In the very short and accessibly priced Atheism and Agnosticism noted philosopher of religion Graham Oppy aims to give a brief description of atheism, agnosticism, and theism, and explain why atheists and agnostics believe as they do. In the first chapter, after noting that atheists are one of the most despised minority groups and in some (mostly Muslim) countries state laws allow for the killing of atheists (1–3), Oppy asserts that atheists ‘believe that there are no gods’ and agnostics ‘suspend judgement on the claim that there are no gods’ (4). Theists, then, ‘believe that there is at least one god’ and monotheists ‘are theists who have the further belief that there is no more than one god’ (5). Not everyone will be happy with such definitions, particularly in the broader field of religious studies, where atheism is often seen as more inclusive, and even in the philosophy of religion, where theism is typically taken to be narrower in scope, with the theistic world-view actually being one of the monotheistic world-views (itself a subset of supernaturalism, for which Oppy does not provide a definition, but which is generally taken to be ‘not naturalism’). However, what matters in philosophical research is that authors are very transparent about how they use such terms; and, as ever, Oppy is very – even tediously – precise about this. Even with these contested definitions, Oppy does a masterful job in explaining the great diversity of agnostics and atheists (7–10). He also easily brushes away the absurd notions that atheism and agnosticism must be religions, even if his definition of religion may be too narrow for many (14–15). The focus of the book then shifts to how atheists and agnostics could arrive at their positions, with Oppy having correctly noted that if absolute proof were required we would essentially all be agnostics (12).

Chapter 2 is concerned with method. I was pleasantly surprised – since in the past we have had minor disagreements about whether his ‘argument from dominance’ approach to arguing for naturalism is effectively equivalent to my explicitly probabilistic ‘Bayesian without numbers’ approach to arguing against theism – that Oppy describes the need to weigh up virtues and data for various ‘big pictures’ (hypotheses); refers to ‘calculations of comparative theoretical virtue’; asserts that ‘The best big pictures are those that make the best trade-off between minimisation of theoretical commitments and maximisation of explanatory breadth and depth’, which, to a Bayesian, sounds very much like trading off between prior probabilities and likelihoods; effectively says just that ‘very crudely’ soon after, even invoking Bayes's Theorem; and agrees that numbers are not always required ‘in order to be able to do Bayesian calculations’ (25–28). Despite this, Oppy decides to proceed without substantially mentioning Bayesianism again, and, as one of the world's foremost experts on the ontological argument for God's existence, ends the chapter with a quick refutation of that argument (31).

The highlight of the book is the third chapter, which has Oppy outlining a very compelling ‘Case for Preferring Atheism to Theism’ (32). In actuality, Oppy presents a ‘case for naturalism’, which is one of the forms atheism can take (34); naturalism being described as ‘the view that (a) there are none but natural entities with none but natural causal powers, and (b) well-established science is the touchstone for identifying causal entities and causal powers’ (13). He explains that naturalism is more ‘minimal’ than theism, in that it makes fewer commitments; further citing creatio ex nihilo as one unnecessary commitment many theists make (34–36). A Bayesian would say that naturalism has won on the prior side of the equation, and that if naturalism wins or at least does not lose on the likelihood side it should be considered the superior hypothesis. Indeed, Oppy goes on to explain expertly that nothing is better explained by theism than naturalism, including the fine-tuned parameters of the Universe, the alleged existence of minds, reports of miracles, claims of religious experiences, and scripture; in fact, much of this is outright better explained by naturalism (37–48). Clearly, then, naturalism is to be preferred.

Less impressive is chapter 4, which has Oppy attempting to construct a case for agnosticism. This may have been difficult because Oppy is himself an atheist, as well as a naturalist, and because an objective examination of the evidence does indeed indicate that naturalism is the better hypothesis over theism, making Oppy's sort of agnosticism untenable. I similarly struggle to imagine a good case for theism, when almost all of my time is spent arguing that naturalistic and supernaturalistic alternatives to theism are far superior probabilistically. The first, and major, problem is that Oppy asks us to put naturalism to one side, so that the agnostic can declare that atheism cannot be said to make fewer commitments than theism (52–53). Of course, this is no longer atheism, then, but atheism′. In a more general sense, if we were allowed simply to ignore evidence, as theistic apologists, conspiracy theorists, and political ideologues are often wont to do, we could argue for any position – something that competent Bayesian reasoning prevents. In any case, the rest of the analysis is thus effectively nullified.

Nevertheless, Oppy returns to the issue of how well atheism and theism explain the evidence, but again will have critics scratching their heads, and not only because this section could almost have been copied from the previous chapter since even on his case for atheism he did not necessarily claim that atheism better explained the evidence. Oppy now wishes us to put the non-naturalistic atheisms to one side and focus on naturalism (53). But now it seems that we are no longer concerned with atheism or even atheism′ anymore, but atheism″. There also seems to be some asymmetry here as cases for theism and naturalism are apparently expected to rely on objectivity, while Oppy is content for the agnostic to be swayed by what they thought was their own religious experience or how they felt about certain holy texts (58–60). This all does seem disingenuous, inconsistent, and idiosyncratic, perhaps leaving Oppy achieving the exact opposite of what he intended: (his form of) agnosticism is simply unreasonable. In reality, however, our evidence is always limited and ever changing, and Oppy knows this, so some form of agnosticism is always viable. He sensibly concludes:

Atheism and Agnosticism is an affordable and straightforward book that provides a framework for discussing and comparing various ‘big pictures’. The concise case for naturalism therein, presented by a secular philosopher at the top of his game, is alone worth the price of admission. It should prove a handy resource for high school students, undergraduates, and laypeople.