My friend and one of my most excellent high school teachers, Johnston Smith, proposed an abductive argument for the existence of God that is a version of the argument from morality. Today, I'm going to discuss what makes for a good abductive argument drawing from précis of philosophical literature on explanationism done by John Danaher and Luke Muehlhauser. Later, we'll take a look at Johnston's argument and see why I think that it, like all abductive arguments for the existence of God, is not very good.
Abductive arguments make an inference to the best explanation for some empirically verifiable phenomenon. They look something like this:
1. H is some hypothesis that would explain some collection of data (such as facts or observations)
2. No other hypothesis explains the data as well as H does.
3. Therefore, H is probably true
Say that you arrive at your car to find that one of your tires is flat. The hypothesis that you ran over a nail or screw at some point is a reasonable explanation based on the knowledge that this is a common cause of flat tires. The fact that you drove through a new home development with plenty of construction the previous night pretty much clinches the nail/screw explanation as the leading one. You're making an abductive argument that it's probably true.
Notice that unlike a sound deductive argument, the conclusion of an abductive argument is not guaranteed. Abduction can be thought of more like an inductive argument in this regard: the conclusion is merely made more probable. After all, some better explanation might come along, or a less good explanation can become the best explanation in light of new evidence. For instance, the nail/screw explanation might not be such a good one if you discover that you failed to notice that all of your tires are flat. Foul play would then start to look more compelling as the best explanation of all of that evidence.
Abduction is how Sherlock Holmes admirably solves crimes. It's also how science works. It's used to develop a hypothesis that explains some evidence. Scientists then attempt to falsify the hypothesis. I'm not suggesting that when you examine your tire and fail to find a nail or screw, falsifying that explanation, that you are doing science, but the principle is exactly the same. Hypotheses that fail testing are rejected or modified while those that pass the most and best tests survive. The best scientific explanations are developed into Theories. Accordingly, the common criticism that the Theory of Evolution is "just a theory" is no criticism at all. It merely reveals a misunderstanding of how science works.
Assessing abductive arguments rests in understanding what makes for a good explanation. Why, for instance, would "magic!" be a poor explanation for the unexpectedly empty bank vault? Why would robbery be a good explanation? The strengths and weaknesses of competing explanations can be assessed by comparing the "explanatory virtues" that they possess.
As you might have guessed, the ability to test the explanation ("testability") is a very important explanatory virtue. To be good, an explanation should not only render testable predictions, but it must pass the tests. Quite simply, better explanations pass more and better tests.
Better explanations are more consistent with our background knowledge. This helps to shed light on why magic is a poor explanation and robbery is such a good one for the empty bank vault. The healing capabilities of acupuncture are not well explained by changes Qi ("vital energy") because Qi has no place in our background knowledge of physiology. On the other hand, placebo effect, which we know works with pills, injections, lasers, surgeries, etc. is an excellent explanation.
Better explanations are simpler. A nail in the tire is a better explanation for a flat than that hundreds of genetically engineered and CIA-trained ants mistakenly did it as part of a botched covert operation. Each component of the explanation has to be true for the whole explanation to be true. Explanations with fewer components or steps are naturally going to be more likely.
Better explanations have better explanatory scope, meaning that they explain a wider variety of data. Changes in Qi fail to explain why acupuncture doesn't "work" to cure widely metastatic pancreatic cancer, blindness due to retinoblastoma, and Ebola infection, for instance, but placebo effect explains these failures perfectly well.
In addition to those, philosophers (like Peter Lipton, Gilbert Harmann, Wesley Salmon, William Lycan, Paul Thagard, and others) have identified other explanatory virtues, like precision, mechanistic insight, unification, predictive novelty, and past explanatory success, but some of them seem to overlap with the four I've discussed here in certain ways. In the interest of simplicity, I'm going to suggest that Johnston and I limit the assessment of his argument from morality to the four virtues I described above (testability, consistency with background knowledge, simplicity, and explanatory scope).
Johnston's abductive argument was proposed as follows:
"I suggest that the idea of a compassionate, loving God WHO SEZ better explains our human yearning for an assertion of justice than does the proposition that there is no such God" - Johnston SmithThe data that needs to be explained is "the human yearning for an assertion of justice". I agree with Johnston that such a yearning does exist and that it requires an explanation. Johnston suggests that the hypothesis that a compassionate, loving God exists provides a better explanation for that data than naturalism does, therefore, God probably exists.
Next time, I'll take a closer look at Johnston's theistic explanation and naturalistic explanations and see how they fare with respect to these four explanatory virtues. Remember, the best explanation points to what's probably true.