhetorical advantage of putting its best foot forward at the very beginning of the debate.

However, is the sort of atheism against which Aquinas argues the same sort of atheism that Flew advocates? Since these arguments for atheism are (as one might say) displayed as specimens of atheological arguments, it is clear that the sort of atheist that Aquinas aims to defeat is, in Flew’s terms, a positive atheist. So, if Aquinas adopts a presumption of atheism, it seems that it is positive atheism to which he grants the presumption. Aquinas seems to be imagining an opponent who presents arguments for the proposition that God does not exist. He is not imagining a task of simply presenting theological arguments to a person who just lacks a belief in God.

One who favors Flew’s approach, faces two questions, the first having two parts.

  1. Question 1:
  1. Can there be arguments of any sort in support taking an epistemic stance of negative atheism?
  2. If so, what conclusion would one be trying to support when one does not deny that God exists but simply lacks a belief in God?
  1. If there can be such arguments, what arguments could there be for taking the stance of negative atheism that would not simply be arguments for positive atheism?

It is in the context of these questions (which Flew does not explicitly raise) that Flew’s comments on Aquinas’ objections become particularly significant. He points out that, in Aquinas’ second objection, the specimen atheological argument makes use of a principle of postulational economy and that by introducing this principle into the dialectical exchange, “Aquinas … indicates a reason for adopting” a presumption of atheism.[22] This principle is one that is to be used in comparing the merits of alternative explanations of the same facts. The specimen atheological argument alleges that naturalistic explanation is epistemically superior to theistic explanation because it makes use of fewer principles.

Apparently, Flew is claiming that one can give reasons why negative atheism deserves a presumption in a debate. One reason, Flew seems to be saying, is that, standard principles of evaluating the merits of alternative explanations, e.g., postulational economy, seem to favor not using theism as any part of an explanation of matters of observable matters of fact. One who lacks a belief in God would, of course, not even formulate such explanations. Thus, a kind of justification for the stance of negative atheism is possible even though it is not an argument for the proposition “God does not exist.” So, questions 1(a) and 1(b) seem to be answerable from Flew’s perspective, even though Flew does not seem to perceive that such questions need to be asked, let alone perceive that answers to them need to be provided.

But the answers I have suggested are not entirely satisfactory, and Flew’s comments on Aquinas’ second objection reveal the reason why these answers will not do as they stand. Flew says:

What is perhaps slightly awkward for present purposes is the formulation of the first objection…. It would from my point of view have been better had this first objection referred to possible difficulties and incoherencies in the meaning proposed for the word “God.” Unfortunately, it does not, although Aquinas is elsewhere acutely aware of such problems. The changes required are, though important, not extensive.[23]

The awkwardness that Flew is referring to is related to the question of the sort of atheism to which Aquinas grants a procedural presumption. It seems to be positive atheism because the specimen atheological argument is one having the conclusion “God does not exist.” To take the stance of negative atheism, Flew has stated, is to proceed in the debate as if one simply lacked a belief in God.

Flew imagines a variation on Aquinas’ specimen atheological argument from evil that refers to “possible difficulties and incoherencies in the meaning proposed for the word ‘God.'” The variation would refer to the detection of an “ostensible” incoherence, “not in the proposed concept of God as such, but between that concept and “another element in the theoretical structure in which it is normally involved,” namely an ostensible incoherence between the idea of a creator on whom creatures depend completely, continually and absolutely and the idea of those creatures being sufficiently autonomous for their faults not to be also and indeed primarily the creator’s fault.[24]

This view of God as creator of such creatures may be, as Flew says, essential to theism. It may indeed be a problematical view of the nature of creation. But to speak of an ostensible incoherence between the traditional theistic idea of creator and the traditional idea of his creatures is quite different from speaking of an ostensible incoherence in the meaning proposed for the term “God.” The difficulty Flew is pointing out would better be classified as an ostensible inconsistency between two ideas rather than an ostensible incoherence in one idea.

