The next x-phi video (freedom)

Ram Neta stars in the next x-phi video about the Aristotelean conception of freedom, ship captains, and drowning wives.

Read the full paper here.

Comments

[Edouard Machery]

Amazing!

[Felipe De Brigard]

Very cool. (Loved the shirt.)

[Rob Wilson]

"Amazing" and "Very cool" indeed, at least about the video, until it gets to the punchline about the only thing that could explain this ... attention grabbing, yeah, but bad philosophy.

In the study, the probe question in Study 1 was:

The captain was forced to throw his cargo / wife overboard.

"Forced" is ambiguous between a non-moral and a moral sense, something suggested by the split in the traditional literature on free will between "the metaphysical types" and the "ethics types". The contexts of the two scenarios (understandable) makes salient each of these. I suspect something like this goes on in the subsequent studies in the paper.

[Joshua Knobe]

Rob,

That's a very interesting suggestion, and I'd love to hear any ideas you might have about how to decide between it and the view we offer in the paper.

In our view, there are not really distinct 'senses' of these various expressions. Rather, to say that a person acts 'freely' always means, roughly, that this person had certain alternatives open.

The question, then, is what it means for there to truly be open alternatives in a given case. Here people presumably just use their normal capacities for quantifier domain restriction, which will of course be very sensitive to context. In some contexts, we might be very liberal in our view about what counts as an open alternative (regarding the person as free whenever there is a metaphysically possible alternative available); in other contexts, we might be far more stringent (regarding the person as free only when there is an ethically permissible alternative).

On this sort of view, there is no need to posit a distinct 'ethical sense' of the relevant terms. All of the effects of moral considerations follow just from the very same sense of the terms we use when speaking metaphysically.

Do you have any thoughts about how one might decide between this approach and a view like the one you suggest above?

[Greg Frost-Arnold]

How about the following for another explanation of the findings?

Take Josh's characterization in the above comment of free action: a person acts freely = (roughly) this person has certain alternatives open. In the cargo case, respondents think that there is only one real option available to the captain. However, in the wife case, something that was previously not considered a live alternative is made salient: throwing a person overboard.

I wonder whether (some) respondents' reaction in the wife case could be explained by something like: "Well, if throwing the wife overboard is a salient, available option, then so is throwing _yourself_ overboard, to save your loved ones."

That is, thinking about throwing the wife overboard makes salient or available a second alternative (throwing yourself) that was not on the table before. Or maybe a third option of going down together -- once the normally unthinkable option of killing your family is on the table, other normally unthinkable/ unavailable options become available.

[john dell]

I had a similar intuition about this case as Greg Frost-Arnold. It seems like underlying both these cases is the assumption that the captain should preserve what he values most. In the first case this is easy; it is his own life. This makes for a simple hypothetical: If the captain wants to survive (and he does), then he must (i.e. he is forced to) throw the cargo overboard.

Demarcating what the captain values most in the second case is tricky. Does he value his own life most, or his wife's life? What is interesting here is that it seems like the inability to distinguish what is most valuable leads to the impression that more options are on the table. Since it is not clear what the captain should value most in the second case, it is not clear that he is forced to do anything. In other words, if we were to try and construct an action guiding hypothetical for the second case, it would not be clear what the antecedent should be.

[Josh May]

I'm loving these short X-phi videos! They sure are a way to make philosophy more accessible to the public. Great work!

BTW, Josh, is that your wife's band playing the music? Good stuff.

[Felipe De Brigard]

I still believe that there is a possible asymmetry between the two cases which may actually be explainable in the terms Josh used above. I think that people's way of restricting the scope of their quantifiers when assessing available options may be sensitive to judgments of value. In the case of the captain and his wife's cargo, there is a difference in terms of how much they are worth. It is a bad thing to throw it overboard, but given the difference in valence between the captain and the cargo, it is best to opt for the cargo (which is worth less). But in the case of the captain and the wife, both being of equal value, there is clearly an option for him: to throw himself. I bet if you describe the case in the same way and put it in terms of the captain throwing himself instead of his wife, people would give similar results as in the case of the cargo: he was forced to do so (and presumably for reasons similar to those operant in the first case). This problem is equivalent to the alternative of throwing oneself in the trolley problem. One usually makes it explicit that such isn't an alternative, but by ruling out a plausible response to a thought-experiment you aren't thereby ruling out a psychological component that may factor in a moral decision.

