In the comments to a recent post, Tim argued that agnosticism should be defined according to its provenance in T.H. Huxley. By agnosticism, Huxley meant the claim that no one can know if God exists because God is inherently unknowable. I argued in the comments that the term agnosticism is much more diverse than this: it can be personal ("I don't know") or universal ("no one knows"), it can be diffident ("we don't know") or assertive ("we can't know"), it can relate to the object ("it can't be known") or the subject ("we're incapable of knowing"). Tim argued to the contrary that while agnosticism has taken on these colloquial meanings, its technical meaning is Huxley's definition, and we should stick with the technical definition because appealing to colloquial definitions is a recipe for disaster.
The problem with this is that Tim is wrong. I'm giving the technical definition of agnosticism. Tim thinks this is due to creeping colloquialism but he's simply incorrect. Here's why: agnosticism is used by philosophers to denote the withholding of belief in any proposition. If you are unpersuaded that there is life elsewhere in the universe but don't actively disbelieve it, you are agnostic about it, you withhold belief. If you think the evidence for string theory is lacking but think it is very possible that forthcoming evidence may shore up the gaps you are agnostic about it. This may also be how it is used colloquially, but that doesn't mean it's not also the technical definition.
Can we say that agnosticism means one thing when it bears on knowledge in general but something else when it bears on knowledge of God? Well we could but that's not how it's used by philosophers (not to mention the fact that using any term in such a way would be a recipe for disaster if anything would be). For example, agnosticism about God's existence can refer to an individual's claim to not know one way or the other or to the universal claim that no one really knows. I've heard this distinction referred to as subjective/objective, personal/universal, individual/general, but regardless this is a real distinction made by philosophers regarding agnosticism about God's existence. (In a debate William Lane Craig once called it ordinary agnosticism vs. ornery agnosticism.) To insist that the technical use of this term should be abandoned in favor of how it was originally conceived is incorrect. Of course we shouldn't read our definition of agnosticism back into Huxley's writings, it just means that regardless of how a term originated, its technical definition is determined by how it is used in technical discussions. I don't know if this is considered a linguistic fallacy -- treating a term's original meaning as its technical definition -- but it strikes me as one. It's a close cousin to the root fallacy where one defines the meaning of a term according to the parts that make it up as well as the fallacy of semantic obsolescence. Perhaps we can call it the fallacy of provenance.
Part of the problem is that Tim tried to define knowledge in an absolute way: "No-one can definitively know if something exists or not, unless they are omniscient." I presume his reason for this is that unless you know everything the possibility remains that one of the items you don't know would invalidate your claim to knowledge. But epistemologists know that knowledge doesn't work that way. The classical definition of knowledge is justified true belief but this has fallen on hard times in the last several decades. Regardless, whatever one wants to call the third condition of knowledge (or fourth condition if one wants to add something to justification) very few philosophers, if any, maintain that it requires the absolute assurance that only omniscience could provide. (Maybe Peter Unger would accept this, but I haven't really read his defense of skepticism.) Many epistemologists do make one of the conditions of knowledge be that one have a reasonably thorough knowledge of the object known in order for it to qualify as knowledge, but nothing like omniscience. Infallibilism is the view that knowledge is ... wait for it ... infallible and so would be in that general direction, but infallibilists don't make omniscience a condition of infallibility; far from it. In fact, as far as I can tell, the majority view among epistemologists today is fallibilism: knowledge doesn't have to be absolute or certain or anything like it. As long as one believes the proposition, even weakly, and the right conditions are met (and the belief is true) then it qualifies as knowledge.
The whole discussion was raised by the issue of burden of proof. Many atheists claim that only the one who is making a positive claim bears the burden of proof, therefore, they do not have to have any evidence or reason or grounds for their atheism. This doesn't really work since the claim that God does not exist is the claim to know something, just as much as the claim that he does. Disbelief is the belief that a proposition is false. Some atheists respond by saying "You can't prove a negative" which was the topic of the post (quick answer: of course you can). So the atheist has to shoulder his share of the burden of proof. Some atheists respond further by redefining atheism to mean the absence of belief in God rather than disbelief. Disbelief is hard atheism, lacking a belief is soft atheism. I've written about that before too: I understand what belief means; I understand what disbelief means. I also understand what the withholding of belief (i.e. agnosticism) means. Further, I understand what it would mean to have no conception of something: prior to hearing about Russell's orbiting teapot or the flying spaghetti monster I had no conception of them. I could see calling this last case "lacking a belief," but this doesn't help the soft atheist since he has obviously heard of the concept of God. Once I've heard of a concept I no longer lack a belief in it: I believe, disbelieve, or withhold belief in it. So I asked Tim, as I've asked others, to define what lacking a belief in God means and how it differs from belief (theism), disbelief (atheism), and withholding belief (agnosticism). Tim, much to his credit, offered the first response to this I've ever received: he said it seemed to be closer to withholding belief. The problem is that this is not atheism, it is agnosticism. Hence the discussion. I kind of suspect that soft atheists disbelieve in God, but weakly; they're right on the border of being agnostic but leaning towards disbelief. But they insist that's not what they mean, they mean they lack a belief -- a concept I can't get them to define.
The believer must shoulder the burden of proof, but so must the disbeliever, the one who believes that the proposition up for grabs is false. I, for example, disbelieve in Russell's orbiting teapot and the flying spaghetti monster. I must therefore shoulder the burden of proof, I must supply a justification or reason for my belief that these proposals are false. I've written about this before too: The reason why I don't believe these claims -- and the reason why everyone else doesn't believe them either -- is because they are completely ad hoc. The more ad hoc, or contrived, a claim is, the less likely it is true. As I wrote before, "The degree to which it is ad hoc is the degree to which it is implausible. This is particularly evident with the absurdly ad hoc propositions mentioned above: we react against such suggestions because they are completely contrived. It's not merely that we have no reason to think they are true; we think, for whatever reason, that they are just "made up," and this is a specific reason to think they are not true." Of course, these ideas were originally conceived in order to try to compare them to belief in God, and to show that, just as the rational response to the orbiting teapot and spaghetti monster is to disbelieve them until we have evidence for them, so the rational response to God is to disbelieve until we have evidence. Obviously this attempt fails: the concept of God is not inherently ad hoc, it has been present in all cultures in all times. Of course the concept of God can be used in an ad hoc way, as can other concepts (Marxists, for example, use economics in an ad hoc way), but this does not mean that the concept itself is ad hoc.
Only the withholder of belief, the agnostic, does not have to bear any burden of proof, because he is not believing or asserting anything (and of course the one who has never heard of the concept does not bear the burden of proof for it). Soft atheism strikes me as an attempt to get the negative position of the atheist and the burdenfree position of the agnostic. It looks like an attempt to disbelieve in God without having to go through all the rigmarole of having any burden of proof placed upon one's shoulders. Until I hear a definition of lacking a belief that is distinct from disbelief and withholding belief, I see no reason to question this impression.
Let me just conclude by reiterating something I posted on Agent Intellect: "Agnostic" is the Greek-based term for someone with no knowledge. The Latin-based term is "ignoramus."
Discuss this post at the Quodlibeta Forum