Default Agnosticism

Abstract

Agnosticism has always had its fair amount of criticism. Religious believers often described the first agnostics as infidels and it is not uncommon to see them described as somewhat dull fence-sitters. Moreover, the undecided agnostic stance on belief in gods is often compared with being unsure about such obviously false statements as the existence of orbiting teapots, invisible dragons or even Santa Claus. In this paper, I maintain that agnosticism can properly be endorsed as a default stance. More precisely, I use a strategy presented by Alvin Plantinga and argue that it is rationally acceptable to be agnostic about the existence of God. I also anticipate and answer a number of objections. Finally, I offer my conclusion.

1. Introduction

Agnosticism has always had its fair amount of criticism. Thomas H. Huxley, who coined the term, was often described as an infidel (Wace 1889, p. 7) as well as an agnostic coward (Magee 1889, p. 45)1 and agnostics in general have always been described as something like mediocre fence-sitters. For example, in The God Delusion, in a chapter entitled the ‘The Poverty of Agnosticism,’ committed atheist Richard Dawkins appears to be sympathetic to a statement of his childhood preacher. He describes his preacher’s statement as follows:

The robust muscular Christian haranguing us from the pulpit of my old school chapel admitted a sneaking regard for atheists. They at least had the courage of their misguided convictions. What this preacher couldn’t stand was agnostics: namby-pamby, mushy pap, weak tea, weedy, pallid fence-sitters…

(Dawkins 2006, p. 69)

Moreover, while religious believers are sometimes ridiculed as being just as irrational as believers in flying spaghetti monsters (Dawkins 2006, p. 15), believers in invisible dragons (Sagan 1997, p. 171) or Great Pumpkin believers (Martin 1990, p. 272) the undecided agnostic stance on belief in gods is often compared with being undecided about such obviously false and irrational claims as the existence of orbiting teapots in space (Russell [1952] 2000, p. 93), believers in fairies or Santa Claus (Holland 2001, p. 85).2

Despite these accusations of cowardliness and irrationality, I maintain that agnosticism can properly be endorsed as a default stance.3 More precisely, I argue that it is rationally acceptable to be agnostic to the claim that there are no gods and that this does not imply that one needs to commit to agnosticism, or being undecided, with respect to other preternatural beings.

Moreover, my focus is on epistemic rationality rather than on pragmatic or prudential rationality. When evaluating if someone is rational in the former sense the focus is on the goal of believing only true propositions. When evaluating if someone is rational in the latter sense the focus is on whether or not believing certain propositions would somehow be beneficial.

I proceed as follows. In Section 2, I describe my understanding of agnosticism. In Section 3, I present what is generally called the presumption of atheism—the claim that atheism should be presumed in the absence of arguments for the existence of god or gods or when it is considered that the evidence is inconclusive.4 In Section 4, I use a Plantingian strategy to show that agnosticism can properly be regarded as a default stance. In Section 5, I anticipate and answer a number of parody objections. Finally, in Section 6 I offer my conclusion.

2. Agnosticism

Arguably, much of the criticism of agnosticism is due to conceptual disagreements about the term. It therefore seems advisable to start by describing how I use the term and why I use the term as I do.5 In short, I describe what I take to be the simplest meaning of the word.

Huxley often understood his word ‘agnosticism’ to refer not to a position or creed, but to a method. He defined the positive version of the method by stating that “in matters of the intellect, follow your reason as far as it will take you without regard to any other considerations” (Huxley 1892, pp. 281–82) and presented the negative version of it by asserting that one should “not pretend that conclusions are certain which are not demonstrated or demonstrable” (Huxley 1892, p. 282). I must admit that I have a tough time accepting at least the negative version of the method,6 but viewing agnosticism as a method has in any case failed to catch on. Instead, Huxley’s use of the word to denote people who, like himself, declared that they are “hopelessly ignorant” about not only the existence of god or gods, but also about a variety of dogmatic and metaphysical matters has gained more attention (Huxley 1884, pp. 5–6).

Here I restrict agnosticism to refer to an answer or actually a failure to answer the ontological question of whether or not there exists at least one god.7 An agnostic, in other words, is someone who after considering the matter is undecided and withholds judgment on the proposition that ‘there is a least one god’ (Oppy 2018, p. 4). More elaborately, agnosticism is defined in terms of being an agnostic and as such of being in a psychological state towards a proposition.8 By contrast, a theist affirms the proposition and an atheist denies it (Baggini 2003, p. 3).9 We should perhaps emphasise the words “at least” one god. A person could deny the existence of an array of gods, but as long as he or she also is agnostic towards at least one of them and does not affirm the existence of any of them he or she would qualify as an agnostic.