Flew’s case would be supported much better by leaving Aquinas’ specimen arguments behind and illustrating some convincing examples of ostensibly incoherent concepts. Indeed, he should provide examples of palpably incoherent concepts that may seem initially to be coherent. After all, that is one his main reasons for preferring a stance of negative atheism for debating purposes. One ought not even to concede at the outset that “God” has a coherent meaning. Conceptual incoherence is not to be confused with propositional inconsistency. Propositional inconsistency is possible only when the propositions in question are at least capable of having truth-values. If a term is found to express an incoherent concept then it is not a term that is capable of serving as the subject of a proposition because the result of putting the concept into the empty part of a propositional function would be something incapable of having any truth value. (Example: “Triangular odors are noisy.”)

In his 1990 paper “The Case for God Challenged,” Flew takes Kai Nielsen to task for not taking enough care to “show how some term or expression may be in some way immediately intelligible while remaining nevertheless ultimately incoherent.”[25] Flew even goes so far as to charge Nielsen with having been “reckless” for not having protected himself against the charge that enormous numbers of people have not known what they were talking about when they used the word “God.” Flew then gives what he thinks would be good examples of such terms. He cites Leibniz’s example of such a term: highest conceivable speed. He also cites Leibniz’s argument that, before demonstrating the existence of God, it is first necessary to demonstrate that the concept of God as a logically necessary being is coherent. Flew seems to cite the Catholic doctrine of transubstantiation as “strictly senseless” on the grounds that it would be impossible to identify any substance by anything other than its accidents.[26]

Later, he cites “person without a body” as another example of a candidate for ultimate incoherence only to withdraw this choice on the grounds that tension apparent in applying this term to human beings might be overcome in the special case of God.[27] Then, Flew announces there are difficulties he does not think can be overcome: those of reconciling certain of the defining characteristics of the theistic God with features of the universe in such a way that God can be not only perfectly good but also the creator and sustainer of the universe. Once again, Flew seems to vacillate between what ought to be classified as cases of implicitly contradictory expressions and expressions lacking an ultimately coherent meaning.

Once we reflect on this point, we may arrive at an understanding that is better than what Flew provides of how the presumption of atheism may be justified, where it is a stance of negative atheism that is taken for debating purposes. One may justify taking this stance without thereby producing arguments for positive atheism. The way to do so, I suggest, is by presenting justified doubts about the cognitive status of the statement “God exists.”

Such doubts need not be of the kind that refers to evidence that puts the truth of a proposition in doubt, i.e., reasons sufficient for denying that the proposition is true or for suspending judgment as to its actual truth-value. One may instead have reasons for doubting whether a given statement expresses a proposition at all, without presenting these reasons as premises of an inference to a conclusion that the statement does not express a proposition. Rather, they may be reasons that justify doubting whether the statement expresses a proposition. Another example of such a doubt would be a doubt as to whether the statement has any definite truth-conditions. One may have unresolved and justified doubts, for example, as to what the truth-conditions of “x is a timeless agent” could be, without needing to show, necessarily, that the term is simply a truth-functional contradiction, a category mistake, or other kind of logically deviant expression.

These doubts would be of a sort that make one uncertain as to what the semantical content of a statement might be. In this way, we may led to realize that such doubts explain how a stance of negative atheism need not be a sort of pretense adopted for the purpose of trying to gain the maximum advantage for oneself in the context of a debate. Flew’s defense of negative atheism and the justification of its presumption is in great need of this sort of supplementation. It may seem, in spite of the analogy with the presumption of innocence, that it is mainly an attempt to gain maximum advantage for the atheist in debates with theists.

Fortunately, in some later publications, Flew provides what may serve as examples of arguments that I have suggested may be used to justify doubts about the cognitive status of a proposition, but I will only briefly allude to some of these arguments to illustrate. He argues at length, for example, that the notion of a disembodied or bodiless person is either flatly contradictory or unintelligible. Core issues of personal identity and the possibility of identification play a central role in his argument. Moreover, Flew points to the fact that there is an ancient and ongoing controversy even among the most sophisticated of philosophers and theologians over whether the term “God” has an intelligible meaning. In view of the fact that the God of classical theism is himself a bodiless person, a general doubt as to the intelligibility of classical theism may be and has been produced from other (familiar) arguments over whether a human person can exist in a disembodied state. The existence of such disagreements are themselves grounds for being consciously noncommittal on the intelligibility of the term. On a more popular level (famously discussed by Descartes, Flew notes), it may be observed that people’s religious views seem to vary mostly according to custom and tradition, rather than according to rationally constructed arguments and definitions. There are many cases of people who believe radically different things about the meaning of key religious terms, even though they may be equally rational and highly advanced in their learning, and they are well aware of each other’s beliefs and doctrines.

