The New Atheism
Where’s The Evidence?
Michael Antony argues that the New Atheists miss the mark.
“A wise man,” wrote Hume, “proportions his belief to the evidence.” This is a formulation of evidentialism – the view that a belief is rational or justified if and only if it is supported by one’s evidence. A more generalized version of evidentialism covers beliefs with various degrees of confidence, as well as other ‘doxastic attitudes’ such as disbelief, doubt and suspension of judgment (doxa is Greek for belief or opinion). It states that the rational or justified attitude to adopt with respect to a claim or proposition is the attitude that fits one’s evidence. Although evidentialism is much harder to clarify and defend than it might seem, there is no denying its prima facie reasonableness.
Evidentialism plays a key role in attacks against religious belief by the New Atheists, as it did for Hume. Belief in the existence of God or other divine realities is criticized on the ground that there is no good evidence for it. Echoing Carl Sagan and Laplace before him, we are told that “extraordinary claims require extraordinary evidence,” and we are assured that there is nothing of the sort when it comes to the divine. The upshot is that religious belief must be judged irrational, epistemically unjustified, or intellectually illegitimate, and it should be rejected. As Christopher Hitchens is fond of saying, “what can be asserted without evidence can also be dismissed without evidence.”
But what of the New Atheists’ atheism – their belief that there is no god or other divine reality? According to evidentialism, that belief (with whatever degree of confidence it is held) also requires evidence in order to be rational. However, the New Atheists tend not to worry much about providing evidence. Although they sometimes offer arguments – ‘the problem of evil’, Dawkins’ ‘Ultimate Boeing 747 gambit’ in The God Delusion, and a few others – overall, those arguments play a minor role in their attacks. Far more central is their repeated insistence that because religious belief lacks evidence, it is irrational and so should be abandoned.
The question I wish to ask is this: How can the New Atheists employ evidentialist principles to argue that religious belief is irrational if they are unwilling to apply those same principles to atheism? If the New Atheists’ atheism is not evidence-based, as Hitchens implies in the above quotation, doesn’t evidentialism entail that atheism is itself irrational or epistemically unjustified? The answer is ‘Yes’; at least if evidentialism is interpreted in the standard way. So it appears that the New Atheists need some fix for evidentialism – a kind of ‘theoretical plug-in’ – which legitimizes their atheism in the absence of evidence. They also seem to be aware of this, since they offer several reasons why atheism requires no evidential support. I will discuss five of the most commonly-offered reasons, and argue that none of them succeed. At the end I will gesture toward what I believe is the right way to view matters.
1. Atheism Isn’t A Belief
It is often said by atheists that atheism is not a positive position at all – a belief or worldview – but merely a disbelief in theism, a refusal to accept what the theist believes, and as such, there is no belief or position for there to be evidence for. Evidence is not needed for ‘non-positions’.
While the word ‘atheism’ has been used in something like this sense (see for example Antony Flew’s article ‘The Presumption of Atheism’), it is a highly non-standard use. So understood, atheism would include agnosticism, since agnostics are also not theists. However, on the common understanding of atheism – no divine reality of any kind exists – atheism and agnosticism are mutually exclusive. Some insist that this non-standard sense of ‘atheism’ is the only possible sense, because a-theism means without theism. But if that were a good argument, the Space Shuttle would be an automobile, since it moves on its own (mobile=move, auto=by itself). Ditto for dogs and cats.
Yet none of that really matters, for even the non-standard sense of ‘atheism’ does nothing to neutralize evidentialism’s demand for evidence. As we saw, evidentialism applies to all ‘doxastic’ attitudes toward a proposition P: believing P, believing not-P, suspending judgment about P, etc. Therefore evidentialism says, with respect to the proposition God exists, that any attitude toward it will be rational or justified if and only if it fits one’s evidence. Now it is true that if one had no position whatever regarding the proposition God exists (perhaps because one has never entertained the thought), no evidence would be required for that non-position. But the New Atheists all believe that (probably) no God or other divine reality exists. And that belief must be evidence-based if it is to be rationally held, according to evidentialism. So insisting that atheism isn’t a belief doesn’t help.
In what follows I will use ‘atheism’ in its standard sense.
