Some Of Our Books


« Gatekeepers to the Profession | Main | A dilemma for interrogational torture? »

February 23, 2009


Feed You can follow this conversation by subscribing to the comment feed for this post.

Here's a possibility.
A lot of people nowadays are pretty familiar with virtual reality. Suppose they interpret the "illusory life" they've been having (in the story) not as a kind of movie they were watching, but as a kind of virtual world they've been interacting with (a la the Matrix, maybe).
In that case, maybe the subjects do not draw a big distinction between the 'virtual' world and the 'real' world. What they care about is the people they're familiar with. If so, the difference is unsurprising: people don't want to be transferred to an alien world where they'd lose touch with the people and places they love.

And before De Brigard, there was a fascinating set of similar 'live-action' thought experiments known widely to the public as Vanilla Sky and The Matrix....ok, that was tongue in cheek; i actually found the article clear and insightful. Some questions though: first, De Brigard's findings do not decapitate Nozick's main point, namely, that life's value is measured in something other than hedons--so it's perahps besides the point to find that this something might not be being in touch with reality. Second, De Brigard admits that his 'status quo' suggestion is just that, one among many potential explanations: but still, it seems unlikely to me. Or at least it can't be JUST that it's the status quo--people's projects and concerns and values are all tied up in this world, this is the base world in the person has become a self, and so it would be difficult to give this up.

An potentially interesting follow up: I wonder what the correlation would be among those who chose to stay in the machine vs. those who chose to leave indexed to those who report satisfaction with their current or likely future life vs. those who are dissatisfied with their current or likely future life.

This is an intriguing idea!

Prof. Knobe writes:

"he found that most subjects who had been given Nozick's case said that they did not want to plug in to the machine... but, interestingly enough, he did not also find that most subjects in the inverted version wanted to unplug. Instead, many of those subjects wanted to hold on to their present, purely illusory lives."

This struck me, because of the absence of any modal terms or operators in the last sentence. But, the question is asked of people in the real world what would happen in some other possible world, a possible world which is not particularly near to this world. Nobody in this world wants to "hold on to their present, purely illusory lives" because no one has one of those. The original experience machine argument asks whether or not one, exactly as one is now, would want to plug in. This asks one to stretch one's imagination a bit, but not nearly as much as the inverted experience machine. Perhaps De Brigard's findings suggest that people cannot usually stretch their modal intutions/decision-making so far. First the subjects have to imagine that their lives are illusory. Seems doable but already a bit strange since their lives are all they know so far. Second they have to imagine they are informed of the possibility of unplugging and returning to the real world. Things become more strained. They have to imagine that they are really someone else than they thought were - this on top of already imagining their lives were illusory. Next, they are asked to deliberate about whether or not they would trade in their imaginary illusory lives in for the imaginary real lives. This seems like a difficult task, particularly when one takes seriously the relationship between imagined altenatives and decision-making. I admit the results are interesting, but I wonder if they say more about the way people generate such judgments than anything about the nature of well-being or about biases with regard to peoples' preferences.

Interesting post.

Two thoughts. First, I'm not sure I like the way the "positive scenario" was described in the experiment. As has been noted here, the reaction most people give to the original EM (at least in my experience) is that the EM is missing something concrete about reality, that is, genuine achievements, relationships, etc. The only thing that the positive scenario suggests is that you are a "millionaire artist". But I suspect that if, say, you had been described as a person with loads of relationships, successful achievements, etc., people's reactions might be different--they would choose to plug out, as it were. I might be all washed up on this, of course.

A second point concerns the relevance of this for Nozick's argument. I think this depends on how you interpret the EM in an argument against hedonism. If the claim is: "people don't desire to enter the experience machine, hence hedonism is false," this move doesn't work at all against an "objective" version of hedonism, which says that the satisfaction of people's desires is neither here nor there. People may not desire to enter the machine, but people's desires are poor guides to the good; this desire is simply not prudentially trustworthy. But I don't think that Nozick's argument should be interpreted in this way. It seems to me his argument shows that psychological hedonism is false, viz., the claim that we only have intrinsic desires for pleasure, which can function in an argument for hedonism of the sort that marries hedonism to a desire-satisfaction view (think Brandt, Sidgwick, various others). But if this is the argument, then I'm not sure that the inverse EM causes any problems for it. It's still true that people don't desire all and only pleasure. Turns out they desire the status quo, as well.

