Cleanthes finally realizes that he has hit a pretty big obstacle to his empirical theism, and he tries to surmount this obstacle by sacrificing the "infiniteness" of God. God, he says, might just be finitely perfect. This might explain why he cannot make the universe exactly as he wants it to be: as a finitely perfect God he must make some compromises. His benevolence is regulated by wisdom and limited by necessity.
Philo, though, shows Cleanthes that this sacrifice will not help him. Even this sort of God (a finitely perfect one) cannot be inferred based on the evidence available to us. It might well be true that ours is the best world that God could have provided for us given the constraints within which he was working, but when we look at the world we cannot see that, and thus the evidence provides us with no basis on which to conclude that God is at all perfect. To illustrate his point, Philo draws an analogy between our universe and a very ugly palace. It might well be the case that the architect responsible is highly skilled and that he managed to produce the absolute best dwelling space he could have produced given the limitations of his materials and space. But still, when we look at the palace and see only that it is terribly ugly, we have no grounds on which to infer that the architect is talented.
Philo then shows us in what ways our universe appears to be like the ill- constructed palace. He identifies the four sources of misery in the world, and shows that (as far as we can tell) they are all unnecessary. First, there is the existence of physical pain. The purpose of pain is to motivate us to avoid certain actions and to engage in certain others. But this same goal could be fulfilled with only pleasure: God could have arranged the world so that those actions that were good for us cause pleasure, and those actions that are bad for us cause a lessening of pleasure. For instance, instead of the sensation of burning, fire could simply make us feel less nice than we would feel otherwise. Pain could then be cut out of the picture entirely. The second source of misery is the fact that the world is conducted by general laws. It would be better for us, Philo claims, if God ran the world by particular volitions, or if he at least regularly suspended or manipulated the laws of nature in order to produce a greater good. For example, whenever a small innocent child found itself in the way of a launched bullet, God could suspend the laws of physics and either change the bullet's trajectory, or cause the bullet to vanish into thin air, or do any number of other things to prevent the tragic meeting of bullet and child.
The third source of misery is the limited abilities of every particular species. God just gave each species what they needed in order to survive; he did not give any of them what they need in order to really thrive and be comfortable, safe, and happy. All human beings need in order to greatly improve their lot in life, for instance, is a much greater diligence. If God had given this gift to us then we would be morally, intellectually, and practically superior beings and most of our problems would disappear. The final source of misery is the fragile nature of the universe. In order for our world to function at its best, the conditions need to be just right. Too little of something and there is a disaster; too much of something and there is a disaster. For example, if there is not enough rain then there is drought and plants and animals die; if there is too much rain then there is flooding and plants and animals die. The same can be said of wind, heat, bodily fluids, and innumerable other earthly variables.
What can we conclude from a world with these four seemingly unnecessary features? So long as we have an independent belief in a benevolent God, Philo assures his friends, these features of the world are not sufficient to disprove our faith, since we can presume that God's goodness and the world's evil can be reconciled in some unknown way. But we certainly cannot infer the goodness of God from such a universe. Actually, if we do try to infer God's moral attributes from the world (which Philo thinks we should not do) then the only proper conclusion to draw is that God is neither good nor evil, but entirely indifferent to these principles—that God, in other words, is morally neutral.
Demea is offended by this last part of the argument and finds an excuse to exit.
Why is God's moral neutrality the proper conclusion to draw from the evidence? Philo tells us that it is because there are four possible hypotheses about the moral nature of the first cause of the universe, and all but one of them can be eliminated. But he does not go into much detail about this surprising claim, and so it will be helpful to go through it ourselves.
These are the four possible hypotheses about the moral nature of God: (1) God is perfectly good. This, naturally, is what the empirical theist wants to conclude from the evidence. (2) God is perfectly malicious. Almost no one would want to hold this view. (3) God is both good and malicious. (Or, alternatively, there are two different forces in the world, one good and one evil). Several sects have believed in a picture like this one, most notably the Manichæans cited by Philo. And finally(4) God is neither good nor malicious.
Given the mixed nature of our world and assuming that God could take away or augment suffering, the first two possibilities are immediately ruled out. We cannot infer a perfectly good God from a world that contains evil, and we cannot infer a perfectly evil God from a world that contains good. The reason that the evidence does not support the third possibility is more difficult to understand. It would seem that a mixed world like ours points precisely to some dueling moral forces, such as God and Satan. This inference certainly seemed strong to many cultures which posited just such a picture. Nearly all ancient mythologies make use of this sort of explanation to account for the mixture of blessings and curses in our life, and there are numerous sects in the Judeo-Christian tradition that uphold this view as one of their central tenets.
Philo points out, though, that our evidence does not point to this picture at all. Good and evil in our universe follow general laws. There is no indication of any struggle. If there were a struggle taking place (between two beings, or between two sides of God's nature) then we should expect to see sudden we should expect to sometimes see miracles performed to benefit the good, and other times see feats performed to help out the evil. Instead what we see is that our world simply operates based on immutable laws of nature, and whether one is harmed by them or benefited by them has nothing to do with how good or evil one is. Bad things happen to good people and bad things happen to bad people, but (as far as we can tell from the evidence) only because both good people and bad people are subject to the laws of nature.
All we are left to conclude, then, is that God is indifferent to good and evil. The laws that he set up affect the good and the evil exactly alike because they were not set up with any moral preferences in mind. Certainly, we might think that if God were perfectly good he would have at the very least set up laws of nature to benefit the good, but he did not even do this. So far as we should trust the evidence, then, we should conclude that God is morally neutral.
The next obvious question to ask is how far we should trust the evidence. Philo claims that we should not trust the evidence at all: instead we should rely on revelation to tell us what God is like. But Hume might not have shared Philo's view on this issue. As a serious empiricist he believed that all we should trust in is the evidence. The wise man, he says in the Enquiry, is the man who proportions his belief to the evidence; in other words, what it means to be wise is to try to believe only what you have evidence to believe.
Philo's argument, taken in this light, has a much stronger conclusion: his conclusion is not merely that we would be forced to believe in a morally neutral God if we were so silly as to try to draw religious truths from experience, but rather his conclusion is that we really should believe, given the evidence, that whatever first cause is responsible for our universe is morally neutral. This conclusion brings us a long way toward atheism. If the first cause of the universe is morally neutral, what kind of a God could He be? The first cause might just as well just be some impersonal law of nature, which is, as we have seen, exactly what Hume seems to think the first cause is, in fact. Even while his characters constantly affirm their belief in God's existence, it is possible to read this text as if Hume were subtly weaving in an argument for atheism.