It may be pointed out in response to this suggestion that these are not arguments justifying doubts about a concept’s intelligibility—they are psychological explanations of how one’s doubts have been caused. It seems to me that they could be both justifying reasons and causally explanatory reasons. (Compare: “I don’t trust him because he always acts so sneaky.”)

According to Flew, the presumption of atheism not only puts the burden of proof on the theist, but also puts on the theist the burden of providing an account of the concept of God that makes its conceptual coherence manifest. According to my suggestions, the policy of this presumption need not be one that is demanded by one who is using a purely self-serving pretense adopted for the special purpose of a debate about the existence of God, or by one who is simply ignorant or naïve. One may take the epistemic stance of negative atheism even if one has no intention to ever take part in a debate about the existence of God. And one may do so even if one is not innocent, ignorant, or naïve about what has been said or written by theists.

Indeed, if I am right, Flew’s own position implies this approach to adopting negative atheism. The very experience of becoming familiar with the claims, definitions, and metaphysical theories of classical theism is more than sufficient to provide one with abundant reasons for doubting whether “God” is semantically coherent or intelligible as it occurs in these contexts. Clearly, Flew himself does have just this type and degree of familiarity with the literature of theism. This sort of experience of becoming familiar with this literature and of talking with theists who may earnestly and sincerely try to explain and justify theism to nontheists may well cause one to have doubts about the intelligibility of the term “God.”
However, as I have suggested, the features of theological discourse that may have this effect also provide one with reasons that justify having such doubts.

These reasons need not be reasons that reflect philosophical naïveté. One need not (and Flew does not) embrace a simplistic sort of verificationist criterion of the meaningfulness of factual statements to have justified doubts about the intelligible content of theological assertions. There is, of course, a vast literature on this subject, too. When one has become familiar with even a small part of this literature, it is quite apparent that there is widespread disagreement about even the most elementary conditions of the meaningfulness of factual assertions. But the existence of this further disagreement should not be of any comfort to theists, or to positive atheists. It is, however, another source of reasons that justify the kinds of doubts about the semantics of “God-talk” that I have suggested makes negative atheism a tenable epistemic stance.

If I am right about this suggestion, then it is altogether fair and reasonable to approach the issue of the existence and activities of God, as Flew did seventy years ago, by beginning with the question to theists of what would have to happen to make anyone logically and rightly conclude that God does not love us or that God does not exist at all. To ask this question is not to employ an empiricist criterion of meaning. It is simply to proceed under a presumption of atheism that entails a burden of proof exclusively for theism. Part of this burden is to explain that the concept of God is coherent or intelligible. Those who have doubts as to whether this concept is coherent ought to have some justification for such doubts, and I have argued, in agreement with Flew, that such doubts are justified.

An important connection between the issue of coherence and falsifiability would seem to be that statements involving incoherent concepts would be unfalsifiable, for the obvious reason that there can be no conditions under which they would be true. Statements that are unfalsifiable may have this property for more than one reason. Conceptual incoherence may be only one reason for a statements’ being unfalsifiable. But it seems to be a useful method to begin by inquiring into the possibility of falsifying a statement and then diagnosing the reason of its being unfalsifiable.

It may be that in the case of theological statements, the reason they are unfalsifiable is undetected conceptual incoherence. One who employs this method and obtains such results will still have some important work to do. One job to explain how enormous numbers of people could use such a concept for a very long time without realizing that it is incoherent, and use it with the attitude that it is coherent. Such a project is beyond the scope of this paper, but it does seem that if the matter of justifying a presumption of atheism comes down to producing explanations of this sort of phenomenon, then significant progress in philosophy will have been made, because these will be explanations that are essentially social and psychological. That is the way that philosophy has progressed for the last 2000 years—by handing over its problems to the sciences that it helps to bring into being.

