Positive Atheism and the Meaninglessness of Theism (1999)

Michael Martin


In Part I of Atheism: A Philosophical Justification I argued that sentences about the existence of God are factually meaningless, hence, they are neither true nor false. However, in Part II of this book I maintained that if the sentences about God are not factually meaningless, "God exists" is false. Readers of my book--both theists and atheists--have been puzzled. Indeed, some critics have maintained that I have contradicted myself and the basic argument of my book is incoherent.

However, my position is not incoherent although perhaps more clarification is needed. In this paper I defend my position by offering a clarification of what I meant.

The General Structure of The Argument and The Fall-Back Position

Critics have been puzzled about my advocating both the factual meaninglessness of theism and the truth of positive atheism, that is, the view that God does not exist. How can I advocate both? Arguing that both "God exists" and "God does not exist" are factually meaningless seems incompatible with maintaining that "God does not exist" is true.

First of all, in order to understand why there is no incompatibility it is important to understand the general structure of my argument that atheism is justified. The argument can be stated in this way:[1]

1. Either "God exists" is factually meaningless or "God exists" is factually meaningful.

2. If "God exists" is factually meaningless, then it is neither true nor false; hence belief in God is unjustified.

3. If "God exists" is factually meaningful, then it is either true or false; but (given the arguments in Parts I and II of my book) one is justified in disbelieving that God exists.


4. Either belief in God is unjustified or disbelief in God is justified.

Therefore, given the definitions of negative and positive atheism,

5. Either negative atheism or positive atheism is justified.


6. Atheism is justified.

Furthermore, in order to understand that there is no incompatibility between Chapter 2 and Part II it is important to see that my defense of the position in Part II of the book is a fall-back or backup position. Although I argued in Chapter 2 that a case can be made for supposing that theism is factually meaningless, I realized that such a case is not strong and that a backup position is needed. I maintained that the case for the factual meaninglessness of religious statements is not strong because a commonly accepted and fully developed theory of meaning is not yet available. Until such a theory is developed we must rest content with a partial theory of meaning and a partial justification of the meaninglessness of religious discourse. I maintained that this consideration tells strongly against negative atheists relying exclusively on the verifiability theory to support their case. In addition, I argued that since even this partial justification could be undermined, atheists who use the verifiability theory to support their view would be prudent to have a backup position.

I went on to say in Chapter 2 that for these reasons I would not assume in the remainder of my book that the sentences "God exists" and "God does not exist" are factually meaningless. I said that I would act as if talk about God is factually meaningful and develop the case for negative atheism on independent grounds. I asserted that in the remaining chapters of Atheism I would assume that the sentences "God existed" and "God does not exist" are factually meaningful.[2]

Puzzlement From Readers

My position concerning the meaninglessness of theism in Atheism has puzzled some readers. Professor John Keller in a review of Atheism in Faith and Philosophy saw "a perplexing implication" of my position: if I am correct in the first part of the book, the second part of the book is meaningless. Keller complained: "So it is difficult to avoid wondering whether [Martin] believes the argument [from meaninglessness] is correct. If he does, then what does he think he is doing in the remaining 80% of the book; if he does not, why include it."[3] He also claimed that I did not indicate what I took to be the relative strength of my arguments.

Keller's comments betray a misunderstanding concerning the structure of my argument and a failure to appreciate my strategy of using positive atheism as a fall-back position. The only way I can understand his perplexity is that he mistakenly interprets me to be making two categorical claims about the sentence "God does not exist (p)":

(a) p is neither true nor false.

(b) p is true.

If fact, however, I am making one categorical claim--albeit tentatively--and another hypothetical claim:

(a) p is neither true nor false

(c) If p is either true or false, then p is true.

Keller is also mistaken in saying that I did not indicate the relative strength of my arguments. I would have thought it rather clear from what I said on p. 77 of Atheism that I regarded the meaningless argument as subject to more doubts than other arguments and that it was therefore unwise for atheists to rest their case solely on it. [4]

Dr. Keith Burgess-Jackson in an e-mail letter to me dated September 13, 1996 maintained that he had no problem with the argumentative strategy of Atheism but wondered what position I really held in contrast to the position I took for the purposes of the argument. He said that if I really believe that "God exists" is factually meaningless, then I cannot be and should not claim to be a positive atheist.

