Friday, January 18, 2013

Why Evidentialism Sucks

Home  >  Philosophy  >  General Philosophy

Why Evidentialism Sucks
< Prev   |   1   |   |   3   | 

Intro

While the term evidentialism can mean somewhat different things in different contexts (as is all too often the case with philosophical terms), here I’ll use the term to refer to the idea that a belief is justified if and only if it is backed by sufficient evidence, such that a belief that has zero evidence for it is not justified.

An atheist might criticize religious people for believing things without sufficient evidence on evidentialist grounds. While there is evidence for theism and Christianity, if we’re honest many if not most Christians really couldn’t cite much evidence for their faith. Some don’t have the time or the resources to do so. Are these Christians irrational for being Christians?

In one sense evidentialism is understandable; there are many cases where it’s important to have evidence, and people often undervalue evidence when making decisions. Doesn’t it seem rational then that all beliefs should have sufficient evidence?

The Regress Problem for Evidentialism

Suppose I believe something on the basis of evidence E1, say, the testimony of my friend Joe. For this to work I have to believe that that my evidence is reliable, e.g. I have to believe that Joe is reliable source of information. But evidentialism says I shouldn’t believe anything without sufficient evidence, so I need to have sufficient evidence E2 to believe that evidence E1 is reliable; I need evidence for my evidence. But if I believe evidence E2 is reliable, I’ll need evidence E3 for my belief that evidence E2 is reliable; I need evidence for my evidence for my evidence. Evidentialism thus gives us a problem of a regress. There are three possibilities here:
  1. A vicious infinite regress (E1 is evidence for E2 which is evidence for E3...).
  2. Circular reasoning (e.g. E1 is evidence for E2 and E2 is evidence for E1).
  3. Some stopping point (something that is not believed on the basis of “sufficient” evidence).
Option 1 is unworkable, and option 2 employs a notorious fallacy. An example of circular reasoning is this:
Alice: What reason do you have for believing the book is a reliable source of information?

Bob: The author of the book says it’s a reliable source of information.

Alice: What reason do you have for believing the author is a reliable source of information?

Bob: The book says the author is a reliable source of information.
Circular reasoning doesn’t really give us any genuine evidence or reason at all.

We believe our evidence is reliable to at least some degree, but because of this, if we accept evidentialism any piece of evidence we have is going to suffer from some form of regress or circular reasoning (though one could conceive of a very complex web of circular reasoning). It seems to be the case that eventually we need to reach one or more stopping points.

< Prev   |   1   |   |   3   | 

4 comments:

  1. Your argument seems to assume that the chain of evidence is sequential. However, often evidence is parallel. We might get evidence from many independent sources that accumulate the probability for belief. This applies to all of life, not just religion. We learn to trust some beliefs based on a broad range of evidence over a period of time.

    ReplyDelete
    Replies
    1. Anonymous below was correct in that I was not assuming that the chain of evidence is sequential. It could be e.g. a short sequential chain supported by a complex web of circular reasoning. But even this sort of non-sequential network won’t really solve the problem of evidentialism being unable to successfully support a belief in a non-fallacious way.

      Delete
  2. I don't think his article was positing that at all, Kevin. He rightly points out that no matter how you look at, it's an impossible epistemology to deny some beliefs as foundational or basic and still claim to have a genuine access to reality. At least that's how I interpret it.

    ReplyDelete
    Replies
    1. Yep, anonymous is right. I just used the sequential example because it’s simpler then using some complex web of evidence. You could have multiple strains of evidence but you would still run into the same sort of regress problem I wrote about above.

      Delete