As the Emperor in Amadeus is wont to say, "Well, there it is." We have scarcely got through Part II of twelve parts, and both the Cosmological Argument and the Argument from Design are destroyed. What is Hume to do to keep his conversation afloat? His answer, to which he returns many times in the course of the Dialogues, is simply not to allow his characters to recognize that they have been defeated.
The absurdity of this interpretation is on its face: it makes the entire unity of the Dialogues consist in the stupidity of its characters, not just the stupidity of Demea (who admittedly is toyed with by Philo), but also the stupidity of Cleanthes (who is in fact treated throughout as an intelligent person) and the stupidity of Philo (who apparently never realizes that he can press these supposed advantages). The catastrophe of this interpretation, that it makes the work nothing more than a string of argumentative episodes strung together in poor literary fashion by nothing more than an unconvincing and implausible refusal on the part of its characters to recognize that they've been refuted conclusively, should in and of itself lead one to be suspicious of it. Wolff makes the comparison to Monty Python; but when we have a philosophical text that Hume obviously took seriously -- he put an immense amount of thought into it over several years, worked on it while he was dying, and did everything in his power to guarantee that it would be published after his death -- treated as something with no more coherence or rhyme and reason than a Monty Python skit, we should perhaps look again.
The fact of the matter is that Philo has not refuted Cleanthes's design argument, and Hume is well aware of it for four reasons. First, because he has a more sophisticated understanding of what is at stake in design arguments than people nowadays usually do; the design argument became so popular, and so widely regarded with respect, because it was closely tied up with philosophical accounts of science itself. In Hume's day this is very clear when you look at who was actually presenting design arguments, and they are the major scientific minds of Britain: Boyle, Newton, Newtonians like Colin Maclaurin, and the like. If you read the Dialogues closely, you notice that scientific inquiry comes up a lot, and there is a reason for it: the design argument is closely connected at the time with people's conceptions of what science does. Science discovers the laws of nature divine providence establishes throughout the world; the world is intelligible because there is an intellect that constructs it intelligibly. It's a common view in the period. And Hume cannot stop the argument before addressing it, because then it leaves up in the air whether and how the world has enough intelligibility to it to allow us to engage in scientific inquiry.
Second, Hume is not like some graduate student in analytic philosophy, amateurishly treating each argument as if it stood on its own, with a stand-alone refutation that has no implications for broader thought. Hume isn't looking for just any answer to the design argument; for very obvious reasons he wants one that is consistent with Humean principles, and does not have the implication that large parts of Hume's own worldview are wrong. And a close look at Philo's argument suggests that it is not consistent with things Hume has said elsewhere -- it presupposes an account of analogical inference that Hume does not actually accept, plausible as it may seem at first glance, and fails to take into account a feature of the design argument -- namely, its apparently intuitive plausibility -- which any moderate skeptic like Philo or Hume himself must take into account. In going for the kill, Philo, on Hume's own principles, has overstepped what he can actually commit himself to -- his 'refutation' of Cleanthes is, as Cleanthes quite rightly points out, inconsistent with his Humean defense of skepticism against Cleanthes's earlier arguments. I have talked about this before. It is also inconsistent with Hume's account of analogical inference in the Treatise.
Third, one of the major themes of the Dialogues is sociable friendship. While Demea is a bit of an interloper, Philo and Cleanthes are good friends; they remain so throughout the Dialogues, and much of Part XII is inexplicable unless we take this to be one of the important points of the work. Hume is not interested in attacking the design argument then calling it a day; he gave an early draft of the book (probably the first three parts) to a friend to make sure that he was portraying Cleanthes in the strongest light, and it is pretty clear that the book is supposed to convey the message that proponents and opponents of the design argument can get along amicably if they are moderate and rational about the subject. This message would not be served all that well if Cleanthes shows up, is roundly shown to be stupid, and the book ends.
And fourth, we have no reason whatsoever to think that Hume thought he ever had a solid refutation of the design argument, one that he himself could regard as solid, anyway. Indeed, in every text we have from Hume on the subject, throughout his entire life, he accepts that it tells us something about the world. Some of these comments are difficult to interpret, but it seems clear that his own view is that the argument works -- it just doesn't yield the very robust metaphysical and religious conclusions the Newtonians assume it does. This is in fact precisely what Philo himself will say in the Dialogues. Hume takes the argument far more seriously than Wolff does to begin with; as someone faced with Newtonian embrace of it, he has to do so, and more than that, he has to take it seriously because he believes that it makes a genuine point that must be taken seriously.
When we recognize that Philo's 'refutation' is not a refutation, and that Hume's recognition of this in Cleanthes's ability to confound Philo is not some arbitrary and senseless maneuver, the real brilliance of the Dialogues becomes clear: it is not a string of argumentative episodes linked implausibly by the stupidity and obstinacy of its characters; it is a crafted and unified text in which Philo and Cleanthes are both portrayed as bright people scoring genuine points off each other, and in which Hume takes the argument being considered as a serious matter with extensive ramifications both for philosophical Newtonianism (which Hume respects) and Humean skepticism itself, not as some set-piece that can be dispatched with a brief bit of undergraduate argument whose broader implications are never explored. The Dialogues don't end at Part III because Philo needs to try a new approach to the argument, which he does with a fair amount of success.
And it shows why Hume should be taken seriously as a philosopher. More than two hundred thirty years in the grave, and he still has a better grasp on the problem, and a more sophisticated approach to it, than academic philosophers usually do.