Cleanthes therefore tries to show that religious belief can be supported by arguments of just the same kind as are used in science; hence the design argument. Philo’s attacks are aimed at showing that the logic of the argument is not of the same type. Already by the end of Part VIII, he has achieved this aim. But at a number of places, and very clearly in Part XII, he accepts that it is, nevertheless, very natural to believe that the world is the creation of ‘a first intelligent author’. What the sceptical technique does is to undermine the claim of the natural theologian that this natural assumption – what Cleanthes calls ‘the religious hypothesis’ – can be proved by reasoning of a kind employed in science, namely, analogical reasoning from effects to causes.

Apart from his own philosophy, Hume used a number of sources to construct the positions of his characters. As was mentioned earlier, Butler is one source for Cleanthes, in his use of analogical argument and his willingness to allow that probable reasoning is a proper method in natural theology. Demea’s cosmological argument is based on sections of A Demonstration of the Being and Attributes of God (1705), by the influential theologian and Newtonian Samuel Clarke (1675–1729). And R. H. Hurlbutt has identified sources for parts of Cleanthes’ speeches in two other Newtonians, George Cheyne (1671–1743) and Colin Maclaurin (1698–1746).21 But it is, in my view, better to conceive of Hume’s characters as representing types of theologians, rather than to think of them as standing each for a particular historical thinker. Of course, Philo’s speeches far outweigh those of the others, both in quantity and in sophistication, and for the most part represent Hume’s views. But Hume does enough, all the same, to give Philo some life of his own, particularly in the last Part. Here Philo expresses some elements of fideism:

A person, seasoned with a just sense of the imperfections of natural reason, will fly to revealed truth with the greatest avidity… To be a philosophical sceptic is, in a man of letters, the first and most essential step towards being a sound, believing Christian.

This idea, that scepticism, by undermining the pretensions of reason, makes it possible for us to accept religious revelation by faith alone, has a long intellectual history. It would be perfectly familiar to Hume, not least from the writings of Pierre Bayle (1647–1706). But there is no doubt that Hume himself had no such tendency.

At the end of the Dialogues, Pamphilus judges that Cleanthes’ principles are nearer to the truth than the others. This, too, is no guide to Hume’s opinion. It echoes the ending of De Natura Deorum, where Cicero claims to have found that the theology proposed by the Stoic speaker ‘approximated more nearly to a semblance of the truth’ than the scepticism of the Academic spokesman. Since Cicero is identified in that work as one of the ‘disciples of Philo, and have learned from him to know nothing’, Hume is using the rhetorical device learned from his classical model. His intention in using the dialogue form is to avoid a direct statement of his position. He could well have had in mind Cicero again:

Those however who seek to learn my personal opinion on the various questions show an unreasonable degree of curiosity. In discussion it is not so much weight of authority as force of argument that should be demanded. Indeed the authority of those who profess to teach is often a positive hindrance to those who desire to learn; they cease to employ their. own judgement…

Additionally, the dialogue form enables Hume to practise the sceptical technique of balancing opposing arguments. According to his own theory of the nature of belief, this leads to a state of equilibrium, freeing the mind from dogmatism. This is illustrated most dramatically in Part XII, where Philo suggests that the end result of the discussion is a verdict so meagre in its content that whether we regard it as favouring theism or not is a purely verbal matter. But it is crucial, in Hume’s eyes, that this minimal conclusion – that the cause or causes of order in the universe probably bear some remote analogy to human intelligence – is strictly a philosophical conclusion. It belongs to what Philo calls ‘the philosophical and rational kind’ of religion, and, as such, has no practical consequences whatever for how we ought to live our lives. And, of course, no sane man will think it worth the spilling of a single drop of blood. Hume consistently argued that where men and women hold religious beliefs as certain truths, regard all who do not share them as in error and seek to enforce religious practices, the consequences are always pernicious. But a purely theoretical examination of natural religion is, by its very inability to achieve results of any consequence, itself an antidote to the dogmatism and passion of popular religion.