Blog Coordinator

« Signing Off | Main | The Featured Author of the Month: Kristin Mickelson »



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

The abstract doesn't mention any experiments where they compare morally neutral actions to morally admirable ones. It could simply be that the moral significance of the tested actions could prime other morality-related concepts, such as free will.

Of course, if they didn't get the effect with morally admirable actions it might be down to dissonance reduction ("I'm not a bad person; that person is just abnormally self-denying."), so I'm not sure they could win....

This research is very interesting (I read the paper a while ago so I don't remember if they have results addressing my questions below). It's plausible that people will want to justify their judgments of responsibility and their desire to punish and may do so with attributions of (increased) free will. But as Mark suggests, it might be that morally valenced actions might prime notions of control and free will, and hence increase attributions, without this being driven by a desire to rationalize judgments of responsibility.

Furthermore, regarding Nietzsche's hypothesis, why would a culture need to create a theory of free will to justify blame? Drawing on Tamler Sommer's sort of examples, there seem to be past and present cultures that are entirely happy to punish the hell out of people without requiring that those people had any or much control over their actions (e.g., killing seemingly innocent relatives for the 'sins' of their relatives). Presumably our hominid and earlier ancestors blamed ("blamed"?) and punished without much consideration of free will.

If so, then perhaps the need to *justify* such practices with a concept of free will would have arisen when a culture started to recognize that it's ineffectual (and unfair) to punish people who do not control their actions in the requisite ways. Such a culture might develop the concept of free will, but not as a sort of post hoc rationalization for punishment, but because of a recognition that there are different degrees of control people have over their actions, and different responses are appropriate (justified) for actions produced with different degrees (or types) of control. (Compatatibilist) free will might be a good label for such control. It certainly picks out many of the distinctions current Western legal practices use to assess guilt and punishment.

(Agent causal free will might be a good label for the sort of control that might--but not really!--justify eternal damnation, exactly Nietzsche's target, but certainly not the notion of free will most of the participants in these experiments are using ;-)

First of all, a warm welcome to Kristin! It's wonderful to have her here as a featured contributor, and I hope that this post doesn't at all detract from the (very interesting) discussion of her work.

Anyway, Mark and Eddy raise some extremely helpful points here. If you read through our paper, you will find a lot of very strong evidence that moral judgments are in some way impacting people's free will beliefs, but you will only find a little bit of evidence that this effect is driven by a desire to justify punishment. It is always possible that the effect will turn out in the end to be the product of some other sort of psychological mechanism.

Eddy's comment actually brings up a very nice way of putting our hypothesis to the test. Suppose that people from different cultures differ in the degree to which they think free will is necessary for punishment. Then our effect should arise only for people who do see this connection between free will and punishment, not for those who don't.

One final quick note: In another study (not reported in the paper), participants were asked about the degree to which various famous individuals had free will. Some were famous for doing good things (Gandhi, Mother Teresa); others were famous for doing bad things (Bernie Madoff, Jerry Sandusky). Surprisingly, the bad ones were judged to have more free will than the good ones! This result provides at least some evidence that the effect is specific to moral badness and does not arise for just any act that has a lot of moral significance.

But Joshua,

Suppose that people from different cultures differ in the degree to which they think free will is necessary for punishment. The cultures that don't think it's very necessary for punishment might not think free will plays much role in morality in general. In which case, any priming effect of morality on free will is reduced or lost. In which case, the same experimental result is predicted.

Joshua, your data on Gandhi and Mother Teresa is interesting, but there is still the dissonance reduction theory to rule out. The upward social comparison is hard on our egos so a corresponding reduction in the importance of free will helps guard us against feeling bad. I'm not sure how you could rule out that possibility.

Thanks for the warm welcome, Joshua. When it comes to discussions about free will, I say the more the merrier!

Thanks so much for all these super helpful comments! They definitely help to get at precisely the issues we would need to address in further work.

Paul is completely right to say that the predicted cross-cultural difference would not help to decide between the competing hypotheses. (It would show that the effect was in some way driven by the fact that people see free will as necessary for punishment but would not tell us precisely why.) Still, the Gandhi/Bernie Madoff study does seem to shed some light on this issue.

