There's an argument for non-mainstream, non-scientific beliefs* that goes something like this: "You can't know anything. You can't really ever know what other people are experiencing, what things happen to them, and what's true for them. So to claim that there's a right answer to questions about reality is just always wrong." It sounds like a stupid argument, doesn't it? But I've had conversations -- I've had many conversations -- where people actually try to defend their supernatural or anti-scientific** beliefs with that argument. Not a more sophisticated version, or a subtle take on that argument, but that argument itself.
So, for a moment, I'm going to take it seriously.
After all, it's a valid proposition. It's true that we can't ever know anything to 100% certainty. It's also true that science frequently 'changes its mind' about things, if only in the sense that new information often subtly and sometimes radically changes our best understanding of the probable reality. For example, apparently in the last two decades, the consensus changed from the belief that dogs were more closely related to cats than bears, to the opposite.
But that doesn't mean everyone's completely off the hook for deciding what to believe in.
Once we've agreed that we can't know anything with perfect certainty, we have to decide upon a method for deciding what we will believe, and what we won't.
I'm going to assume, for the sake of argument, that being able to function in life is among your goals. We've already agreed that rational argument is a valid route to truth, because if we don't concede that things rationally make sense, then concepts like "100%" and "absolute certainty" don't belong in the argument.
The next step to take is to agree that our experience of the world outside ourselves, our sensory experiences, have at least some connection to reality, however vague. Without this belief, we are so deeply without information that we can't take any sort of action that could conceivably have meaning.
As far as I can tell, the earliest point of functional acceptance travelling from radical doubt to absolute credulity is an acceptance of science where it's available, less-than-scientific evidence where it's not available and a conclusion is absolutely necessary, and a whole lot of "I don't know."†
Alternately, we could go with any other level of doubt -- levels which range from crippling solipsism to self-endangerment. But I think the most damning argument against more credulity than I've proposed is that not much farther past the position I've proposed, you're faced with the reality that there are several, if not hundreds or thousands, of roughly equally valid positions.
There are a great many reasons why this is philosophically unsound, but my biggest objection to it is that it makes a person ethically suspect. Because their behavior is based on essentially unknowable variables, what they believe is 'ethical' becomes both inarguable and, usually, nonsensical. Greta Christina has argued, and I agree, that this is the greatest danger of religion.
One of the great benefits of science is that it narrows down to a very small number of potentially right answers. Even if you don't believe that science is always right, or that it's the only valid window into the truth, it certainly offers the best odds.
*I'm not saying that all mainstream, scientific beliefs are true -- they're often proven wrong, mainstream beliefs especially, and one should be duly credulous of all claims not clearly supported. But there are arguments that make sense for deconstructing beliefs in that category. This one doesn't. **For the purposes of this blog post, I'm defining "Anti-scientific" as "explicitly contradicted by accepted science, and not supported by any cutting-edge science." Examples: Homeopathy, vaccines as a cause of autism, '9/11 truth' conspiracy theories. †I hope to elaborate, in future, on my views about levels of doubt. I think the issue is more complicated than the points I've laid out here, although I don't think any of the complications are particularly relevant to this argument.