Nevertheless, instead of distinguishing between atheism and agnosticism and understanding the terms as I do here a distinction between ‘positive atheism’ and ‘negative atheism’ is sometimes introduced. The former refers to the denial of the existence of all gods while the latter refers to being in a state of not having a belief about gods at all (Martin 1990).10 Since the introduction of such a distinction does not allow for any difference between atheism (as in negative atheism) and agnosticism, some people instead define agnosticism by appealing to the etymological meaning of the word. In its ancient Greek meaning, an agnostic is someone who is without γνῶσις (gnōsis), or knowledge, about gods. The definition suggests that there can be both theists and atheists who also endorse agnosticism, for the simple reason that there might be (or almost certainly are) believers and disbelievers who would not claim that they know that at least one god exists or no gods exists; they simply claim to have a specific belief about the matter.11

Now the distinction between positive and negative atheism introduces an overly broad atheistic category. Indeed, I doubt that many of those who would qualify as atheists (that is, negative atheists) would be comfortable with such a description. Moreover, it is not at all obvious that the position of not having a belief about the existence of gods should be sorted under the category of atheism. Scott Yoder, for example, distinguishes between ‘agnosticism of not knowing’ and ‘agnosticism of not believing’ and thus sorts the position of withholding judgment and not having a belief about the claim that there are no gods under the category of agnosticism (Yoder 2013, p. 141).

However, my own reason for defining agnosticism simpliciter, as withholding belief with respect to the proposition that ‘there exists at least one god’, is pragmatic. That is to say, it makes pragmatic sense to separate the ontological question about whether or not there are no gods from, for example, the epistemic question about whether or not one knows, or indeed can know, that there are no gods. It should also be recalled that in this paper my focus is on epistemic rationality and not on knowledge. Arguably, the work of the philosopher would generally be easier if epistemic qualifiers like ‘knowledge,’ ‘rationality,’ ‘justification,’ ‘reasonability’ and ‘warrant’ are only introduced when needed.12

Nevertheless, it might be suggested that the pernicious problem for the view of agnosticism used here is not whether or not both ontological and epistemic terms should be included in the definition of agnosticism; rather, the problem relates to how we define a god. It is indeed difficult to know what properties an entity would need to have to properly be called god. Here I will rely on our prior understanding of the term, but add that I mainly focus on what I call source theism and perfect being theism. The god of source theism is a being who fulfils the roles of being the ultimate cause of the universe as well as the ultimate source of love and morality.13 The god of perfect being theism is a being who is all-powerful, all-knowing, perfectly good and the creator of the world. The latter god is a more specific god than the former but I strongly believe that most people would agree that both beings are serious candidates for being gods.14 However, it is actually atheism rather than agnosticism that comes up against a problem here. If one denies the existence of all gods one needs to know what qualifies as a god. Without this knowledge, the agnostic attitude seems more plausible.

Moreover, even if an atheist is able to define a sufficient set of conditions for being a god another problem arises. Atheism, as it is most often understood, is rather sweeping in it its claim. The atheist is not agnostic, but denies the existence of all potential gods.15 The atheistic claim thus amounts to the denial of a disjunction. Specifically, the claim is that it is not the case that god₁ or god₂ or god₃ or god₄ or god_n exists, and such an extensive claim is rather difficult to justify (Diller 2016, p. 8), unless atheism for some reason should be presumed, even without evidential support.

3. The Presumption of Atheism

The claim that atheism should be presumed before the evidence comes in or when the evidence is ambiguous is actually fairly common. According to Robin Le Poidevin, the “almost orthodox view” is the following:

Atheism doesn’t require a defence. Rather, it is up to theists to convince us that there is a God. Unless they can do so, we can remain comfortable in our disbelief. Only if they produce a really compelling argument in favour are we obliged to stir ourselves and show just where the argument fails. If there’s room for doubt […] then the rational thing is to be an atheist.

(Le Poidevin 2010, p. 56)

This is what is often labelled the presumption of atheism and the presumption makes agnosticism redundant. That’s why Le Poidevin is correct (I think) in stating that any practicing agnostic needs to have an answer to it (Le Poidevin 2010, p. 56).

