The Experience Machine Part 2: Now We Poke (A Lot Of) Holes

Difficulty: What the fuck

This is Part 2 of a series on Robert Nozick’s Experience Machine Thought Experiment (EMTE), so you might want to read Part 1 first. This series is a simplified, slightly profane version of the Internet Encyclopedia of Philosophy’s article “The Experience Machine.”

Now things get more complicated. We will backtrack a little and look at some of Nozick’s stipulations for EMTE and see if he’s cheating dialectically. We’ll then look at a few objections to the abductive interpretation of his argument.

Nozick’s stipulations

The stipulations: Nozick tells us to set aside questions like “What is reality?” and “How can we know it”? The first question is what philosophers call a metaphysical question; that is, it deals with what things are on a fundamental level. The second question is an epistemic (knowledge) question: how do we know that something is true? How do we know that life outside EM is any more real than life inside EM?

On to the first question: If there is no such thing as reality because all reality is is mental states, stuff that goes on in my head only, then EMTE is pointless: life inside and outside EM would be equally virtual and “fake.”

And about the epistemic question: René Descartes famously advanced the Evil Demon problem to argue for a view called global skepticism: How can I know that I’m not dreaming? How can I know that I’m not a madman, and my “hands” are, in reality, two deformed pumpkins or some other crazy shit? How can I know that anything I perceive is real at all?

If global skepticism (or a more local skepticism that targets EMTE) might be true, then without Nozick’s stipulation to set aside the epistemic question, EMTE would also be pointless: how can we say that the “virtual” world of EM is more, or less, real than the world outside EM?

Lastly, there is the contextual stipulation: Nozick says we should set aside the question of context when we think about EMTE. For instance, if a possible context is that life outside the EM is torturous and unbearable, that might make us decide that plugging into EM is preferable. Nozick tells us to set the contextual question aside and assume that the pleasures we experience outside EM are what the average Joe experiences, so, nowhere near super pleasurable, but also nowhere near unbearable.

Is Nozick cheating?

So is Nozick cheating here?

Well… maybe. But not in a stupid way.

Let’s look at what he’s doing. He tells us to ignore:

• metaphysical questions (what reality really is)
• epistemic questions (how we know what’s real)
• contextual factors (like whether real life is miserable)

That’s a lot to set aside.

So one reaction is: “Hold on, aren’t those exactly the kinds of things that might affect my decision?”

For example, suppose:

• there is no meaningful difference between “real” and “virtual” experience
• or we can never know which one we’re in
• or real life is absolutely terrible

In those cases, plugging into the machine suddenly looks a lot more appealing.

So by ruling out those possibilities, Nozick might look like he’s stacking the deck, designing the thought experiment so that we’re more likely to say, “No, I wouldn’t plug in.”

But here’s the other side.

If we don’t set those questions aside, the whole thought experiment collapses.

If we start worrying about whether reality exists, or whether we can know anything at all, then we’re no longer evaluating hedonism, we’re doing skepticism and metaphysics.

And if we allow extreme contexts, like a life of constant torture, then of course we might choose the machine. But that doesn’t show that pleasure is the only thing that matters. It just shows that sometimes pleasure outweighs everything else.

So Nozick’s stipulations aren’t obviously cheating. They might just be a way of isolating the variable he cares about:

Does pleasure alone determine how well our lives go?

Still, the worry remains. If a thought experiment works only after you strip away enough complications, you might wonder whether it’s telling us something deep, or just guiding us toward a pre-loaded intuition.

The abductive argument, revisited

Now that we got that out of the way, let’s revisit the abductive interpretation of Nozick’s argument:

What’s the best explanation for the fact that many people care about reality, not just pleasurable experiences? One possible answer is that reality itself has intrinsic value.

The question we now focus on is whether “reality itself has intrinsic value” really is the best explanation for why many people care about reality. There are at least three possible sources that make us question the abductive argument: the hedonistic bias, imaginative failures, and the status quo bias.

Hedonistic bias

The hedonistic bias is the most speculative of these three potential biases. Really, it’s controversial as hell, because its core claim is:

You think the reason you don’t want to plug into EM is that you value not just pleasure, but also reality. But actually, all your desires, including the desire not to plug in, are still best explained by the fact that what ultimately drives them is pleasure.