2. You Can’t Prove A Negative
Another common claim of the New Atheists is that you ‘can’t prove a negative’ – where what is typically meant is a negative existence claim of the form ‘X does not exist’. Rhetorically, this claim functions to legitimize the idea that evidence needn’t be provided for God’s nonexistence. After all, if evidence cannot be provided for a proposition it would be irrational to expect one to provide some, and so reasonable to believe that evidence isn’t needed. But the claim that you can’t prove a negative cannot help the atheist. That is because, on each of two possible ways of interpreting what it means to ‘prove’ something, it is generally false that you can’t prove a negative (and often true that you can’t prove a positive).
Consider first, proofs which deliver certainty, as in mathematics or logic. Such proofs are sometimes possible for negative existence claims, such as the claim that there is no greatest prime number. One can also prove with certainty that there are no Xs whenever the concept X can be shown to be incoherent (like the concepts round square, or 3pm on the sun). Of course, it is true that many negative existence claims cannot be proved with absolute certainty, but the same holds for positive existence claims, for example, from science or common sense, such as that there are electrons or tables and chairs. So there’s nothing special here about negative existence claims.
Turn next to proofs which aim to establish only the probable truth of their conclusions. These are the sorts of proofs which result from successful scientific and other empirical investigations. In this sense of ‘proof’, it is easy to prove the non-existence of many things: for example, that there is no pomegranate in my hand, or no snow-capped mountains in the Sahara Desert. And while it may be difficult or impossible to even in this weaker sense prove the non-existence of many things – goblins, sombreros in the Sombrero Galaxy – the same goes for many positive existence claims – that Aristotle sneezed on his 20th birthday; that there is a transcendent deity; that there is a sombrero somewhere in the Sombrero Galaxy. So, again, there is nothing unique about negative existence claims. The unfortunate saying that one can’t prove a negative should be dropped.
3. The Burden of Proof Is On The Believer
Another familiar strategy of atheists is to insist that the burden of proof falls on the believer. If that’s right, it may allow the New Atheists to avoid evidentialism’s requirements, and rationally maintain atheism without evidence. But is it right?
The concept of ‘burden of proof’ (Latin, onus probandi) originally goes back to classical Roman law, and it remains important in legal theory. Who has the burden of proof, and what it consists of, is determined by a judge or by established rules which vary across legal systems. The same is true of formal debates which occur in a variety of formats. The idea of ‘burden of proof’ also has application in non-formal settings; for example, in academic disputes or public controversies. However, without a judge or rules to determine who has the burden and how it is to be discharged, it becomes unclear how the concept is to be applied, or even whether it has clear application.
Yet although the concept of burden of proof in informal settings is ill-understood, that does not stop many from confidently proclaiming how the burden of proof should be assigned. The most egregious mistake is to think that it is a matter of logic. Rather, the burden of proof is a methodological or procedural concept. It is, in Nicolas Rescher’s words, “a regulative principle of rationality in the context of argumentation, a ground rule, as it were, of the process of rational controversy” (Dialectics, 1977). Another error is to presume that the burden falls on whoever is making the grammatically positive statement. However, positive statements can often be translated reasonably faithfully into negative statements, and vice versa: the statement ‘everything happens for a reason’ can be expressed as ‘there are no coincidences’, and ‘there is nothing supernatural’ can be restated as ‘reality is wholly natural’. A third problem is that to be taken seriously many negative statements – ‘there are no atoms’, ‘there are no coincidences’ – require evidence, whereas the corresponding positive statements do not.
It is sometimes said that one acquires a burden of proof if one’s statement runs counter to received opinion, and it does seem that burden of proof often falls in this way. But this proposal has problems too – one being that a person can legitimately take on a burden of defending a widely-held position to those who are ignorant of it or its defense (teachers do this, for example). It may be that the best we can hope for is something like the following: in situations in which participants to a discussion are expected to take seriously the claims made by other parties, all participants bear a burden to provide support for their claims, if asked (see James Cargile’s paper ‘On the Burden of Proof’ in Philosophy 72, 1997).