Still, the presumption of atheism is not endorsed entirely without reason. The standard reason given is that if one endorses agnosticism about gods when one does not have arguments or is unable to judge the (relevant) evidence then one is also committed to being agnostic about many other things that we generally would deny the existence of. By avoiding such irrational commitments, atheism would therefore be the rational presumption. This reason can at least be traced to the atheist Bertrand Russell. In Is There a God?, he writes:

If I were to suggest that between the Earth and Mars there is a china teapot revolving about the sun in an elliptical orbit, no-one would be able to disprove my assertion provided I were careful to add that the teapot is too small to be revealed even by our most powerful telescopes. But if I were to go on to say that, since my assertion cannot be disproved, it would be intolerable presumption on the part of human reason to doubt it, I should rightly be thought to be talking nonsense.

(Russell [1952] 2000, p. 93)

Similarly, Norwood Hanson espouses the strategy that Le Poidevin calls the “almost orthodox view” but uses the examples of green goblins and the Loch Ness Monster:

I will urge that God does not exist precisely because the reasons theists advance for supporting that he does exist are all poor reasons […] for that conclusion. This is just the way in which one supports claims like ‘It is not the case that green goblins exists’ and ‘It is not the case that the Loch Ness Monster exists. One simply evaluates the reasons certain believers offer in support of the claim that such entities do exist, and then show them to be poor reasons.

(Hanson 1971, p. 313)

Other entertaining examples found in the literature are Santa Claus (Rosenberg 2011, p. xii; Scriven 1966, p. 103) and flying spaghetti monsters (Dawkins 2006, p. 15), but let me stress again that the point here is that atheism is the default stance and that even agnosticism about gods would commit one to agnosticism about other obviously irrational beliefs. The self-proclaimed atheist Michael Scriven even thinks that we have an obligation to adopt atheism in the absence of evidence for god (here he thinks of something like the god of perfect being theism). He writes:

[W]e need not have proof that God does not exist in order to justify atheism. Atheism is obligatory in the absence of any evidence for God’s existence […] The proper alternative, where there is no evidence, is not mere suspension of belief, e.g., about Santa Claus; it is disbelief.

(Scriven 1966, p. 103)

At first glance Scriven’s claim that we are obliged to adopt atheism in the absence of compelling evidence seems to be taken from thin air. As Alvin Plantinga has observed, there certainly is a lack of parity in that atheism does not have to meet the same evidential standard as theism (Plantinga 1983, pp. 27–8). Instead, Russell, Hanson and Scriven seem to rely on their comparisons between believing in a god and believing in orbiting teapots, green goblins, the Loch Ness Monster and Santa Claus.16 However, there are obviously quite striking differences between teapots, whether orbiting or not, green goblins, Loch Ness monsters, Santa Claus and (say) the god of perfect being theism. If we also throw into the mix fairies, Halloween visiting Great Pumpkins and invisible gardeners—all examples from the literature on the subject—an underlying principle, rather than a comparison, to help us affirm or deny the existence of these beings would be appreciated. To find such an underlying principle we need to move beyond Russell, Hanson and Scriven.

Robert McLaughlin, for example, adduces the well-known principle of Occam’s Razor. Employing (yet another) amusing example, he writes:

‘A three-headed hippogriff is alive and well right now on the fifth planet of the Vega system.’ If I follow this claim with the frank admission that I had no grounds at all for it, you might be inclined to reject it out of hand, despite your presumed lack of any grounds for its negation. To do this would be to invoke a version of Occam’s Razor.

(McLaughlin 1984, p. 198)

McLaughlin does not really offer a clear definition of Occam’s Razor but seems to endorse a rather strong version of it (McLaughlin 1984, p. 202). Such a strong version would stat that one should reject (or assert the non-existence of) unnecessary entities (Brinton 1989, p. 6). McLaughlin’s purpose is to argue for a wide applicability of Occam’s Razor17 but he states that his reasons for adopting it are pragmatic and omits to expand on what these pragmatic reasons might be. However, even if there are pragmatic reasons for denying existence claims without evidence pragmatic reasons are generally not regarded as truth-indicative. Typically, they are regarded as indicative of other values or benefits besides truth. So, if it is the truth that we are searching for it seems odd to claim that for pragmatic reasons we have an obligation to deny the existence of things, even when we have no reasons supporting their non-existence.