Unlike Keller, Burgess-Jackson understood what my argumentative strategy was. What perhaps he did fully understand, however, is that positive atheism is a fall-back position and that I did not intend it categorically. My position is something like this. Yes, "God exists" is probably meaningless. But I am not completely sure. If I am wrong, there is still no reason to believe in God (negative atheism) and good reason not to (positive atheism). So my argumentative strategy reflects to some extent my actual belief. I would believe that God does not exist if "God exists" is not meaningless (which it may not be). [5]

Professor Theodore Drange, an atheist, in his excellent unpublished work Nonbelief and Evil writes:

I understand and sympathize with Martin's appeal to a backup (or 'fall-back') position. . . .

However, the use of atheism as a backup position for religious noncognitivism strikes me as less legitimate than any of my maneuvers since noncognitivism declares atheism to be flat-out meaningless. Martin says he agrees with the view that "God does not exist" is meaningless in the sense of being neither true nor false. But then he says in the rest of his book he will disregard that and proceed as though the sentences were meaningful. How can a meaningless sentence be treated as meaningful? That is, how can a sentence which cannot be used in arguments be used in arguments? Some further clarification is needed.

One thing that Martin could have done to make his position clearer would have been to take a more neutral stand on the issue of whether or not the sentence 'God does not exist' is meaningful. He could have said that he 'leaned towards' the view that the sentence is meaningless but had not yet made up his mind on his matter. [6]

There seems to me to be a close similarity between what Drange says I could have done and what I did do. Although I did not say that I leaned toward the view that the sentence "God does not exist" is meaningless, I certainly made plain why I thought the meaninglessness claim was not fully justified and that in order to justify it fully more argument would be needed. There may not be much difference between:

(d) I lean toward position P.

(e) I hold position P but not very strongly.

However, that point aside, it is not clear to me why I could not have a strong belief that sentences about God are meaningless and yet for the sake of the argument suppose they are meaningful. We can believe strongly that a statement is false yet, for the sake of the argument, suppose it is true. It is unclear why a similar hypothetical position cannot be entertained with respect to meaningless sentences. Suppose we read in a faculty handbook that all members of the faculty will wear academic gowns at graduation. One might believe strongly that this is a command and reject it as being inappropriate. However, since commands lack truth value they are meaningless in the sense of lacking truth value. But one might also suppose, for the sake of the argument that what one has read is a prediction and, on this supposition, reject it as being false.[7]

In an e-mail letter dated September 19, 1996 Drange drew a distinction between sentences that lack both truth value and meaning and sentences that just lack truth value. He maintained that my example of "All faculty will wear academic gowns (S1)" taken as a command lacks truth value but not meaning. He admitted that one could treat S1 as true even though one believed that it lacks truth value but he contrasted S1 with sentences that lack both truth value and meaning such as "GJB is a trodeliptic granfy (S2)." Unlike S1, S2 cannot be treated as true.

Drange may be correct that the term "meaningless" is misleading in this context of religious language. Nevertheless, by a meaningless sentence I meant lacking in truth value. Thus on page 42 I defined literal meaning as: "A statement has cognitive or literal meaning if and only if it is either true or false." Moreover, in Atheism I made a distinction between the meaninglessness of sentences like S2 and statements about God. Indeed, I explicitly distinguished the meaninglessness of sentences such as:

(1) God has no body and yet acts in the world.

from the meaningless of sentences such as:

(2) God is gluberfied.

(3) God big impossible.

(4) Goo Foo is gluberfied.

As I said on page 45, "Terms such as 'God, ''has a body', 'acts', 'in the world', unlike 'Goo Foo' and 'is gluberfied', have uses in our language." Moreover, I pointed out that in contrast to (3) (1) has no syntactical irregularities. In addition, "God" has a fixed syntax making possible certain logical inferences from (1). Of course, when I said in Chapter 2 that, although a case can be made that statements of religion are meaningless, in the rest of the book I will suppose they are not meaningless, I meant sentences like (1) and not (2) or (3).

