Wednesday, August 30, 2006

Opacity and Anti-Realism

Something I probably should have thought about before, but in fact only noticed in conversation with Daniel Stoljar on Friday, is that (by my lights) '... is an anti-realist about ...' can create opaque contexts, on at least some uses. I reckon that one way to to be an anti-realist about ethics is to say that the ethical facts are identical with certain facts about what we approve and disapprove of. But anti-realists about ethics needn't (even by their own lights - i.e. even given that their identity claim is right) be anti-realists about approval.

4 comments:

Anonymous said...

"Anti-realist" is, pretty uncontroversially, a term with many meanings. Whether reductionism is a kind of realism or a kind of anti-realism depends on who you're talking to. This is true even in ethics; I have encountered a form of straight up subjectivism which was called by its advocate "dispositional ethical realism."

The case of causation may be instructive. Some, e.g. Tooley, seem to think you don't believe in causes unless you take causes as primitives. Others think that reductionism which makes causes involve the action of primitive laws of nature is a form of causal realism, while reduction of causes to regularities is anti-realism. Still others claim that even regularity theories are realists, leaving only those who are anti-realists about everything to be anti-realists about causes.

I think the way it works is that people have intuitions of various kinds and various strengths concerning what something must be like in order to be a proper X. Taking it as primitive that there are proper Xs is always realist, and denying that there's anything like Xs whatever is always anti-realist. Whether reducting Xs to Ys makes you a realist about Xs depends on whether you think Ys are enough like Xs to play the X role, which of course depends on what your intuitions about Xs are, and how high your standards of being close enough are.

Whether you view a particular form of reductionism as a form of realism is not necessarily determined by whether you accept that form of reductionism. Mackie and Hume are both, roughly, subjectivists, but Mackie clearly considers subjectivism a form of anti-realism, while Hume is probably best interpreted as considering it a form of realism.

Carrie Jenkins said...

Hi Andrew,

Well, I don't want to saddle anti-realists with having to deny that the things they are anti-realists about 'really exist', though. At least, I think there are uses of the term 'anti-realist' about which are perfectly consistent with thinking the things in question are real: mind-dependence anti-realists about mathematics, for instance, don't necessarily have to deny that numbers/mathematical facts etc. really exist, they can just say that they are (e.g.) mental contructions.

Carrie Jenkins said...

I see - he intends to define up a technical sense for 'really exists'. In that case, does Fine's view do anything for us over and above saying that the anti-realist doesn't think numbers are 'fundamental'?

(Incidentally, I think I'd want to allow that someone could think that the mind and all its constructions are totally fundamental and still be a mind dependence anti-realist about numbers because he thinks numbers are mental constructions.)

Anonymous said...

Carrie,

I'm not sure how Stoljar's opacity claim is supposed to connect up with the claim about approval. I suppose that we can construe the approval claim as something along the following lines: whether something is ethical depends on whether it's approved (thus the ethical anti-realism) but whether something is approved doesn't itself depend on being approved, i.e. the approval requires no additional approval (thus the realism about approval). Consider, by analogy: whether something is red may be response dependent, but whether being red is response dependent is not itself response dependent.

What I take these kinds of remarks to suggest is that the compatibility of ethical anti-realism and approval realism is due to differences in various relata (e.g. whether something is approved vs. whether something's being approved is itself approved) and not whether the contexts are opaque. This is not to deny the opacity, but only to suggest that it's not particularly relevant here.