Unfortunately, an agnostic version of Occam’s Razer does not do any better with regard to the issue in question. Such a version would only state that one should not assert the existence of unnecessary entities, not that one should deny that they exist (Brinton 1989, p. 6) and even though it might suggest agnosticism with respect to the god of perfect being theism it also suggest that we should be agnostic about for example Leprechauns and McLaughlin’s hippogriff.

Another principle is presented by Thomas Morris, who adds a clause—let us call it the Morris Clause—to the view stated by Hanson and Scriven. Hanson and Scriven, it will be recalled, subscribe to ‘the orthodox view,’ according to which one ought to deny the existence of things, including gods and other preternatural entities, unless one has evidence supporting their existence. Morris adds the proviso that we only should accept such a denial if we believe ourselves to be in a good epistemic position relative to the existence claim in question. He provides the following definition of what it means to be in a good epistemic position:

[B]eing in a good epistemic position relative to a proposition is being in such a situation that should there be any positive epistemic considerations for the truth of the proposition, one would have them or most likely have them.

(Morris 1985, p. 222)

Morris goes on to argue that since the proposition ‘god exists’ (he is here thinking of the god of perfect being theism) is a metaphysical proposition we are not in a good epistemic position to make any judgment about the truth of the proposition. In other words, we cannot reasonably believe that we are able to make the relevant truth considerations (Morris 1985, p. 223).

The Morris Clause has some intuitive force but it also seems to have some of the consequences we wanted to avoid. P. J. McGrath writes that Morris’s position:

Implies that we should be agnostic about almost all preternatural beings. For example, I know of no good reason for thinking that the gods of ancient Greece and Rome exist and I am not in a good position for assessing their existence, since there is no guarantee that if they did exist, evidence to support that fact would be available to me. The same may be said of the existence of tribal deities, elves, hobgoblins, leprechauns or indeed of the existence of any preternatural being one may care to invent.

(McGrath 1987, pp. 54–55)

To be fair, Morris does not think that we need a “guarantee” that if the ancient gods exist, for example, we would have access to evidence supporting their existence. As we have seen, Morris puts it in terms of “reasonable belief” about being in a good epistemic position and how “likely” it is that one has access to the relevant truth considerations (see the quote from Morris above). Recognising this is important because one might be inclined to think that if the ancient gods existed, we would most likely have access to evidence supporting their existence, even without a guarantee of their existence. Contrary to what Morris himself argues, this recognition might also help us when denying also the god of perfect being theism, since it seems reasonable to believe that if this god exists, he would most likely make himself known to us.

Nevertheless, even if the Morris Clause might help us to deny the existence of some preternatural beings, like the ancient Greek gods and (perhaps) the god of perfect being theism, other beings whose existence we would like to deny seem to fall into a category which the clause suggests we should be agnostic about. Take (again) McLaughlin’s three-headed hippogriff on the fifth planet of the Vega system, for example. Are we really in a position to deny the existence of this specific (and peculiar) hippogriff? Certainly, it is not reasonable to say that if McLaughlin’s three-headed hippogriff existed somewhere in the Vega system we would be likely to have evidence suggesting that it does.

Recently, it has been suggested to me that one should actually bite the bullet here. More precisely the suggestion is that if we cannot make a judgment about whether or not we have or are likely to have access to the relevant truth considerations about leprechauns, green goblins, hippogriffs, etc. we should accept and commit to agnosticism about these supernatural beings. The idea being that we can be agnostic about, say, the god of perfect being theism while also simply accepting agnosticism with respect to strange and fairy tale-like beings. I find it difficult to commit to such agnosticism; if a principle suggests that we must accept such agnosticism something must be wrong with the principle, or perhaps Hanson and Scriven are actually correct in their claim that we should accept atheism and consequently deny the existence of all the other preternatural beings.

More recently some philosophers of religion have taken a different approach and started to evaluate the issue of which position—atheism, agnosticism or indeed theism—should be presumed by using the terms prior or intrinsic probability. Robin Le Poidevin, for example, defines prior probability as (roughly) the probability that a proposition has “before the evidence starts to come in” (Le Poidevin 2010, p. 49) and proposes that we measure what probability value an existence claim has by determining how much the claim (or proposition) rules out. The more it rules out, the lower the prior probability, and vice versa. In other words, the specificity of the claim determines its probability value. An unspecific existence claim, for example, has more ways to be true than a very specific one (Le Poidevin 2010, pp. 49–50).