Even more bluntly:

You don’t want to plug into EM not because you value reality, but because you unconsciously believe that reality would get you more pleasure than EM would.

Speculative and controversial as that claim might be, if it’s possible that the hedonistic bias might be true, then it’s possible that Nozick’s EMTE does not, contrary to what many philosophers in the 1990s believed, provide a knock-down argument against hedonism.

Sharon Hewitt shares similar views, arguing, in her 2010 paper “What Do Our Intuitions about the Experience Machine Really Tell Us about Hedonism?,” that it’s not enough to merely stipulate that our practical hedonistic reasons for not plugging in to EM are irrelevant; we must also “fill in the concrete details that would make them irrelevant.” In other words, you can’t just tell people to ignore their practical concerns; you have to show why those concerns don’t actually matter. Further, our intuitions about EM might not be reliable indicators of what is intrinsically valuable: rather than telling us about what truly matters, our intuitions about EM might just be telling us something about how our brain is wired.

Or to flesh it out a bit more:

1. Thought experiments target gut reactions: we’re trying to see what we intuitively prefer.
2. But those reactions are “pre-reflective”: we didn’t think it through, we just reacted.
3. The problem is that we don’t really know why we reacted that way.
4. So our reactions are not reliable evidence of what actually has value.

Imaginative failures

Let’s move on to imaginative failures. This is another source of bias, which can be divided into two separate confounding factors: imaginative resistance and overactive imagination.

Imaginative resistance

Empirical evidence shows that imaginative resistance, namely, the phenomenon whereby people disregard the stipulations of a thought experiment, is an actual problem when people decide whether they would want to plug into EM.

In his 2014 paper “Nozick’s Experience Machine Is Dead, Long Live the Experience Machine!,” Dan Weijers found that 34% of the participants who did not want to plug into EM were affected by imaginative resistance, disregarding some stipulations of EMTE by worrying about things like whether EM would malfunction, even though Nozick had explicitly stipulated that EM would never malfunction and would provide a perfectly blissful experience.

(This is where you might pause and say, “Oh, those people don’t know how a thought experiment works. You can’t disregard stipulations; that’s the whole point of a thought experiment!” But then again, even philosophers make that mistake at times. The philosopher Ben Bramble, in his 2016 paper The Experience Machine, ignored Nozick’s stipulation that EM provides a perfect simulation: Bramble argued that artificial intelligence is too primitive and that it thus cannot possibly give us a convincing simulation of loving someone and being loved in return. But with recent advances in AI, of course that argument sounds stupid, in hindsight.)

Overactive imagination

Then there’s the problem of overactive imagination, which, according to Weijers’ study, happened for 10% of people who didn’t want to plug into EM. People who have an overactive imagination when thinking about EMTE might be biased against plugging in due to reasons not stipulated by EMTE. For instance, if they’ve watched The Matrix one too many times, they might say, “Oh, no thanks. I don’t want to plug in because I don’t want to be enslaved by aliens.” This, clearly, goes beyond what EM stipulates.

Also, considering that a whopping 46% of anti-plug-in philosophers and 39% of anti-plug-in laypeople were biased by at least one of these two imaginative failures, it’s quite possible that EMTE is simply too far-fetched for us to reliably conclude anything about what we actually value. In other words, our judgments about extreme scenarios like this might be less reliable than we think.

a. Memory’s erasure

Now here’s another complication of possible imaginative failure, particularly one of imaginative resistance: Bramble argues that another reason people might not want to plug into EM is that doing so forces EM to erase some of our memories, making it seem like we would die if we plug in. And since most people don’t want to die, the threat of death might be their reason that they wouldn’t plug in.

If this is hard to wrap your head around, it might be easier to understand what Bramble is saying by thinking of the TV series Severance (if you’ve watched it), in which employees working at a company undergo a “severance” procedure that not only erases their memories of their lives once they enter the workplace, but also makes them forget, once they’ve entered, that they underwent the procedure at all. So if, hypothetically, an employee were to enter the workplace and never come back to the “real world”, the only place they can regain memories of life outside work, then their non-working identities would, in a sense, be dead.