Sort of like Jamie's point, I care about the people I love whatever they are made of. If you tell me they are illusary and that there are (unknown to me now) real people I really love, is it a huge surprise I'd stick with the ones I know and care about?

It seems to me that so far (and I haven't read the actual paper) I'm not sure what I am being asked to believe when I accept that my experience is an illusion. If I really wind up believing that the people I know don't exist, then (given that there is no way to join them since they aren't anywhere at all) I might as well join the actually living. OTOH, if all I accept is a theory on which the people I know aren't real (of the sort various philosophers sometimes believe) I might still prefer to stay in touch with those I care about.

Interesting paper.

I suspect he would also get similar results if he asked people about switching from their current life to a different life if 1) both lives were real, or 2) both lives were virtual. That would further support the claim that people's intuitions are not primarily driven by an attachment to reality.

This is a very interesting idea and it does seem like the status quo bias could explain the effect Felipe got in his experimental study.
Another possible explanation is that in both Nozick's original experiment and Felipe De Brigard's new experiment, participants were simply demonstrating risk aversion. That is, there seem to be very serious risks involved in both connecting yourself to the machine and disconnecting yourself from the machine, i.e., you don't actually know what your life will be like once you make the switch.
One way to test these two hypotheses against each other would be to imagine a situation in which the risk averse behavior would actually be to accept being disconnected from a machine. So, for instance, if instead of having a normal life, you were an illegal gambler continuously escaping from law enforcement officers and debt collectors, would you then be willing to unplug yourself from a machine and resume your real life which was much less risky, even if equally dishonorable? What do other readers think?
One further possible way to test between the two hypotheses would be to have a more fully described actual life to resume once you unplug yourself from the machine. It seems that if the two lives were exactly the same, then the status quo bias really would show itself. However, I wonder if in the original study, risk aversion would be a better explanation than the status quo bias.

I am tempted, initially, by the kind of explanation Jamie and Mark offer—that there is a desire for the continuation of my current relationships. And I suspect, along these lines, that Dale is wrong that offering a "fuller" life-description would entice most people to unplug. I do feel that such descriptions would add some force, but it seems to me that force stems from the initial excitement of having a full, new life on offer—an excitement that would wear off once I remembered what I was about to lose.

That being said, just what is it I am so worried about losing? If what I valued were my experiences of my relationships, then I should be willing to plug into an EM that has perfect physical and psychological duplicates of my friends and family with whom I can continue to relate. But I wouldn't be. So it is tempting to say that it is the relationships themselves that I value. But in the inverted EM, all there is to my relationships is my experiences of them, because the people I am relating to don't actually exist.

Mark, you write: "I care about the people I love whatever they are made of." Perhaps this hints at the answer—that I value the entirety of my relationships, whatever they turn out to be. So, if I'm not in an EM, I won't plug in because I want to keep the full, real relationships I have. But if I am in an EM, I still want to keep what I have, even if it's an illusion. Here, I wrote about de dicto victims in response to Doug's worries about Carl's future child. Could it be that what I value here—or even what's actually of value—are de dicto relationships?

Thanks for the interesting post, Josh. A few brief comments/questions:

1) Leaving the experience machine would produce the experience of a sudden and wrenching disconnect with one's virtual relationships, whereas entering Nozick's experience machine need have no similar effect. My actual desires include desires for the wellbeing of my loved ones, and I would look forward to these desires seeming satisfied if I entered the experience machine. I imagine that my plugged in life would develop organically from my present one, only better, in Nozick's machine. So I don't really see how status quo bias is supposed to dissuade me from plugging in in the same way as it very clearly could dissuade me from unplugging.

2) Surely the question De Brigard asked should not have been (a) "Would you unplug?", but rather (b) "Would your life, on having learned this information, seem to you suddenly less valuable in any way?". If De Brigard has an interesting philosophical point to make, it must surely be not just that status quo bias exists, but that it is status quo bias rather than the way in which contact with reality matters to us that explains our reluctance to plug in. There's no evidence of this, though, in the absence of a negative answer to (b). But I don't think anyone sufficiently reflective will give a negative answer to (b). Clearly, one's life would seem less valuable if one's "relationships" were shown to be all along illusory.

