Sunday, February 17, 2008

Process Reliabilism

In a comment on Stephen Law's blog, Timmo propounds the epistemic theory he labels Process Reliabilism:
(R) If an agent A believes that p, then A is justified in believing that p if and only if there is some reliable-belief forming process P and A's belief that p is a result of that process P.
(For some reason, Timmo labels this statement as externalist, however, it seems entirely internalist to me.) His view, however, is deficient for a number of reasons.

Most importantly, we have no rigorous account of what constitutes a "reliable belief", or even a rigourous definition of "reliable": We're simply moving the lack of rigor from "justification" to "reliable belief". And why abstract epistemology at the definitional level to a process? If we can tell the difference between a reliable-belief forming process and a process that does not form reliable beliefs, then we can tell the difference between between a reliable belief and an unreliable belief, and why not just base our epistemology on that faculty.

I can conclude only that this sort of philosophical mumbo-jumbo is employed so that an advocate can simply define some process as reliable-belief forming, and thus justify the beliefs formed by this process. As Timmo notes:
Ibrahim can hold that he recognizes the truth of the Koran's teachings through the testimony of his sensus divinitatis. Moreover, because this faculty, which has been given to him by God, is a reliable one, he is justified in believing in the Koran.
Timmo also notes that all our beliefs, including those about external reality are, in some sense unreliable:
All of our belief-producing faculties fail under certain conditions: memory suffers from confabulation; the testimony of others might be lies; vision fails under low-light conditions; our conscience is distorted by own own wicked motives and by the powerful cultural moores which surround us; etc. The analogy between conscience and the sensus divinitatis is worth pressing. Just as our own wicked motives or corrupt social moores distort the moral intuitions emanating from our conscience, so does our sin and alienation from God distort beliefs produced by our sensus divinitatis. The disagreement here merely reflects the fact that we are sinful creatures and require the assistance of God.
But the problem is even more severe than Timmo notes. On the one hand, our memories might be always false; our vision might be nothing but specious illusion. Our sensus divinitatis might be a sensus demonitatis. On the other hand, we might hold that our memories are perfectly reliable, our vision always veridical (i.e. if we cannot see an object, regardless of the light level, it is not actually there), our conscience perfectly reliable. As Quine noted, and Monty Python humorously illustrates we can take any number of statements as axiomatically true "come what may" and adjust the rest of our beliefs around these axioms. You can even "prove" that the mythology and cultural prejudices of Iron-age slave-owning genocidal savages is absolutely, factually and literally true and reconcile the worst of its barbaric ethics with modern Western political reality (so literalists don't go to prison for stoning their disobedient children).

These sorts of axioms — empirical, religious, or even schizophrenic paranoia — cannot be disproved just by an appeal to internal consistency. It's worth noting that, in a sense, even empiricism forms an "intellectual black hole": Once you've "bought into" the premises of empiricism to the exclusion of all others, there's no way for any reasoning on those premises to throw you out of the system.

Empirical rationality does not possess the advantage that it can be somehow "proven" to be reliable (it can be defined without contradiction to be perfectly reliable, completely unreliable or anywhere in between). Its chief advantage is that it is more pragmatically effective at generating agreement between individuals. It generates agreement on two principles. First, it relies upon only generally agreed-upon statements of perceptual fact. The empirical project is not to "believe the truth" but to generate agreement, therefore we embed agreeability into the definition.

Secondly, the veracity of any set of statements, including those of perceptual fact, is evaluated according to Occam's razor: A statement-set is veridical if and only if:
  1. It does not contradict any statement of perceptual fact
  2. it is compatible with all statements of perceptual fact
  3. there is no simpler statement set with features 1 and 2
It is the third qualification is both the greatest strength and greatest weakness of empiricism and evidentiary foundationalism.

Because we assume simplicity as a metaphysical criterion, we are able to generate agreement on what specifically counts as a reliable or unreliable (a.k.a. veridical or mistaken) perceptual experience. If, alone in my house, I remember that my car keys are in my jacket pocket, but they're not there but instead lying on the kitchen counter, I conclude that this instance of my memory is unreliable: It is simpler to hypothesize that my memory is false than to overturn all the laws of physics to hypothesize that my keys moved on their own when I wasn't looking. Note that I am still assuming my memory is authoritative: I don't deny that I really do remember my keys being in my jacket pocket, and my overall theory must account for this memory and why I conclude it isn't veridical.

