…but I can’t really help it. I’ve been interested in Nozick’s Experience Machine objection in Anarchy, State and Utopia since early in grad school. Now David Friedman blogs about the experience machine, with some insightful discussion following, and at Catallarchy, Scott Scheule both comments on Friedman and offers a very thoughtful criticism to my own comment. I started to respond to Scott there, but then realized that, given the dearth of material here recently, I’d write it here instead. Shameless, I know. I don’t really care.
Anyway, here’s the response. First, thanks, Scott, for the really interesting and thoughtful reply. You raise some good points–several of them, actually–and there’s no real way that I can hope to do justice to all of those points, at least not without taking up way too much of the time that my University thinks I am spending writing about just war theory. So I’ll try here to get at what I think might be the heart of the difference between my rejection of the egoist’s use of the term “selfish” and my own use of the term “real.”
My initial response here, I think, is that it’s not so much that I’m trying to redefine or use the word “real” in some strange way as it is that the word “real” is already pretty hopelessly vague, and that it tends to get used in ways that aren’t really rigorous in the first place. “Selfish” by contrast, does have a pretty well-understood meaning. So let me offer some rigorous (or semi-rigorous; it’s only 9 a.m. so I’m not currently up for really rigorous) attempts to articulate what “real” might mean in this particular context.
- An experience that corresponds to actual phenomena in the world.
- An experience that can be shared or verified by others around us.
- An experience that is not produced through direct manipulation of my brain states.
- True and actual; not imaginary, alleged, or ideal.
- Genuine and authentic; not artificial or spurious.
- Free of pretense, falsehood, or affectation.
- Existing objectively in the world regardless of subjectivity or conventions of thought or language.
- Of, relating to, or being an image formed by light rays that converge in space.
Okay, I cheated a bit. Most of those come from here, though I broke the first one into two as they raise, I think, very different points.
Now, though, the fun part. Which of these uses of “real” supports Scott’s (and Nozick’s) intuition that real experiences are to be preferred to experiences in the experience machine? I think that (1) pretty clearly won’t do. After all, as Scott admits in his post, though in a different context there, the experiences that I have inside the experience machine pretty clearly do correspond to phenomena in the actual world. They just don’t correspond to the phenomena that I think they correspond to. Thus (1) is out. (2) doesn’t fare much better; if we just want the experience to be verified, then that will require only that I hook at least two people up to the same experience machine. Experiences verified. (3) I think fails for two different reasons. First, it begs the question entirely. And second, it would rule out things like getting drunk (which, after all, is an experience brought about entirely by a direct manipulation of your brain) as real experiences. (8) is equally easy to reject; as Scott mentions, the experience machine could just as easily be a holodeck, making the creations very much images that are formed by converging light rays.
Number (4) seems a bit more troubling at first glance, but I think that a deeper look reveals that that definition is superficial. It defines “real” by offering an equally poorly understood synonym (“actual”). The clarification, though, reveals that there is little to help out those wanting to reject the experience machine. Experiences contained therein are not imaginary (the experiences are actually happening inside my head, and I am in no way in control of their happening), alleged (they really are right there) or ideal (no one says that the experiences have to be perfect in any way).
Of the eight definitions, 5, 6, and 7 present the most difficulty. I think that one can, at the end of the day, reject all of them as displaying a bias for the organic and the natural that doesn’t really seem to be all that justified. If we’re talking about the experiences themselves, then what difference, really, does it make how that experience is generated? There isn’t, I think, an in principle distinction. Think here about certain kinds of experiences.
Suppose, for example, that I were to decide to obtain the services of a prostitute. And let us further suppose that I have two possible avenues. The first is a flesh-and-blood prostitute; the second is a virtual prostitute. Now in each case, I will end up having a certain kind of experience, namely that of having sex with a willing but (really under it all) entirely uninterested partner. Sex with the flesh-and-blood prostitute will be a “real” experience (let’s be clear: it will be a real experience of sex-with-a-prostitute) in the sense of being a genuine and authentic and not spurious or artifical experience of sex-with-a-prostitute. But why, exactly, is that a better experience? Both experiences (i.e., the virtual and the flesh-and-blood prostitutes) are already substitutes for a different experience, namely, sex with a willing and interested partner.
It seems to me that the only difference between the two encounters lies in the genesis of the experience. And I can’t see any good reasons for thinking that that the way that the experience gets generated is all that relevant, at least not in this particular sort of case. Virtual prostitute sex and real prostitute sex are pretty much the same experience. What that shows, I think, is that it is not the way that the experience gets generated that really drives our intuitions about experience machine examples. What we’re concerned about is that the experiences inside an experience machine will lack meaningfulness. Why? Well, because when both of the experiences in question already lack any real meaningfulness, we don’t really care which way they are generated (and may indeed have other reasons for preferring the artificial–virtual hookers rarely give one STDs).
The problem, then, I think, is that people reject the experience machine because they think that it fails to be meaningful, that unless their actions somehow connect up to other people’s lives in the right sorts of ways, the experience will have no value. That objection to the experience machine really boils back down to a version of (2), though. We can fix it by simply plugging lots of people into the same experience machine. Now my experiences do connect up to the lives of others. So what’s the problem?
At the end of the day, I don’t think that I’m distorting the word “real” to mean things other than it usually means. Rather, my claim is that ordinary understandings of the word “real” don’t actually support the Nozickian intuition about the experience machine. This is an instance of intuitions that aren’t really designed to cope with the sort of example that he offers. We have no real grasp of an illusion that, ex hypothesi, does not ever go away.
Let me put the point another way. We’ve all heard (way back in intro to philosophy) Cartesian arguments for skepticism–you know, the brain in a vat, evil demon sorts of things. The argument here is that, for all we know, the world we’re actually living in now is already a virtual world. We simply have no way to tell. Well, that’s what the experience machine would be like. It would be exactly like the world that we’re in right now. Only, again ex hypothesi, everything would go way better for us. I really just can’t see what the problem is.