Nozick’s experience machine

David Friedman writes:

Robert Nozick, in Anarchy, State and Utopia, offers an interesting hypothetical; I don’t have a copy of the book ready to hand so will give you my version:

Someone invents an experience machine; get into it and you will have a fully convincing illusion of experience. Somehow, the inventor figures out about what your life is going to be like and makes you the following offer:

Get into my experience machine, spend the rest of your life there, and I will give you the illusion of a life slightly better than the one you would otherwise live. Your average income in the illusion will be a few thousand dollars higher than it would have been in reality, your wife a little prettier, your children slightly better behaved, your promotions just a little prompter. Your illusory summers won’t be quite as hot, or winters quite as cold.

As I (and Wikipedia) remember the thought experiment, your life [sic] is much better, a bit like the Hollywood movie of your choice, at the very least.

This thought question is supposed to refute utilitarianism, or at least its hedonistic version.  Pleasure isn’t everything, and authenticity counts too.

But for an economist, Friedman is oddly non-interested in the marginal questions.  At what age should you opt for the experience machine?  If it can give you eighty more years of subjectively perceived time on your deathbed, that is a no-brainer.  At what per capita income level should you prefer the machine?

Or say you don’t want the machine, but your acceptance will save five other people’s lives.  Would you proceed without guilt?  And how many lives should be needed to push you over the edge?

Is the experience machine example so compelling as a refutation of hedonism?  I think it puts pleasure squarely on the map as one value which matters and which is even undervalued in many circumstances.  Many of us are too reluctant to step into the machine rather than too ready.  Isn’t our general tendency to overvalue the illusion of control? 


Tyler's last paragraph is rather odd.

On the one hand, he recognizes that the most people would have an aversion to using the experience machine. On the other hand, that aversion flies in the face of the logic of hedonistic utilitiarianism, and hedonistic utilitiarianism is such a powerful tool for economic
theory that he doesn't want to give it up. So in a rather pathetic attempt to avoid the logic of Noziak's argument, tries to argue that people have the wrong preferences. They should be hedonistic utilitarians, even it they aren't. (Because, turning empericism on its head, then our theoretical apparatus would work better.)

Tyler, don't we endeavor to beat this kind of thinking out of our students in the first semester?

Since Dr. Cowen is already inside the machine (it is called The Mind) it is the discussion about the existence of a choice that is the illusion.

One version of the machine is called heroin.

The original thought experiment is that the experience machine offers you the best possible set of experiences for your personal preferences. So if you need a little adversity to overcome, it gives you that. If you prefer to have to work to win your mate, then you experience the effort.

Nozick's point is not that we prefer other things over pleasure all the time. Rather, pleasure is just one of many values. Most people would prefer to have an average but real life to an experientially better but fake life. We prefer not just to have the experience of being witty or smart or morally good, but to actually be witty, smart, and morally good. We would prefer not just having the experience of friendship but to actually have friends. Nozick's thought experiment is broader than an attack on hedoism. It shows, if successful, that we value things other than experiential states. We have a plurality of values that experiental theories of value fail to do justice to. The experience machine doesn't refute, say, preference utilitarianism, because preference utilitarianism (crudely put) holds that the best state of affairs consists in the maximum satisfaction of all preferencess. The experience machine doesn't satisfy all preferences.

And, of course, as Nozick recognized, pleasure is a value. In some cases, like when one is about to be tossed in a gulag, the experience machine is a good deal. Most of us prefer our average but real lives to the experience machine life, but that doesn't mean we prefer real torture followed by real death to the experience machine.

That makes me think--Suppose one wanted to become a martyr for a cause. The experience machine could give you the experience of being persecuted, tortured, and killed, but you'd be mad to learn that it's fake and you haven't actually died for anything.

A good point to watch Vanilla Sky / Abre los Ojos, I think.

Another question you might ask is this. If you were in the machine and came to know it, would you want to leave it? Tyler's questions about the margin might apply here too. So, you're in the machine in which you live a vibrant life as a 30 year old double Nobel Prize recipient in Economics and Literature, but you find out that, were you to unplug, you would be a 60 year old work-a-daddy suffering from loneliness, moderately low income and mild bout of diphtheria. Would you leave?

I think that it's absolutely clear in this case that in the "reversal test" version we would be foolish to leave.

To my mind the biggest downside is the possibility of mechanical failure. If the machine breaks, especially after a long time, I don't return to my real life in progress, I return to a world I haven't even been truly inhabiting. My family, friends, and worldly possessions are for all intents and purposes gone.

So to me, it's more of a perception-of-small-probabilities-of-horrific-events quandary.

Hi there, Alex.

The experience machine discussion remind me of another discussion about how to judge people's welfare posthumously (I believe Parfit wrote something about this). One argument advances a non-experiential conception of welfare, from which it follows that a person's life can become worse after he is dead (if, say, his children become dropouts or drug addicts). This argument always seemed fairly untenable to me.

Authenticity seems to be valuable only as an ex post realization. Our ex ante judgments about authenticity always refer to our expectations about ex post feelings. (We would be angry if we found out we didn't really win a Nobel prize, etc.). If there were no chance that this discovery would happen, then, it seems almost indisputable that we ought to plug in.

One more argument - if it could be proven that the ex ante disutility from knowing you were giving up reality outweighed the positive utility from a great virtual life, then it could be argued that the experience machine is a bad choice. I doubt this is true for most people, though.

the premise of the Matrix movies comes to mind; Joe Pantoliano's character in the first movie (dinner scene) and in response, the indian family in the third movie

Start with these two questions:

Why would anyone bother to write a will?
Why on earth do we bother to honor them?

The answer is intuitively obvious- we care about what happens after we're gone, particularly to the people we care about. We form the habit of honoring wills because it's necessary for everyone to believe that their own will will be honored.

What's the relevance to Nozick's machine? Once you enter the machine, you've foreclosed all possibility of leaving anything lasting behind you. This is why the gut-reaction to this choice is that it would be "empty" or "meaningless". Wanting to leave a lasting mark on the world (for most people the most lasting thing they leave is their children), and the desire to be remembered for something are powerful emotions, for which people have often risked their lives.

Sure, if the machine made you forget the real world, then you're sort in Matrix-land (someone mentioned Joe Pantoliano's character- remmeber his vehement insistence that he have no recollection of his decision). If someone knocked you out and you woke up in one, well then you'd have no way of knowing the difference and the decision is moot. However, when you're faced with the choice, you have choice A: much increased pleasure, but no chance whatsoever of leaving anything behind from this point on, or B: take your chances on the real world, with the expectation of lower pleasure and the possibility (but no guarantee) of leaving something behind or being remembered for something. Most people, we assume, would take B (Everquest and World of Warcraft are popular, but not that popular, and anyway work on the conceit that in someway, everyone else there is a real person). It could be rational for a number of people- what if you're old and childless, say? But if you have the bulk of your life in front of you, you'd probably rather take your chances with reality.

Should we page Robert Barro about inheritance choices?

You're assuming that "Tyler Cowen's life" is not already a single "thread" within some kind of "experience machine". Extended self-inquiry (often derided as "navel-gazing" by those who wish to avoid it) might lead to a different conclusion.

Comments for this post are closed