Le Poidevin goes on to define a god, let us call it the god of source theism, as a being who fulfils the roles of being the ultimate cause of the universe as well as the ultimate source of love and morality. He then points out that the difference between source theism and atheism is simply that a source theist thinks that there is a being that fulfils these roles while an atheist denies that there is a being that fulfils them. Such theists and atheists thus make similar claims that are not very specific, yet at the same time equally specific and therefore equally probable. He adds, however, that when we are more specific about how god fulfils the above roles the prior probability decreases, but that the same goes for atheism. That is to say, when atheists give other, more detailed explanations of what fulfils the role of being the ultimate cause of the universe as well as the source of love and morality the prior probability of this more specific atheism (or perhaps more appropriately, ‘naturalism’) also decreases (Le Poidevin 2010, pp. 52–53).

A consequence of Le Poidevin’s view seems to be that the god of a more specific theism, like the god of perfect being theism, has a low probability value and that therefore atheism (albeit a non-specific version of atheism) should be presumed. Moreover, the existence claims about other more specifically characterised gods, as well as other preternatural beings like spaghetti monsters, Santa Claus, leprechauns, Loch Ness monsters and happy hippogriffs in the Vega system, also have a low probability value.

Paul Draper points out that there is good reason for thinking that prior or intrinsic probability does not only depend on the specificity of the existence claim made, even if it is uncontroversial that this is an important factor (Draper 2017, p. 9). Instead, besides specificity, Draper adds factors like number of ontological commitments, elegance, coherence and intelligibility, all of which increase the probability value of a proposition, and then formulates his own argument. He does so by first comparing what he calls source idealism with source physicalism. According to the former, the mental world existed before the physical world; indeed, it is the source of, or caused, the physical world. This view is consistent with ontological physicalism, but physical entities are causally dependent on mental entities. The latter view is the precise opposite of source idealism. This being the case, source idealism and source physicalism are equally probable. They are symmetrical views and equally specific. They also make the same number of ontological commitments, are coherent and equally intelligible. Draper then goes on to argue that since theism, like that of perfect being theism, is a more specific version of source idealism its probability is much lower. Moreover, given that source idealism and source physicalism are equally probable, perfect being theism is quite improbable. Atheism should thus be presumed. The specificity factor also seems to lower the probability of the existence of other preternatural beings like fairies and leprechauns.

In my view, at least Draper’s approach and argument are convincing. He establishes (I think) agnosticism with respect to a non-specific conception of god (call it the god of source idealism). However, there is an altogether different approach that we can take in order to show that one can rationally accept agnosticism as a default stance with respect to a specific version of theism, such as perfect being theism, while at the same time not committing to agnosticism with respect to intergalactic hippogriffs, leprechauns or the man in the moon (to add yet another amusing example).

4. Agnosticism as a Default Stance

One way to show that agnosticism with respect to the more specific god of perfect being theism can be a regarded as a default stance is to change the approach altogether. Until now we have, in Rodrick Chisholm’s terms, used the strategy of ‘methodism’ rather than the strategy of ‘particularism’ (Chisholm [1966] 1989, p. 7). In other words, we have started by trying to formulate a principle of rationality (or conditions or factors determining prior probability values) and then applied these principles to the existence claim ‘at least one god exists.’ Whenever we have found that a principle leads to absurd consequences, we have tried another principle or added a clause of some kind (I am of course thinking of the Morris Clause here). However, the strategy of particularism seems just as adequate as that of methodism. Why not start with propositions one can accept as rational or probably true and then use these examples to formulate a principle of rational acceptability, or perhaps prior probability? Arguably the most well-known philosopher of religion who used this latter strategy of particularism is Plantinga. In his now classic essay ‘Reason and belief in God’ Plantinga argues that the appropriate way to find a criterion for what he calls ‘properly basic beliefs,’ i.e., beliefs that are not evidentially supported, based on or inferred from others in one’s noetic structure of beliefs, is not deductive but broadly inductive. He writes:

We must assemble examples of beliefs and conditions such that the former are obviously properly basic in the latter, and examples of beliefs and conditions such that the former are obviously not properly basic in the latter. We must then frame hypotheses as to the necessary and sufficient conditions of proper basicality and test these hypotheses by reference to those examples.

(Plantinga 1983, p. 76)

According to Plantinga, a criterion for proper basicality, or in our case a principle for determining one’s default stance regarding the proposition that the god of perfect being theism exists, cannot be obtained ex cathedra. It must be argued, and tested, by appealing to a relevant set of examples.