There are those who classify themselves as atheists on the grounds that all
putative statements containing the term “God” are neither true nor false because they are unintelligible or conceptually incoherent. These philosophers are quick to point out that, for this very reason, they must not be classified as positive atheists since neither the assertions nor denials of unintelligible statements are either true or false. This may be called a noncognitivist version of atheism, and it may also be said to be a version of negative atheism.

If one says that “God” is a contradictory concept, then, whatever one says about God will be vacuously true if it has the form “(x) (x=g→p)” because the antecedent will be necessarily false. Does this result coincide in a nonaccidental way with the unfalsifiability of God-talk? Perhaps only in an oblique way. Everything one can say about God is vacuously true if the concept of God is contradictory. But vacuous truth is not the same as unfalsifiability. Unfalsifiability results from the degree of ad hoc qualification made for God-statements by theists. Yet the two fates of theological assertions are oddly parallel.

Speculation: covert incoherence of concept makes it psychologically possible and easy to introduce ad hoc qualifications to prevent falsification.


[1] Antony Flew, “Theology & Falsification: A Golden Jubilee Celebration.” Philosophy Now, Issue 29 (October/November 2000).

[2] Flew, “Theology and Falsification” in New Essays in Philosophical Theology ed. Antony Flew and Alasdair McIntyre (London, UK: SCM Press, 1955): 96-98, p. 98. (It was first printed in an undergraduate journal called University in 1950 at Oxford University.)

[3] Flew, “The Burden of Proof” in Knowing Religiously ed. Leroy S. Rouner (Notre Dame, IN: University of Notre Dame Press, 1985): 103-115, p. 110.

[4] Flew, “The Burden of Proof,” p. 110.

[5] Flew, “The Burden of Proof,” p. 111.

[6] Michael Martin, Atheism: A Philosophical Justification (Philadelphia, PA: Temple University Press, 1990).

[7] Theodore M. Drange, “Atheism, Agnosticism, Noncognitivism” (1998). The Secular Web. <https://infidels.org/library/modern/theodore_drange/definition.html>.

[8] Flew, God, Freedom, and Immortality: A Critical Analysis (Buffalo, NY: Prometheus, 1984), p. 14.

[9] Flew, God, Freedom, and Immortality, p. 15.

[10] Flew, God, Freedom, and Immortality, p. 15.

[11] Michael Scriven, Primary Philosophy (New York, NY: McGraw-Hill, 1966), p. 100.

[12] Scriven, p. 94.

[13] Scriven, p. 94.

[14] Norwood Russell Hanson, “What I Don’t Believe” in What I Don’t Believe and Other Essays ed. Stephen Toulmin and Harry Woolf (Dordrecht, Holland: D. Reidel Publishing Co., 1971): 309-331, p. 311.

[15] Hanson, p. 315.

[16] Flew, God, Freedom, and Immortality, p. 15.

[17] Flew, God, Freedom, and Immortality, p. 15.

[18] Flew, “The Presumption of Atheism” in A Companion to Philosophy of Religion ed. Charles Taliaferro, Paul Draper, and Philip L. Quinn (Malden, MA: Wiley-Blackwell, 2010): 451-457, pp. 451-452.

[19] Flew, “The Presumption of Atheism,” p. 452.

[20] Flew, “The Presumption of Atheism,” p. 452.

[21] Flew, “The Presumption of Atheism,” pp. 452-454.

[22] Flew, God, Freedom, and Immortality, p. 29.

[23] Flew, God, Freedom, and Immortality, p. 29.

[24] Flew, God, Freedom, and Immortality, pp. 29-30.

[25] Antony Flew, “The Case for God Challenged” in Does God Exist: The Great Debate ed. J. P. Moreland and Kai Nielsen (Nashville, TN: Thomas Nelson, 1990): 162-176, p. 166.

[26] Antony Flew, “The Case for God Challenged,” p. 166.

[27] Flew, “The Presumption of Atheism,” pp. 454.

Copyright ©2020 Charles Echelbarger. The electronic version is copyright ©2020 by Internet Infidels, Inc. with the written permission of Charles Echelbarger. All rights reserved.

all rights reserved