And (to forestall some objections) there is nothing at all mysterious or non-rigorous about "simplicity": You basically count the irreducible hypothetical premises; the fewer the premises, the simpler the theory. (There are some sophisticated tweaks, but the basic idea of counting premises still applies.)

But the criterion of simplicity is also empiricism's greatest weakness. We cannot ever prove that the theory we have is actually the simplest. We cannot even prove that the simplest theory is actually true.

But that's not problematic. It's a sucker's game to try to talk about what's really true independent of our knowledge: By definition we cannot know anything at all independently of our knowledge. The universe might just be a program running on a mad scientist's giant computer, or just one thought among many in the mind of God. What appears to you, gentle reader, to be external reality — including this article — might be nothing more than the your own fevered imagination.

But so what? Who cares? I'm not interested in playing a sucker's game, where nothing can be known, where knowledge is irrelevant, or defined to be what you want to believe. Such metaphysical fantasies, where the all beliefs are equally "true", can do nothing but generate bullshit, albeit often bullshit of considerable artistry. But, at the end of the day, it's just bullshit.

Our project as human beings is to live and be happy. A big part of that project is not having the perceptual experience of falling off a cliff or being eaten by a tiger, or starving, or stubbing your toe. Pain still hurts, regardless of whether it is a specific neural state, "objectively" indistinguishable from any other neural state, or electrons in a mad scientists computer, or God's thoughts, or one's own solipsistic imagination.


  1. I'm no philostomopher, but isn't the problem with Timmo's argument just the same old, "How do I really know I'm not a brain in a vat in the lab of some evil scientist who has all my nerves hooked up to a computer that is delivering appropriate signals to make it seem like I live in a world"? At the end of the day, there is no satisfying response to that concern, right?

  2. [I]sn't the problem with Timmo's argument just the same old, "How do I really know I'm not a brain in a vat..."

    I'm not sure that this is a problem with specifically Timmo's argument. The brain in the vat problem applies to all forms of axiomatic foundationalism; it's a problem in philosophical epistemology in general.

    The problem with Timmo's argument is that it doesn't really solve anything. It just moves our confusion about "justification" to confusion about "reliable belief" (or, if you move Timmo's hyphen, to confusion about "true belief" and "reliable process).

    The intent of such a maneuver is to exempt belief-formation processes from criticism: It might be the case that a belief-formation process is reliable, and thus beliefs formed by that process justified, without actually being able to prove that the process is reliable. You have to have, dontcha know, a reliable belief-formation process to form reliable beliefs about belief-formation processes.

    If the circularity and self-referentiality above makes your brain hurt, you're in good company.

    There's no way to break the deadlock except by Pragmatism... a Pragmatism that goes considerably beyond Pierce, James and Dewey's formulations. (In much the same way we have to go considerably beyond Popper's formulation of falsifiability to get a working theory of scientific epistemology.)


Please pick a handle or moniker for your comment. It's much easier to address someone by a name or pseudonym than simply "hey you". I have the option of requiring a "hard" identity, but I don't want to turn that on... yet.

With few exceptions, I will not respond or reply to anonymous comments, and I may delete them. I keep a copy of all comments; if you want the text of your comment to repost with something vaguely resembling an identity, email me.

No spam, pr0n, commercial advertising, insanity, lies, repetition or off-topic comments. Creationists, Global Warming deniers, anti-vaxers, Randians, and Libertarians are automatically presumed to be idiots; Christians and Muslims might get the benefit of the doubt, if I'm in a good mood.

See the Debate Flowchart for some basic rules.

Sourced factual corrections are always published and acknowledged.

I will respond or not respond to comments as the mood takes me. See my latest comment policy for details. I am not a pseudonomous-American: my real name is Larry.

Comments may be moderated from time to time. When I do moderate comments, anonymous comments are far more likely to be rejected.

I've already answered some typical comments.

I have jqMath enabled for the blog. If you have a dollar sign (\$) in your comment, put a \\ in front of it: \\\$, unless you want to include a formula in your comment.