In his 2012 paper “What’s Wrong with the Experience Machine?,” Christopher Belshaw raises an even stronger objection: for EM to work as intended, it would require not just some memory erasure, but invasive memory erasure. Since EMTE stipulates that you wouldn’t remember that you plugged yourself in, for the pleasures to be meaningful in EM, scientists would have to recreate your identity for the pleasurable experiences to be realistic enough.

At this point, the IEP article accuses Bramble’s and Belshaw’s arguments of being victims of imaginative resistance themselves. That is, although EMTE doesn’t explicitly stipulate that identity would not be erased, nowhere does it insinuate that your identity would be erased. In EMTE, identity preservation is implicitly assumed, so by disregarding this implied stipulation, Bramble and Belshaw might be suffering from imaginative resistance themselves.

b. Moral concerns

Moral, as opposed to prudential, concerns are also stipulated as irrelevant in EMTE, though people disregard these stipulations and decide not to plug in anyway. This, again, is a case of imaginative resistance.

For instance, Weijers reported that a participant said that “I can’t [plug in] because I have responsibilities to others,” and Guido Löhr, in his 2018 paper “The Experience Machine and the Expertise Defense,” wrote that anti-plug-in philosophers stated, “I cannot ignore my husband and son,” “I cannot ignore the dependents”, and “Gf[girlfriend] would be sad.”

But EMTE stipulates that we should ignore moral concerns and think only of prudential reasons, that is, reasons for why something is good for me. So this disregard for this stipulation shows just how both laypeople’s and philosophers’ seemingly muddled thinking might render conclusions about what we value unreliable.

And imagine this: if someone told you to disregard moral concerns in a case of, say, sexual violence, would you be able to do so? Probably not. In short, the fact that someone stipulates the rules doesn’t mean you’ll follow the rules.

So what does all of this suggest? It suggests that all our crappy imaginative failures might be a big reason that most of us choose not to plug in. If, on the other hand, we were not affected by any of these biases, then our strong intuition against plugging in might weaken, or at least become less reliable as evidence. That would weaken Nozick’s argument that hedonism is false.

Status quo bias

Lastly, we come to the status quo bias, which means that we humans have a tendency to abide by the adage “when in doubt, do nothing,” especially when faced with complex decision making. We see this happen when people vote for an incumbent politician or decide not to trade in a car. And some studies have exposed just how powerful the status quo bias is during EMTE-like decision making.

Take, for example, the reverse experience machine thought experiment (REM), where participants are asked whether they would want to unplug from EM if they were already living in EM. Most people’s answers: no. Don’t unplug.

And then there’s the “stranger No Status Quo” (stranger NSQ) scenario, which attempts to remove the status quo bias from REM and see how people would respond. The stranger NSQ scenario is based on the idea that the more detached we are from the subject of a thought experiment, the more rational, and thus less biased, we would be. So, researchers asked people to make a decision for a hypothetical stranger: should the stranger, having spent 50% of his or her life in EM, be unplugged and enter reality? Given this scenario, only 55% of participants were pro-plug-in.

This is a problem for Nozick because the same basic preference can generate opposite answers depending on how the question is framed. That suggests the intuition may not be tracking value at all, but bias.

Conclusion (for now)

So what does all this say about Nozick’s abductive argument against hedonism? Recall that the abductive argument goes:

What’s the best explanation for the fact that many people care about reality, not just pleasurable experiences? One possible answer is that reality itself has intrinsic value.

Sure, that’s a possible answer, but is that really the best explanation for why most people choose not to plug into EM in EMTE as formulated by Nozick? Probably not. We’ve seen that there are many other good explanations for why people choose not to enter EM: the hedonistic bias, imaginative failures (including imaginative resistance and overactive imagination), and the status quo bias. And if any one of these explanations is better, or even comparably good, then the abductive case for saying that reality itself has intrinsic prudential value becomes much weaker.

In other words, EMTE does not straightforwardly show that hedonism is false. At the very least, it is not the philosophical knock-down argument that many people once thought it was.

That’s all for now. In Part 3, we’ll delve into EMTE even further. Meanwhile, keep thinking.