[Bryce Huebner]

FYI: We have a bunch of data on throwing yourself in front of the trolley! Hopefully i will have it written up soon and will be able to put it up on the blog...

[Jonathan Phillips and Joshua Knobe]

A number of commenters have pointed out that, in the second version of the case, the captain could just throw *himself* overboard. This is a good point, which one could explore in further studies. For example, suppose one explicitly says that the wife is much fatter than the captain (so that only the act of throwing her can prevent capsizing). Or suppose that the wife does not know how to sail the ship (and hence would die anyway if he went overboard). Would people still say that he was not forced to do so? When we try these revised cases on ourselves, we still get the same answer, but it is certainly worth testing experimentally.

In any case, the actual paper presents a second experiment that is not plagued by this sort of confound. Subjects were told about a doctor who is ordered to prescribe a certain medicine. In one case, the medicine will help the patient; in the other, it will harm the patient. Subjects are asked whether they agree with the statement that he 'did not really have the option' of not prescribing the medicine. Here too, we get an asymmetry, with people being more inclined to say that he did not have the option when the action is good than when it is bad.

Does that help to address the worry?

[john dell]

Hi Jonathan and Josh

As far as future constructions of further studies go, I think the capsizing case might yield a more balanced set of intuitions with the cargo case. The reason for this is because in the cargo case it's, all things considered, throw the cargo over or die. In the capsizing case, it's throw your wife over or you both die. In a way, this eliminates the conflicting value judgments (see my post above) and puts the captain and his wife on a level playing field. In other words, if the captain's wife is going to die anyway, then there should not be as much dissonance in his deciding to save himself. If there is not an evaluative conflict then it should be clear what *must* done. When what must be done is salient, then I think this is what gives rise to the intuition that the captain was forced in his actions.

On a separate line of thought, though one not unrelated to the evaluative idea I have been pushing. I wonder what would happen if you kept the cargo case and the wife case the same, and used a gradient of cases between those two. For example, in between the two original cases you might add a case where the cargo is extremely valuable, a case where the captain has his dog with him, a case where the captain has an ordinary ship mate with him, and a case where the captain has his best friend with him What would be really interesting is if the respondents increasingly said the captain was not forced as the item on the ship become more and more valuable.

[john dell]

Sorry, I meant to say in my first sentence above that I think a capsizing case where it is clear the captain's wife will die no matter what, might yield a more balanced set of intuitions with the cargo case.

[Joshua Knobe]

John,

I really like the idea of looking at intermediate cases. If the theory is actually correct, one might predict that judgments about whether the captain was forced to throw his dog overboard would end up being intermediate between judgments about throwing the cargo and judgments about throwing the wife. (Moreover, one would predict that people's judgments about whether he was forced to throw the dog would be correlated with their *moral* views about the status of dogs.)

As for the point about the moral differences between the two cases, you are certainly right to say that these differences exist, but perhaps they should be seen not so much as confounds but as aspects of the very phenomenon under study here. After all, the main point is to show that people's moral judgments are influencing their intuitions about whether the agent is free, and that is precisely the heart of the hypothesis you are proposing.

[john dell]

Hi Josh, what you said sounds absolutely right. I guess now all that's left is to go out and see if the hypothesis holds up!

johnldell@gmail.com

[Jonathan Phillips]

Felipe,

You made a really good point about the difference in value between the two items thrown overboard in the experiment we conducted.

However, I am not sure that the general value of the object thrown overboard could account for the asymmetry in the study. If we imagine that instead of his wife's cargo, the ship captain threw overboard a heavy box of priceless 12th century artifacts, would that at all change whether or not he was forced to do it? My intuition is that no matter how valuable we made the object thrown overboard, the captain was still forced to do it. So I don't think it was just the wife's general value that influenced people's judgments that the ship captain wasn't forced to throw her overboard.

But my theory here is probably influenced by some earlier research I did on political freedom (about whether or not the value of an option which is restricted affects the extent to which freedom is subsequently lost), it turned out that the general value of the restricted option did not significantly influence participant's judgments at all.

The basic design was that participants were given a vignette in which a dictator arbitrarily banned one of three things from his country: (1) well-written newspapers, (2) soap operas, or (3)violent pornography.
Participants were asked both about the extent to which they valued the thing restricted and the extent to which freedom was lost as a result of the restricting law. Although the judgments about the value of the restricted thing were highly varied, the judgments about the loss of freedom were almost identical.