The concept of burden of proof in informal settings is too complicated to sort out here, but fortunately, we don’t have to, because the question of which side has the burden of proof in an argument is largely independent of the question of what evidence is required to rationally believe any of the positions. Suppose for example that someone claims that there are no electrons, and that person bears the burden of proof. It’s not the case that so long as their burden hasn’t been discharged people can rationally believe that electrons exist without evidence. On the contrary, as evidentialism says, evidence is required for the belief to be justified even if there is no burden to defend the belief. This means that even if the burden of proof never falls on the atheist in disputes with theists (something we have so far found no reason to believe), it does not follow from that fact that atheists can rationally believe without evidence that there is no God or other divine reality. Consequently, the concept of burden of proof is also of no use to the New Atheists in avoiding the demands of evidentialism.
4. Ockham’s Razor
What about Ockham’s Razor, the principle of parsimony associated with the medieval philosopher and monk, William of Ockham? His principle is often expressed as, “Do not multiply entities beyond necessity”? Can this help? The idea, presumably, is that if a conception of reality without any divine being contains the resources necessary to explain everything that needs explaining (a proposition Ockham would have vehemently denied!) then Ockham’s Razor licenses us to exclude all references to the divine in our explanatory accounts. Might this maneuver justify atheism without evidence?
No. The trouble is that Ockham’s Razor is of little use in disputes over whether some entity X exists. That is because it is typically an open question in such disputes whether everything that needs explaining can in fact be explained without X. Theists believe, or at least suspect, that there are features of reality which are inexplicable without appeal to a divine being: the existence of a contingent universe, the fine-tuning of physical constants, etc. We need not decide here whether a divine being is needed to explain these things: what is important is just that the Razor itself cannot decide such matters. It comes into play only assuming that a complete explanation of the relevant phenomena is possible without X; at which point it licenses us to eliminate X from our ontology. But theists will not accept that a complete explanation of reality is possible without appeal to a divine being, so long as no compelling case for that claim has been made. So Ockham’s Razor can have no persuasive force in this debate.
5. Absence of Evidence is Evidence of Absence
To retain evidentialism in the absence of positive evidence for atheism, the New Atheists appear to need a principle which states that, in the absence of good evidence for theism, atheism is thereby evidentially supported. This may seem like magic, but a major theme of Norwood Hanson’s 1967 essay ‘What I Don’t Believe’, is, “When there is no good reason for thinking a [positive existence] claim to be true, that in itself is good reason for thinking the claim to be false.” Michael Scriven proposed a similar principle. So following Thomas Morris, I’ll call this the Hanson-Scriven Thesis, or ‘HST’. HST is a version of the idea that absence of evidence is evidence of absence.
Hanson defends HST in some of the ways we’ve already rejected. However, his rhetorically most effective defense involves pointing to things for which we have no good evidence – the Abominable Snowman, the Loch Ness Monster, Shangri-La, goblins – and which we also believe do not exist. His idea is that we believe these things don’t exist because we have no good evidence for them. However, he offers no argument for this latter claim. Presumably the examples are meant to just show that we reason in accordance with HST.
More recently the New Atheists have employed Hanson-like examples to defend atheism. We now hear of Zeus, the Tooth Fairy and the Flying Spaghetti Monster; then there is Bertrand Russell’s example of a china teapot orbiting the sun between Earth and Mars, too small to be detected by our telescopes. In spite of our being unable to disprove the existence of such a teapot, this doesn’t mean we must take its existence seriously. On the contrary, the rational attitude to adopt is that the teapot doesn’t exist. Russell’s point, according to Dawkins, “is that the burden of proof rests with the believers, not the non-believers” (The God Delusion).
To evaluate this example-based defense of HST, I want to distinguish two broad types of evidence. Let us call evidence for a proposition P which is usually insufficient on its own to persuade a disbeliever that P is true, weak evidence. Weak evidence, however, can accumulate to make a compelling case, and it can also support different or even incompatible propositions (think of facts in a criminal case which are cited in arguments for incompatible conclusions). By contrast, strong evidence comprises sufficient or compelling grounds for rational belief, or at least, powerful considerations which competing theories cannot account for. It’s strong evidence we’re after when we ask, “What is your evidence for that?”
This distinction is important because the “good reason” in HST must be understood as strong evidence if HST is to apply to the case of divine reality. That is because there is weak evidence for a divine reality – religious experience, the fine-tuning of physical laws and constants, the apparent contingency of the universe, etc. These and other points, although far from decisive, and although explicable in other ways, could conceivably be mentioned in a compelling argument for the existence of a divine being. Therefore, if HST is about the absence of weak evidence, one cannot infer from HST that no divine being exists. So for HST to stand a chance of applying in the atheist case, ‘good reason’ must be understood as something closer to strong evidence.