Some examples are quite clear. Plantinga suggests, for example, that ‘being appeared to redly’ might justify the belief that ‘one is appeared to something red.’ If so, ‘being appeared to redly’ is both a necessary and sufficient condition justifying the belief in question. To put it differently, when ‘being appeared to redly’ it is not only rational and permissible to believe that ‘one is appeared to something red.’ The belief in question would also be considered as appropriately grounded in how reality really is (Plantinga 1983, p. 77).18 With respect to other beliefs, it might be more complicated to find the right criterion or principle.

Nevertheless, Plantinga goes on to argue that just as ‘being appeared to redly’ confers justification to the proposition that ‘there is something red in front of me’ other conditions, or being in other circumstances, might bring forth the belief that ‘god exists’ (here he is thinking of a specific conception of god, something in the line of perfect being theism), or beliefs which entail that ‘god exists.’ However, there is no particular reason to think that everyone would agree with the examples given. Christian theists, Plantinga claims, might have their own criteria, inductively established by their own set of beliefs and conditions, that bring forth those beliefs. Plantinga mentions a few such conditions and beliefs. ‘Reading the Bible,’ for example, might give rise to the deep sense that ‘god is speaking to me,’ or ‘looking at the beautiful mountains’ might give rise to the belief that ‘god has created all this.’ Moreover, these beliefs self-evidently imply that ‘god exists.’ These conditions and specifically Christian beliefs are also analogous to the condition of ‘being appeared to redly’ and ‘there is something red in front of me,’ and many Christians, Plantinga writes, would think that these circumstances in which Christians might find themselves also confer justification for their beliefs. They are justified because they are appropriately grounded in how reality really is. Again, many would disagree, but why should that matter?

To my mind, Plantinga has shown that not only Christian but also Islamic or, say, Hindu communities would, at the very least, not be doing anything wrong in establishing their own set of criteria or principles for their religious beliefs. Although not everyone can be de facto right, it is entirely rational to take this initial stance; they would not be failing to meet any moral nor epistemic obligation.

Still, this little foray into Plantinga’s views on Christian perfect being theism as being properly basic and rational without evidential support has not helped us to see why agnosticism with respect to the god of perfect being theism can be a default stance. Let us therefore first look at atheism, before turning our attention to such agnosticism.

What kind of circumstances could give rise to beliefs that (say) self-evidently entail, or at least suggest, atheism with respect to the god of perfect being theism? One obvious example, to my mind, is circumstances of horrific and seemingly gratuitous suffering. I think it is safe to say that many atheists would claim that just as ‘being appeared to redly’ grounds the proposition ‘there is something red in front of me’ being confronted with cases of ‘seemingly pointless suffering’ gives rise to the belief that ‘that particular case of suffering is pointless,’ and such a belief self-evidently implies that the god of perfect being theism does not exist, since a good, almighty and all-knowing god would not allow suffering that does not serve any purpose or does not lead to a greater good.19 Not all theists would agree, but why should that matter? Atheists might have their own criteria that have been inductively established by their own set of beliefs and conditions that bring forth those beliefs. They may also believe that in such conditions their atheistic belief would be appropriately grounded in reality, perhaps reliably produced by their cognitive faculties or their capacity of intuition.

Figuring out or suggesting circumstances that might give rise to an agnostic attitude, or more specifically, that might make a person not formulate a belief about the specific god of perfect being theism at all is more difficult. The best thing would of course be to let the agnostics themselves use Plantinga’s inductive procedure, but here is a suggestion that I think is lurking in the bushes. The suggestion is that one might have both an inclination naturally to form a belief about the god of perfect being theism and the belief that that god does not exist. More specifically, one might have a sense of awe and wonder at the world or the starry heavens above and therefore be inclined to form a belief about god. At the same time, one might recognise the existence of seemingly gratuitous suffering in the world and therefore be inclined to form a belief about god’s non-existence.20 Where both these inclinations are present agnosticism with respect to the god of perfect being theism seems to be a natural position to hold, as it may not be possible to determine which inclination is reliably produced and appropriately grounded in reality.

It is important to note here that all these situations that may or may not give rise to belief about the god of perfect being theism are not being construed as arguments or propositional evidence. They are simply situations that naturally give rise to belief or disbelief about the existence of a specific god, and believers or disbelievers in this god may think that the belief or disbelief that is formed in these situations is just as true as a perceptual belief formed when, say, ‘being appeared to redly.’