3) What does the method of experimental philosophy actually lend to De Brigard's argument? There don't seem to be any surprising findings about differences in intuitions between groups here. Rather, we have a number of thought experiments each yielding fairly uniform reflective intuitions, and we want to know what the best explanation of these intuitions and their similarities and differences between experiments is. Couldn't this investigation be just as well conducted in the traditional manner, by the philosopher and her peers, rather than by survey?

As far as I know, Nozick's argument was mainly targetted against hedonistic accounts of well-being. As you say:

"Nozick therefore infers that we not only value having positive subjective experiences [but also having a life that is in reality a good one]".

Here we seem to have two claims:
1)not only positive subjective experiences have value
2) the part I put into brackets: that also having a life that is in reality a good one.

1) seemed to be more important to Nozick than 2) if I remember this right.

As far as I can see, Felipe's results support 1). As you write:

"Perhaps they reflect, at least in large part, a desire we have to maintain the staus quo -- a desire just to leave everything more or less as it was."

If the last conclusion is right, then 1) above must be false. If Status Quo matters, then not only subjective experiences matter. And, so we have more support for Nozick's original thesis rather than argument against it.

Also, I wonder how much changing the story can affect intuitions. What if you were told that in reality your parents, sisters and friends are missing you badly, your wife needs your help to recover from serious illness and so on. If that was the case, would you want to unplug? I think you would.


You write: "Would you choose to plug in? Many people say no, and Nozick therefore infers that we not only value having positive subjective experiences but also having a life that is in reality a good one."

Of course, although many say no, some would and should say yes, for, clearly, some people's reality (including their subjective experiences) are so bad as to make plugging in the prudent thing to do. Indeed, if my life were really bad in reality, I would choose to plug in. So, in this experiment, I'm told that I chose to plug in. Shouldn't I, then, infer that things on the outside are really bad for me? Of course, I'm told that I'm a millionaire from Monaco, but we know that money can't buy you love, friends, or happiness. So if things were bad enough for me to want to plug in in the first place, I must have had few to no friends, no lover, no family, no achievement, and no prospect of this changing. So might the decision not to unplug in the inverse experience machine reflect the fear of how bad things must be in reality to have wanted to plug in the first place despite being a millionaire from Monaco? So what does the author say about this fear of the unknown? In the experiment, do I know what life on the outside will be like in the inverse EM scenario as well as I know what life inside the EM will be like in the regular EM scenario?

This actually happened to me once. Man, was that guy high.

In my article "Mental Statism and the Experience Machine," (see below), I argued precisely that the inverted experience machine thought experiment casts doubt on the strength of Nozick's original example:

(I didn't have any experimental data, though.)

Have to say that the data is pretty undecisive. If your life outside the machine was bad most people would not want to unplug. If it was like your life in the machine or the Monaco scenario (why is that good?), then roughly half of people would want to unplug and the other half would not. I cannot see how you could conclude on this basis more than something like 50% of people value something to do with the status quo and 50% people value something like a contact with reality. So people disagree about what is good...


You write: "in this experiment, I'm told that I chose to plug in. Shouldn't I, then, infer that things on the outside are really bad for me?"

In the story, participants are told that they are plugged to the machine due to a "terrible mistake" (p. 7; cf. pp. 8, 11); their being connected is not the result of a past voluntary decision on their part. Accordingly, they are not entitled to draw any inferences about how their lives were in the real world apart from what they are told explicitly.

Chalmers argued somewhere, in considering the Matrix, that people's beliefs in such an experience machine (one they had been there a while) would largely be true and that there was no good reason to prefer interacting with what we would normally call reality rather than continue interacting with the computer generated "reality". I don't recall arguments for that last part, but it might be worth checking.

Chalmers develops the argument in "The Matrix as Metaphysics", in Christopher Grau (ed.), Philosophers Explore the Matrix, Oxford, 2005, pp. 132-176. His point, roughly, is that the hypothesis that we live in a matrix/EP is a a metaphysical rather than an epistemological hypothesis, since it is concerned with the ultimate or fundamental nature of reality.

