Sunday, May 18, 2008

Sinking the Raft I’m Standing On

In The Non-Existence of God, Nicholas Everitt says, “the right way for the open-minded enquirer to approach the question of God’s existence is to look for grounds or reasons or evidence for thinking that God does exist, then to do the same for thinking that he does not exist, and finally to perform a metaphorical subtraction of one from the other. This will then yield the net grounds, or the grounds all-things-considered, for believing that God does or does not exist.”

One thing that humans are highly prone to do regarding matters of cognitive and emotional importance is to form a view, often before considering the evidence, and then backfill that conclusion with reasons, evidence, and arguments that corroborate it. That is, it is quite common for us to form our important beliefs in precisely the opposite manner that Everitt suggests. We believe, then we gather our evidence. We leave off before seriously considering evidence that would disconfirm the cherished view. When we do encounter negative evidence, we scrutinize it with abnormal levels of skepticism and hold it to inconsistently high levels of proof. We take a liberal and forgiving view of the sources of evidence that support the conclusions we favor. We have a powerful disposition towards confirmation of conclusions that we arrived at before we considered the evidence. We blur and sift the evidence in our favor by confusing the difference between propositions that support a favored conclusion with ones that are merely consistent with it.

But that’s all reasoning backwards from the conclusion to satisfy your gut. That’s enslaving reason to the passions. Atheists in particular loudly and proudly proclaim that they are the reasonable ones, they follow the evidence, they live by rationality, not by superstition and myth. But the real acid test here is whether we’re willing to ride the boat of reason to whatever shores it takes us to. Reason doesn’t lock onto atheism. There have been far more careful, reasonable, and very smart people in history who believed than who did not. What if in the end reason really does indicate that there is a God? Will we accept that conclusion? Or will we find a way to avoid the implication?

The only problem with Everitt’s passage the suggestion that once one has performed this investigation and metaphorical subtraction, then the matter can be settled. But what science has taught us is how to resist the temptation to settle on any answers, even ones we think are decided. The most important lesson we can learn from science is how to actively and perpetually seek out disconfirming evidence.

Lots of believers and non-believers go to the experts who favor their view, who agree with them, to get corroboration for what they already think is true. And when that outspoken and articulate champion of your view presents it in a pithy, clear, powerful, or insightful way, we feel vindicated. But we’ve got to resist the urge to seek out backfill for the conclusions we want to draw. We need to openly and honestly confront contrary views, not for the purpose of refuting them or scoring rhetorical points, but in order to give them long and serious consideration, and give them credit where credit is due. It’s more important to know the truth than to feel vindicated. It’s more important to consider every available source than to draw a premature and tilted conclusion.

The question is, how hard have we tried to understand the opposite viewpoint? How much serious consideration have we given to the counter-evidence?

3 comments:

mikespeir said...

"Atheists in particular loudly and proudly proclaim that they are the reasonable ones, they follow the evidence, they live by rationality, not by superstition and myth. But the real acid test here is whether we’re willing to ride the boat of reason to whatever shores it takes us to."

It's unquestionably true that atheists make this mistake a lot. (Not me, of course.) If we have an edge, though, it's in that we're likelier to recognize the problem and try to avoid it.

"The question is, how hard have we tried to understand the opposite viewpoint? How much serious consideration have we given to the counter-evidence?"

Good question. An excellent exercise is to write out a short essay arguing for the opposing point of view, really, sincerely trying to make the case. It's illuminating, to say the least.

David B. Ellis said...


"The question is, how hard have we tried to understand the opposite viewpoint? How much serious consideration have we given to the counter-evidence?"


Quite a few of us atheists, myself included, were once theists.

For myself, I know I tried very hard, once I began to question my religious upbringing, to find a rational basis for what I had previously believed on authority.

But I never could. Much as I tried.

Bryan Goodrich said...

One thing I see Everett having left out of the equation is to consider the entire domain of possible evidence. The equation calls for "evidence for God" and "evidence against God" and then, essentially, balance the two for the best answer. It doesn't take in account "how much evidence are we lacking?" A non-response in statistics is important. If all you consider is that your sample was those who responded, then you get a bias. The goal in these kinds of inferences-from-evidence is to make sure we're not biasing our results. Everett does not provide that by what you have stated here.

To make it in another logical fashion, the domain is not partitioned by these two sets of evidence, i.e., there's more out there not considered in this equation that may or may not apply. As you conclude, what if, in the end, the evidence supports God?

Let's say our current evidence against God outweights the evidence for God, but the evidence for God is only 5% of the total, while the evidence against is 95%. If we theoretically had 100% of the evidence for and against it shows we ought to be for God, then our current evaluation by Everett's model is just wrong. How might we determine our level of evidence? How might we consider what a "true model" might be? i.e., the ideal model with 100% evidence. That is, essentially, what would make Everett's equation hold (fails if we're any less)?

Therefore, I'm not suggesting we shouldn't appeal to the evidence, I just think it's too simplistic to use like a standard because it does not provide any meaningful insight into considering the sample itself. Without that kind of meta-analysis we will have biases and not have considered them.