Perhaps a key difference between Drange and me on the meaninglessness of religious sentences is this.[8] He believes that to call sentences like

(5) Close the door!

and (1) meaningless on the grounds that they lack truth value is a mistake whereas I do not. However, I am happy to concede Drange's point for it does not affect my views in any significant way. Religious sentences like (1) would still lack truth value and be noncognitive.

A criticism similar to the ones outlined above was raised by Greg Bahnsen in a taped lecture.[9] Bahnsen claimed that my position that religious statements are meaningless is in conflict with my defence of positive atheism. He said that I cannot "have it both ways": I cannot defend the factual meaninglessness of theism as well as the falsehood of theism. Indeed, Bahnsen argued that if my argument is correct in Chapter 2, I should not have written the rest of the book.

My answer to Bahnsen is similar to the answers given above. Bahnsen does not seem to understand the structure of the argument of my book and failed to understand that positive atheism is a backup position. For this reason he seemed to confuse the logically unacceptable position of holding both (a) and (b) and the logically acceptable position of holding (a) and (c).


I hope that the comments provided here eliminate the perplexity of readers of Atheism: A Philosophical Justification and allay suspicion that there is an inconsistency between my thesis that religious language is meaningless and my defense of positive atheism.


[1] I owe this way of stating my position to Keith Burgess-Jackson who suggested an argument similar to this in an e-mail letter to me on September 13, 1996. In an e-mail letter of 30 July, 1999, Burgess-Jackson suggested that since I mean by "factually meaningful," either true or false that I drop the term "factually meaningless" and reformulate my argument without it . I have not followed his useful suggestion in the text on the grounds that it would depart from historical usage and the terminology used in my book. His reformulation. is:

1. Either religious language has truth value or religious language does not have truth value. (Tr v ~Tr)

2. If religious language has truth value, then positive atheism is justified. (Tr -> Jp)

3. If religious language does not have truth value, then negative atheism is justified. (~Tr -> Jn)


4. Either positive atheism is justified or negative atheism is justified (from 1, 2, and 3, constructive dilemma). (Jp v Jn)

5. If positive atheism is justified, then negative atheism is justified. (Jp -> Jn)


6. Negative atheism is justified (from 4 and 5, double negation, material implication, tautology, constructive dilemma, and tautology). (Jn)

[2] Michael Martin, Atheism: A Philosophical Justification (Philadelphia: Temple University Press, 1990), pp. 77-78. This is not the only fall-back position that I took in Atheism. In Chapter 11 I argued that the lack of evidence and convincing reasons for believing in God demonstrated in the earlier chapters was a good reason to suppose that God does not exist. Since I could be wrong about this I developed other arguments. In Chapter 12, for example, I argued that the concept of God is incoherent; consequently, we have a priori grounds for disbelief. If I was right, no other arguments would be needed to be justified in disbelieving in God. But I might not be. So I developed a posteriori arguments in later chapters.

[3] James A. Keller, Review of Atheism: A Philosophical Justification, Faith and Philosophy, 10, 1993, p. 113.

[4] I made these points to Keller in a letter dated July 27, 1993 in response to his review.

[5] I conveyed this reply to Dr. Burgess-Jackson in an e-mail letter date September 13, 1996.

[6] Theodore Drange, Nonbelief and Evil, unpublished, pp. 34-35.

[7] I expressed these ideas in an e-mail letter to Drange dated September 17, 1996.

[8] In a letter dated September 24, 1996 Drange suggested that a major difference between us is that, unlike me, he sees the ascription of truth to sentences as only derivative from the ascription of truth to propositions. He may be correct. But I doubt that this difference accounts for the difference between us in this case.

[9] See Greg Bahnsen, Michael Martin Under the Microscope, tape 12, (Nash, TX: Covenant Tape Ministry), audiocassette.

"Positive Atheism and the Meaninglessness of Theism" is copyright © 1999 by Michael Martin. All rights reserved.
The electronic version is copyright © 1999 by Internet Infidels with the written permission of Michael Martin. All rights reserved.