Hi everyone:
Thanks for all those comments! I don't have a lot of time right now, so I can't reply as carefully as I want to, but let me say two things. Many have pointed out that my experiments do not speak directly against Nozick's original view. That is quite right. According to his original formulation, all he needed was some people not plugging in, not that everyone would decide against plugging in. His point was, after all, against hedonism, and my point has nothing to do with hedonism. It has to do with the way some people take this alleged intuition as evidence that people value reality. What I am saying is that there is another possible psychological cause for their preference to remain unplugged, viz their reluctance to relinquish they life they have had so far, regardless of whether it is real or not. The thought experiment may not have been initially intended to make this point, but some take it to be. And I don't think it is a good thought experiment for that purpose--I also don't think it is a good experiment for the purpose of rejecting hedonism either, but that's another story. I just don't think it is a good thought experiment at all.
The second point has to do with a possible follow up mentioned by Michael at the beginning: finding out what would happen with these sorts of scenarios with depressed people, say, or even with people that for whatever reason score low in subjective scales of well-being before taking the test. I thought of doing that originally, but I have never gotten around to actually test this vignettes with specific populations. It is certainly a fine follow up. Okay, thanks for all those comments. More later.


Thanks. I hadn't looked at the paper and was relying solely on Josh's description. I have now looked at the three vignettes.


I think that the experiment might have been better had you not specified in the Positive Scenario that one is, in reality, an artist. It's hard for me to imagine being an artist and still being the same person. Being a philosopher is so central to my identity that I might prefer to be a virtual philosopher to being a real artist, and not because I have some status quo bias. By contrast, I can imagine being the same person with the same self-conception while being in a maximum security prison. I just have to imagine that I've been wrongly convicted or imprisoned for, say, corrupting the youth. It would be interesting, I think, to run the experiment again but this time without specifying anything substantive about the nature of one's life in the negative and positive scenarios. For instance, I think that it would have been better had you said in the Positive Scenario that your subjective experiences would be on the whole twice as good as they are now. And then in the negative scenario you could say that your subjective experiences would be on the whole half as good as they are now.

"Here's a possibility.
A lot of people nowadays are pretty familiar with virtual reality. Suppose they interpret the "illusory life" they've been having (in the story) not as a kind of movie they were watching, but as a kind of virtual world they've been interacting with (a la the Matrix, maybe)."

I would add to that advanced gaming and role-playing in which people really DO live and interact with other 'people' in virtual lives.

My guess, and it's just a guess, following Jamie, is that more gamers would be amenable to the experience machine than non-gamers.

I freely admit that I don't play any games of the sort, but I judging this on the people I know who do game and seem to get overly wrapped up (and are often unwilling to unplug) in the virtual world to the detriment of the real world.

Thanks for all these helpful comments! It would definitely be interesting to pursue this issue further.

A number of readers have suggested that the experiment might not raise any serious philosophical questions, since it is still clear that people are valuing something other than subjective experience, namely, maintaining the status quo.

It seems to me, though, that there really is a difficult question here about the normative significance of the intuitions people are offering. It might generally be true that the fact that people value something gives us reason to suppose that the thing itself is valuable, but in this particular case, many researchers would conclude that the values driving people's intuitions are simply the result of a bias.

So, for example, if researchers randomly assign subjects to receive either a mug or some chocolates, subjects usually end up concluding that whichever item they received was in fact more valuable, and they are reluctant to switch for the item they ended up not receiving. Most researchers see this as an irrational tendency -- what has come to be known as *status quo bias*.

Now, De Brigard's study certainly does reveal that people's intuitions about the experience machine involve valuing something other than pleasure, but one might legitimately wonder whether we should continue putting our faith in these intuitions, once we see what they are really about. If we find that our intuition in these cases is driven in part by a valuing of the status quo, should we conclude that this intuition is not in fact tracking something of real normative importance?

(As I consider this question myself, I find that I am torn in conflicting directions and don't know quite what to say.)

If we find that our intuition in these cases is driven in part by a valuing of the status quo, should we conclude that this intuition is not in fact tracking something of real normative importance?