Yet, there is an interesting sense in which agnosticism with respect to the god of perfect being theism has an advantage compared with its theistic and atheistic counterparts. It could be argued that perceptual beliefs like ‘there is something red in front of me’ are universal. However, when forming beliefs or disbelief about the god of perfect being theism, in the Plantingian way presented above, one must at least have some kind of story that explains why there are so many people who are not inclined to form the same belief or disbelief. Plantinga, for example, appeals to Calvin and thinks that everyone might have a special capacity to sense the divine reality and to form a belief about the god of perfect being theism. Sometimes, however, that capacity is not functioning properly because of sin and our ability to resist such belief (Plantinga 2000, p. 205). An atheist, on the other hand, might explain why there are people who naturally believe in some kind of god by appealing to self-deception and Freudian theories about the infantile need for a god who serves as a father figure in life even when this god does not exist. However, agnostics, who have an inclination to believe as well as disbelieve, do not have to come up with a story explaining why other people believe or disbelieve in the god of perfect being theism. Rather it is precisely because they do not have such a story explaining why some believe and some disbelieve in the specific god of perfect being theism that agnosticism is a rational position especially when there is an inclination both to believe and disbelieve.

Yet there is a problem here. One might wonder how I can argue that agnosticism with respect to the specific god of perfect being theism can be rational and at the same time agree with Draper’s argument that the (intrinsic) probability value of the existence claim that the god of perfect theism exists is low and that atheism should therefore be the default option. Indeed, using the Methodist and particularist strategies, we seem to end up with contradictory views about whether or not atheism or agnosticism, with respect to the god of perfect being theism, should be the default stance. The answer to the problem is that it is precisely when we are put in two conflicting situations where we are naturally inclined both to form a belief about the god of perfect theism and a disbelief in that god (as well as not having a story telling us that one of the two is not appropriately grounded in reality) that agnosticism with respect to the god in question is the rational stance. However, if we do not have any natural inclinations on the matter whatsoever then the issue of prior probability presented by Draper comes into play. Until the evidence comes in, and if we are not in a situation where we are naturally inclined to form an agnostic attitude, then Draper’s argument should make us conclude that the prior probability of the god of perfect being theism is low.

It seems, therefore, that when one uses the strategy of particularism, or at least a Plantingian version of the strategy, one can arrive at the conclusion that agnosticism even with regard to a specific conception of god can be rationally acceptable. Indeed, it would seem odd to claim that agnostics who find themselves in conflicting situations that naturally bring forth their agnostic attitude, and who cannot form a judgment as to which belief or disbelief is appropriately grounded in reality actually have a, say, Scrivenesque obligation to endorse atheism. However, there are a number of objections.

5. Objections

Before presenting and answering two objections to thinking of agnosticism as a default and rationally acceptable stance, I want to stress that even if the agnostic state can be rationally acceptable and permissible, evidence might very well change things. To cling to one’s original stance no matter what evidence is presented would certainly be irrational.

Having said that, I want to focus on the problem that has been the central problem in this paper, namely that agnostics about the claim that there is at least on god is committed to other clearly irrational agnostic attitudes towards the existence of invisible gardeners, hobgoblins, etc. One could still object that if an agnostic is to be considered rational in the Plantingian sense presented above then the same agnostic is still committed to agnosticism with respect to other clearly false beliefs. Plantinga himself has named such objection to the proper basicality of theism the Great Pumpkin Objection. Here is what he writes:

If belief in God can be properly basic, why cannot just any belief be properly basic? Can we not say the same for any bizarre aberration we can think of? What about voodoo and astrology? What about the belief that the Great Pumpkin returns every Halloween? Could I properly take that as basic?

(Plantinga 1983, p. 74)

To make the objection an objection against agnosticism as rationally acceptable, the last two questions in the quote can be reformulated as follows: What about the belief that the Great Pumpkin returns every Halloween? Must I rationally endorse agnosticism with respect to that belief?