So maybe the same rule would apply in this study as well. Anyway, your suggestion about the variation in which the captain throws himself overboard is really interesting and seems worth pursuing further. My thought is that people would not say the ship captain was forced to throw himself overboard, but maybe we will soon profit from the research Bryce has already done.

[Victor Caston]

This is drive-by history of philosophy.

It actually pays to look for a moment at Nicomachean Ethics 3.1. (Phillips and Knobe cite it in the paper, but little more.)

The case of the ship captain is not used to "illustrate the distinction between the voluntary and the involuntary," as Neta claims in the film. And Aristotle doesn't ever describe the action as 'forced.'

After drawing the distinction in a preliminary way, he raises cases where people's intuitions differ -- x-phi? -- such as when a tyrant orders you to do something despicable when he has control over your spouse and children (1110a4-9). He then compares it to the case of the sea captain in the storm and declares all such cases "mixed," though "more like voluntary actions" (a11). The agent acts willingly and source of whether to act or not lies in the agent (a15-18). The action in these is not something *in general* an agent would choose willingly, not because he is forced, but because it is not something he in general wants (a18-19). But the end of an action depends on the occasion, and *these* actions are choiceworthy when done in those circumstances (a12-14).

Aristotle would not regard the second case brought up in the film as choiceworthy even in a shipwreck, precisely because of the morality of the action. If someone performed it, it would be done willingly and wrongfully: as he says just a few lines later, "in some cases it is surely not possible to be compelled, but one ought instead to die suffering the most terrible things" (a26-27).

Aristotle does not consider any of these mixed cases forced, but "voluntary" -- that question depends on whether "the origin of the action is in the agent" (a15-18). This is still more clear if one renders the Greek terms 'willingly' and 'unwillingly.' All of these actions are willing, but reluctant. Whether they are chosen reluctantly depends on whether the end is worthy of choice in general and in the circumsances. (Or never worthy of choice, as in the second case brought up in the film.)

[Sven Nyholm]

Victor,

towards the end of your comment you say ...All of these actions are willing, but reluctant. Whether they are chosen reluctantly depends on whether the end is worthy of choice in general and in the circumstances... On Aristotle's view, does this depend on whether the end is worthy of choice in general and in the circumstances according to the agent who does the acting? Or does it depend on what has these features when evaluated from some objective/impartial or some other perspective?

It is interesting to think about this feature of Aristotle's view in relation to the so-called Knobe-effect. Because if the latter of the two possibilities above is what's the case---that is, if whether willingness on Aristotle's view depends on whether some action is objectively choiceworthy in the sense that its choice-worthiness does not only amount to whether the agent who does the acting finds it to be so---then Aristotle seems to have anticipated Josh Knobe's findings. (I'm thinking, of course, of the so-called chairman case where I am substituting "voluntarily" or "willingly" for Joshua's "intentionally".)

[Jonathan Phillips]

Victor,

Thanks for your comment; you are bringing up a very good point about the passage from Nicomachean Ethics. This passage is definitely worth really considering. If the paper is more focused on the psychology side of the issue rather than the history of philosophy side, it is only because it was written for a psychology journal.

On our reading, the passage which we referenced is actually pretty complex and Aristotle seems to invoke two very different conceptions of freedom, both of which you mention.

On the one hand, he suggests, 'One acts voluntarily, because in such actions the origin of one's moving one's limbs is in oneself; and where the origin is in oneself, it is also up to one to act or not.' (1110a 15-18) On this reading, it seems that both actions were voluntary because the origin of the movements of the ship captain's limbs were clearly within him.

On the other hand Aristotle also suggests, 'Actions of that kind are mixed, but they more closely resemble voluntary actions; for they are choiceworthy at the time when they are done.' (1110a 11-14) So on this reading, only one of the actions was done voluntarily, that of throwing the cargo overboard.

What is interesting is that subjects seemed to disagree with both of these theories. However, in the paper, we do offer a way in which all of this can perhaps be reconciled. If, as Aristotle seems to suggest, an act is voluntary to the extent that the origin is within the agent, then it seems plausible that in the case of the cargo, participants understood the origin for the action as *outside* of the sea captain, while in the case of the wife, participants saw the impetus for the action as *within* the sea captain. It seems like this asymmetry can easily be seen as a consequence of the moral status of the sea captain's actions.