Isn't this what friendship is about? I value the people who are actually my friends more than the people who I could possibly have a relationship with. If I didn't they would not be my friends. I guess you could call this bias, but in that sense a claim of bias would be purged of all negative connotations, no? Or is it too late for me to be thinking about these issues so that I'm missing something obvious? (I don't mean that ironically - it is a hedge in case I really am too tired to see something obvious.)


I'm not sure about that in the dialectic. Nozick's original argument was in part an ad hominem; look hedonists - even you value something other than pleasure. Whatever other than pleasure people, hedonist philosophers included, intuitively value is evidential support for that claim. Now, of course, the hedonist could reply that what people value is distinct from what is valuable. But this seems to something the non-hedonist will welcome. In that case, it is much easier for her to argue that other things than pleasure count. Much of the support for hedonism as a view of well-being has after all traditional come from psychological forms of hedonism.

Also, if there is that sort of divide between what people value and what is really of value, then doesn't that spell trouble for Felipe's experimental method? Anything he ever could have found was what people valued. If that's not even good evidence for what is valuable, then experimental ethics seems to be a not a good project.

Just a quick note in response to these last comments. There does seem to be deep value in preserving relationships that we truly have with other people (just as Mark points out), but the results from the study call into question the claim that this sort of deep value is what is behind people's intuitions in Nozick's original experience machine case. After all, the results show that if we change the story so that people are supposed to imagine that they don't have any true relationships with other human beings in their present life, they *still* want to keep everything just as it is. This result suggests that people's intuitions in the original case might not reflect an interest in maintaining real relationships but rather hinge simply on a desire to maintain the status quo.

Also: I look forward to taking a look at Adam Kolber's paper (which he links to above). He has done some really interesting work on related topics, and I'm excited to see what he comes up with here.

So you think that people are widely mistaken about what is valuable or plain irrational? Deep relationships have value but people do not choose to remain in such relationships for that reason because either they don't recognise that value or because they are motivated by something other than what they think is good? hmh. I thought charity required understanding people as rational, informed and lovers of the good.

I do agree with Joshua that status quo bias is a form of irrationality. I even find it misleading to characterize it as a case of one's valuing something that is not actually valuable. It's only in a very weak and dubious sense of “valuing” that anyone values a cup that was declared at the beginning of the experiment to be theirs more than a cup that is given to them at the end of the experiment. Status quo bias is a form of irrationality because your choices fail to conform to what you actually value. Just like mental accounting doesn't show that you are particularly fond of money saved from small purchases, status quo bias doesn't necessarily show that you have an immoderate passion for your current state; you might just not excel at making decisions that involve potential losses or tradeoffs. Of course all this needs more argument, but some evidence that SQ bias is really a bias is that people are embarrassed to find out, at least in many cases, that their choice exhibits this kind of bias.

Now no one would want to suggest that choices based on friendships, attachments, etc. are to be regarded as status quo bias. And given that it is probably very hard to internalize the idea that all these attachments are mere illusions and completely worthless, it would seem to reasonable to think that people don't want to disconnect in part because they don't want to leave behind their (virtual) friends, drinking buddies, etc.

However, I do think that it is hard to deny that SQ bias (and loss aversion more generally) is playing a large role in people’s aversion to disconnect in the experiment. The modified neutral vignette invites exactly this kind of bias. I know only that my real life is very different; it could be exactly the kind of life that I would rather not face. But for that very reason I don't understand how this experiment could provide a competing explanation to Nozick's thought experiment. Nozick purports to show that hedonism is false and that, at least, people “care about living in contact with reality”. If people don't “care about living in contact with reality,” why would status quo bias or loss aversion explain why they don't step into the machine? Status quo bias doesn't predict that we'll turn down sheer gains (it’s not that people value “lack of change”; they overestimate the disutility of losses). People don't typically turn down lottery earnings or inheritance money, huge promotions with no strings attached, marriage proposals from the love of their lives, etc. If failure to go into the machine is a manifestation of SQ bias then there must be something that people think they are losing when they are getting into the machine. And since they can program any experience they want to program, they'll only lose something if they care about something beyond the quality of their experiences. SQ bias might lead us to believe that the value of the contact with reality is being overestimated, but not that people don't value it at all. On the other hand, the fact that SQ bias is at play at these experiments conjoined with the fact that still a large proportion of people (41%) still want to disconnect even in the second neutral vignette, suggests that they value contact with reality quite a bit. Even when the scenario makes us tremendously prone to SQ bias, many of us still choose to disconnect.