Well, the best answer in my opinion is to argue that an agnostic is not committed to agnosticism about Great Pumpkins, fairies, green goblins and so on because the agnostic may at least believe that all of these beliefs are not appropriately grounded in reality while at the same time being unsure whether or not the belief that a specific god exists is grounded in reality. Or, to use Plantinga’s own broader epistemology of proper functionalism, the agnostic might think that beliefs about Great Pumpkins, spaghetti monsters and hippogriffs—to take some other examples—are not produced by properly functioning and truth-conducive human faculties while nonetheless being uncertain and agnostic about whether or not belief in the god of perfect being theism would be produced by such properly functioning faculties.21

There is, however, another objection labelled the Son of the Great Pumpkin Objection (Martin 1990, p. 272). A version of this objection is presented by Keith DeRose. He argues that the real problem for Plantinga is that hypothetical Great Pumpkin believers, or Pumpkinites, can use the same particularistic strategy as Plantinga when defending the rationality of their beliefs. He writes:

The question is whether Plantinga’s use of strategy is any more successful, or cogent, perhaps, than is the Pumpkinite’s. To the objector, it seems that, and there seems to be nothing to block the conclusion that, to use Plantinga’s own phrase, Plantinga’s defense is “no better than” the Pumpkinite’s defense.

(DeRose 1999, p. 9)

Now, in order to recast the objection, we must think of someone who is agnostic about whether or not the Pumpkin comes every Halloween or, if we want to, about someone who is agnostic about invisible dragons, Santa Claus or what have you. Could such an agnostic really use the above Plantinga-style strategy to claim that they are rational in their specific agnosticism? I think he or she can and I do not regard it as a specifically difficult bullet to bite. It is still the case that the agnostic about the god of perfect being theism could quite properly claim that Pumpkin agnostics, Santa agnostics and hippogriff agnostics are agnostics about existence claims that are not or cannot be appropriately grounded in reality. In other words, while they are committed to other people using the Plantinga strategy to defend other beliefs or types of agnosticism they are not committed to adopt other beliefs or other types of agnosticism themselves.

6. Conclusions

In this paper, for pragmatic reasons, I defined an agnostic as someone who after considering the matter withholds judgment on the ontological question of whether or not there exists at least one god. More elaborately, I argued that as long as one is undecided with respect to at least one particular god and does not affirm the existence of any god one qualifies as an agnostic.

I then evaluated a number of principles in order to see whether atheism or agnosticism should be adopted as a default stance. Referring to Le Poidevin and Draper, I argued that we should at least believe that it is as likely as not that the god of source theism exists but that before the evidence comes in one should deny the existence of the god of perfect being theism. One should deny the existence of this god because the factors of specificity, number of ontological commitments, elegance, coherence and intelligibility decreased the prior probability of such a god.

However, I then invoked a strategy devised by Alvin Plantinga to argue that there is a sense in which one can be rational when being agnostic with respect to the god of perfect being theism even before the evidence comes in. More specifically, I argued that one can have a sense of awe and wonder at the world or the starry heavens above and therefore be inclined to form a belief about this god. At the same time, one might recognise the existence of seemingly gratuitous suffering in the world and therefore be inclined to form a belief about god’s non-existence. When both these inclinations are present agnosticism with respect to the god of perfect being theism seems a natural position to hold. I also made it clear that being in situations where one is naturally inclined to form beliefs and disbeliefs about god is not the same as alluding to propositional evidence. Instead, a proper comparison is beliefs formed from perception. We do not form perceptual beliefs by evaluating evidence. Rather, we naturally form such beliefs in situations where we are appeared to (say), something redly.

Finally, I answered two possible objections. According to the first objection, if one adopts the Plantingian strategy in order to show that agnosticism can be rationally acceptable one also commits to agnosticism with respect to ghosts, intergalactic hippogriffs or the Great Pumpkin that comes every Halloween, to name a few examples. I answered by saying that an agnostic might very well be in a situation where he or she is inclined to form beliefs or disbeliefs about the god of perfect being theism and therefore end up an agnostic, but there is no reason to think that the same agnostic would be naturally inclined to form an agnostic attitude to Great Pumpkins, hobgoblins and suchlike beings and think that such beliefs are appropriately grounded in reality.

According to the second objection, a person can at least use the same Plantingian strategy that I have used, but do so in order to argue that Great Pumpkin belief can be rationally acceptable. I answered by saying that Pumpkin agnostics, Santa agnostics and hippogriff agnostics and so on can indeed use the Plantingian strategy but this does not commit others to adopt such agnosticism. An agnostic about the god of perfect being theism can still believe that beliefs about most or all other preternatural beings are not appropriately grounded in reality while welcoming others to use the Plantingian strategy as they please.

Thus, as long as one does not believe in any god one can be rational in adopting agnosticism with respect to the god of perfect being theism and is not committed to agnosticism with respect to clearly irrational beliefs about the Loch Ness Monster, fairies, invisible dragons, hobgoblins, hippogriffs in space, the man in the moon, ghosts and what have you.