We can now see why HST is false. Consider the claim that earthworms have a primitive form of consciousness. There is little evidence for this, certainly no strong evidence. Nevertheless, many consciousness researchers believe it (with varying degrees of confidence). Or take the proposition that physical reality is much richer and more mysterious than our current physical theories represent. There is no strong evidence for this either, but it is believed by many (the astrophysicist Martin Rees, for one). Or consider string theory. Again, there is nothing that could properly be called strong evidence for it, yet many physicists believe it. Such examples could be multiplied. Yet if we were to take HST seriously, given that there’s no strong evidence for any of the above propositions, we would rationally have to conclude that the negations of the propositions are true: that earthworms are not conscious, that physics is not far from completion, and that string theory is false. But that is absurd! These negative conclusions can be believed – indeed, many people do believe them – but there is no reason to suppose that they must be believed.
It gets worse. For whenever the negations of propositions like those above can be rephrased as positive existence statements lacking strong evidence, HST will counsel us to believe contradictions. For example, the statement ‘earthworms are not conscious’ can be substituted with ‘the boundary between conscious and non-conscious creatures is above the level of earthworms’. Since there is no strong evidence for that, according to HST we should believe there is no such boundary – which means believing that earthworms are conscious! So, according to HST, to be rational we should believe that earthworms are both conscious and not. This is a reductio ad absurdum of HST.
It is now easy to see where Hanson and the New Atheists go wrong with their example-based defense of HST: they select examples that conform with HST and ignore cases of the sort just offered that conflict with it. Not only does this generate the false impression that HST is true, it suggests that religious belief, because it lacks strong evidence, must be judged to be just as ridiculous as the Tooth Fairy or goblins. But given that there are numerous non-ridiculous beliefs that lack strong evidence, it remains open that belief in a divine reality is more like those than like the ridiculous beliefs. Certainly neither Hanson nor the New Atheists have said anything to argue otherwise. Moreover, it is clear that they have no argument that religious belief is ‘ridiculous’: If they did, they would have no need to justify atheism without evidence – the argument would itself be the evidence. Here it may be objected that believers have no argument that religious belief is serious rather than silly either. That may be true, but it is irrelevant. My point is just that, in presenting ridiculous examples and ignoring non-ridiculous ones, Hanson and the New Atheists create the misleading impression that the silliness of religious belief is a result of their reasoning rather than an unsupported presupposition.
We have surveyed five ways in which the New Atheists attempt to exempt themselves from the demands of evidentialism while criticizing religious belief for failing to satisfy those demands, and we have seen that they all fail. Therefore, on matters concerning evidence and justification, the New Atheists have no good reason to treat their atheism differently from how they treat belief in the divine.
How could the New Atheists respond to this conclusion? One option is to accept that evidentialist principles apply to atheism too. Another is to reject evidentialism. Since we cannot examine these options here in any detail, let me end with a brief sketch of how I view the situation. I believe that the dispute between believers, atheists and agnostics can be modeled on disagreements in the sciences, philosophy and other fields in which there is insufficient evidence to clearly favor any position. In many such disputes, all positions have a kind of intellectual legitimacy (which doesn’t eliminate the disagreements, of course). Think of the range of legitimate positions that can be taken on the question of string theory, or on whether earthworms are conscious. Saying what ‘intellectual legitimacy’ amounts to, and on what it depends, is a difficult task. It may fall short of epistemic justification, and instead involve a kind of instrumental or practical rationality. It may also depend on inquirers recognizing the distinct value of strong and compelling evidence, and accepting that such evidence must be the final arbiter on theoretical questions. However, the main point to be emphasized here is this: the various positions that can be taken on the existence of a divine being – theism, atheism, agnosticism, and variants – are in principle no less intellectually legitimate than positions in disputes in the sciences and other fields in which none of the positions enjoy strong evidential support.
© Dr Michael V. Antony 2010
Michael Antony is a Senior Lecturer in the Department of Philosophy at the University of Haifa, Israel. He is writing a book on how to approach the question of whether there is a divine reality, and what it might be like.