I think Dan made a good point early on, one that sometimes my students make about this sort of argument. Suppose you could switch lives with some actual person, like in the movie Freaky Friday. Stipulate that whatever you think makes for a good life, that other person has lots more of it than you do. I think most people wouldn't switch. At least, some of my students say they wouldn't. (I haven't done any experiments.)

As Dan says, this would indicate that connection to reality may not be what's driving people's judgments. Furthermore, it would show that what people would choose in such cases has nothing to do with which theory of well-being is true, contrary to the line Jussi is pushing. This just shows we must value things other than our own well-being. People care about keeping the friends they have, etc., even at a cost to their well-being.


If *my relationships* to *my friends* make *my life* better, I'm not sure how other person's life could have lots more what makes *my life* better unless the other person was much more friends with my friends than I was. If that's right, then that I wouldn't switch is compatible with me being concerned about my well-being. Switching lives would make my life worse because I would lose my friends. I might acquire new ones eventually as I live the other life but whether they would compensate for the losses would not be very certain.

I agree though that in many cases what we choose doesn't have much to do with our well-being. We make choices for instance for moral reasons and non-well-being based personal reasons.

Ben, to add to Jussi's reply:
Nozick's thought experiment has *nothing to do*, so far as I can see anyway, with the experience of switching lives with someone else. If what you want is what you've got, except better, then that's exactly what you can look forward to experiencing in the experience machine. (I made this point in a comment above).
De Brigard might argue against this that what motivates us is not a bias in favor of *merely experiencing* the status quo being maintained, but rather a bias in favor of the status quo actually maintaining, irrespective of how our experience goes. And this could motivate us not to plug in to Nozick's machine (we'd *really* lose touch with our friends, even if we never experienced this). But there's no evidence that such a bias is at work to be drawn from De Brigard's thought experiment. So most of the discussion here seems to me to be beside the point.

A couple of points:

First, Nozick seems to argue against hedonistic theories of value, not well-being, which may or may not be identical with value. One could accept Nozick's contention that pleasure is not the only thing of value, that well-being is solely constituted by pleasure, and that well-being/pleasure, "being in touch with reality," and/or various other things are valuable. I recognize that many people have applied Nozick's argument to hedonistic theories of well-being, but I don't think that Nozick intended it that way.

Second, even if we did value "being in touch with reality," this valuing would not (indeed, could not) track whether or not we actually were in touch with reality.

For example, suppose that Frank, after seeing The Matrix, thinks, "I surely do value the fact that I am not living an illusory life!" Of course, Frank is a brain in a vat, so he can't value the fact that he isn't living an illusory life, but he values and approves of his belief that he isn't in much the same way that you value your (probably true) belief that you aren't, if you do value it.

For a different sort of example, consider Francine, who is not a brain in a vat but who deeply values the protection of the all-powerful (but non-real) Odin, just as she would if she lived in a universe that contained a real Odin. If we asked Francine whether she would prefer to live in the Odin-containing universe, she would probably say yes, she's glad she does.

The fact that I wouldn't want to be hooked up to a machine by "superduper neuropsychologists" doesn't seem clearly to illuminate anything. I might already be hooked up to a machine or be a brain in a vat. If so, I might conclude that since I value writing superserious blog entries more than drinking wine and watching Comedy Central (I'll get to that in a minute) I must value things other than pleasure (or I must take pleasure in things other than booze and laughs).

I could also toss in the conclusion that I value really writing blog entries, but what would that add to what I had already said? That is, how would vat-me's valuing differ from real-me's valuing? Does real-me have superior well-being or a more valuable life? If so, I hope I'm real-me, but I suppose I shall never know.

It's not obvious how we should answer these questions, but the Experience Machine gimmick seems to me to obscure at least as much as it illuminates when we think about these matters.

The comments to this entry are closed.

Ethics at PEA Soup

PPE at PEA Soup

Like PEA Soup

Search PEA Soup


  • Unless otherwise indicated, the views expressed in any given post reflect the opinion of only that individual who posted the particular entry or comment.