It seems to me that Torley clearly has the upper hand in this exchange so far. As a debate judge, I would “flow” the entire “debate” to Torley up to this point. But that doesn’t mean game over for Loftus, however. In each case, I think Loftus has strong replies available.
Here are my brief comments on Torley’s points.
Mistake #1. Loftus’ failure to take account of prior probabilities
As a Bayesian, I agree that taking account of prior probabilities is essential. Of course, I also think metaphysical naturalism, which entails atheism, has a higher prior probability than theism. And there is no good reason to think that theism has a significantly higher prior probability than naturalism. So this “mistake” should be easy enough for Loftus to overcome.
Mistake #2. Loftus’ illegitimate narrowing of the evidence set
Again, I agree that, if one is to going to conclude that a hypothesis (such as theism) is probably false, one must consider the total relevant evidence. And, again, I think this mistake should be easy enough to correct by demonstrating how typical theistic arguments commit the fallacy of understated evidence. Once one considers the total evidence, it’s far from obvious that God exists.
Mistake #3. Loftus overlooks the fact that his “no-God” hypothesis explains senseless tragedies, only if physicalism is true
At the outset, I’d like to compliment Torley for this creative objection, which I hadn’t run across before. In essence, it asks us to treat physicalism as an auxiliary hypothesis and to consider the evidential implications if physicalism is true and if it is false.
Let’s distinguish between two different ways in which any hypothesis may “explain” the data. One way a hypothesis can explain the data is to make the data more probable than not. Arguments which attempt to show this are called “P-inductive” arguments.
The other way, however, does not require that the hypothesis makes the data more probable than not. Instead, the idea is to show that the data increases the probability of a hypothesis. These arguments are called “C-inductive” arguments. One way to defend a C-inductive argument is to show that the data is more probable on one hypothesis than it is on another, rival hypothesis (i.e., a second hypothesis which is logically incompatible with the first).
Torley may well be correct that, on the assumption that physicalism is false, there is no good P-inductive argument for atheism. But he overlooks the possibility that a strong, C-inductive argument for atheism remains: it may be the case that (and I think is the case) that tragedies like the Sandy Hook massacre are much more probable on the assumption that naturalism is true than on the assumption that theism is true.
Furthermore, Loftus could point out three additional facts which would strengthen his argument. (All three of these facts are pretty much plagiarized from Paul Draper’s writings.) First, not only does the world contain much horrific suffering, but it contains relatively little glorious pleasure. Second, horrific suffering often destroys a person, at least psychologically, and prevents them from growing morally, spiritually, and intellectually. Third, God is silent in the face of tragedies like Sandy Hook, in the sense that victims of tragedies rarely report feeling God’s comforting presence. All three of these facts are more probable on the assumption that Loftus’s “no-God” hypothesis is true than on the assumption that God exists.
Loftus has two replies available which, I think, render this objection moot. First, Loftus could point out that Torley has committed the fallacy of understated evidence. Let’s assume that the “no-God” hypothesis indeed fails to explain the universe in the first place. But the fact that the universe exists hardly exhausts the available cosmological evidence. Given that the universe exists, the fact that it began to exist with time, rather than in time, is evidence favoring naturalism over theism. Or again, given that the universe exists, the fact it does not exist on a human scale is evidence which weakly favors naturalism over theism. So it’s far from obvious that the available cosmological evidence, once fully stated, favors theism over naturalism.
Second, Loftus could point out that the evidence from tragedies is prima facie evidence, i.e., that all other evidence held equal, tragedies such as Sandy Hook make the “no-god” hypothesis more probable than the god hypothesis. If other evidence favors God’s existence, then so be it. But that other evidence, if it exists, doesn’t deny the fact that tragedies are more probable on the assumption that naturalism is true than on the assumption that theism is true.
Mistake #5. The physicalistic version of Loftus’ “no-God” hypothesis fails to explain the emergence of life
I think Loftus can use the same two replies to mistake #4 here. In addition, Loftus has a third reply available: cosmological fine-tuning arguments and biological design arguments are at odds with one another.
Mistake #6. Predation is not senseless, but a necessary fact of life
Again, I think Loftus has two replies available. First, predation is a necessary fact of life only because of the laws of nature. If theism is true, however, God could have designed the laws of nature differently so that predation of sentient animals is unnecessary. Second, facts about the biological role of pain and pleasure and the flourishing and languishing of sentient beings are antecedently very much more probable on naturalism than on theism. And note these facts are logically independent of Loftus’s appeal to facts about tragedies, like Sandy Hook.
Mistake #7. Loftus fails to account for the marvel of the human brain
Again, Loftus can point out that, at best, Torley is committing the fallacy of understated evidence. Assume that the complexity of the human brain requires a designer and so is evidence favoring theism over naturalism. Given that a human brain exists, however, the fact the mind depends upon the physical brain is evidence favoring naturalism over theism.