Mark is also right to say that this latter effect could be driven by people's desire not to see themselves as blameworthy. (If there is no free will, maybe I can't be blamed for not being like Gandhi.) Yet, though this hypothesis differs in some ways from our own, it is clearly in the same general family. Both hypotheses say that people's judgments are being driven by a motivation to justify certain claims about blame -- either the claim that others are blameworthy or the claim that they themselves are not.

Here's another idea that occurred to me: maybe a bad outcome highlights the role of earlier choice, in a way that a good outcome does not. And incidentally, the goodness need not be moral goodness. I often find myself going over and over past decisions when I regret the outcome, but not otherwise.

If this hypothesis were true and important to explaining the results, we would expect it to apply at least as well in one's own case. So give subjects vignettes in which they either do something which successfully advances their career, or something which accidentally damages it. Then ask about the agent's freedom in the scenario.


Excellent point. We actually did run a study in which participants were asked to recall a time when they did something that was either morally good or morally bad. Strikingly, in that case, the effect flipped direction. People were actually *less* inclined to believe in free will after writing about a time when they did something bad.

This result seems to provide at least some evidence that the effect is driven by a desire to blame, rather than just by moral badness per se.

Thanks for the heads up on this article, and for the research itself. This is one of the most fascinating and stimulating set of studies I have seen in a long, long time; remarkably fine research, beautifully designed, and leading to very important results. One quibble in study 5, the item a173: “Some people feel they have complete free choice and control over their lives while other people feel that what they do has no real effect on what happens to them.” Does this get more at a question of locus-of-control (internal vs. external) instead of free will and the larger issue of control? Someone might believe that he or she has very high internal l-o-c, and that the major events in his or her life are largely the result of his or her own efforts; and yet also believe that he or she is remarkably incompetent when it comes to being effective in exercising that control. I feel that way most of the time: the major events in my life are largely under my control, but I am pathetically bad at managing them (I am high in internal l-o-c, but low in sense of self-efficacy for managing my own life). What happens is up to me; but I don’t have much effective control over it (imagine yourself being thrust into the position of the starting quarterback of the New England Patriots – probably one of Tamler’s fantasies – in which the outcome of the game is really up to you, but you can’t control it effectively; what you do has genuine effects, but not effects you can successfully control).
I particularly like Eddy’s point, that people punished enthusiastically long before there was any justification based on claims of moral responsibility or free will (as Tamler makes admirably clear). It was (according to Bernard Williams) when people began to think of the world as just (as morally ordered) that they began to look for moral justifications for punishment. You note the just world research (in relation to Study 3); it would be great to have experimental data that manipulates just world beliefs and examines the effect on belief in moral responsibility and free will.
Finally, Study 5 was intriguing; as you note, it raises questions about the Vohs-Schooler result (though I have some doubts about that experiment anyway: they may be testing the effects of belief in fatalism or even learned helplessness, rather than the loss of any real notion of free will). There are lots of variables operating, as you point out; but in any case, this interesting result would not be surprising to criminologists such as Cavadino and Dignan, who would attribute the high crime rate and the high free will belief rate to a common underlying force, the neoliberal social-political view that celebrates “rugged individualism” and thus exaggerates the powers of free will (into a radical libertarian version) while also promoting greater disparities of wealth and a much weaker social support system (both recognized factors in contributing to crime).
Just some quick thoughts, on a wonderful set of studies and a wonderful paper.


First off, thank you so much for all of your kind words. In general, I feel like the community of people working in the theory of agency tend to be incredibly supportive, but you somehow manage to go even beyond what would be expected here.

You bring up a whole bunch of interesting questions, but I wanted to focus on two in particular.

1. In the actual studies presented here, we tried to manipulate people's desire to punish, but just as you say, we could try going in the opposite direction. Instead of manipulating the desire to punish, we could manipulate the degree to which people think that free will is necessary for punishment. (For example, participants in one condition could get a passage from one of Tamler's papers about cultures in which people punish without ascribing free will.) One might then predict the corresponding change. The more people became convinced that punishment was justified in the absence of free will, the less sure they would be that free will existed at all.

2. You are completely right to distinguish between the belief that something is 'up to' the agent and the belief that the agent actually has the efficacy to succeed at something. It would be very interesting to look further at the question as to whether the effect was only on the former sort of judgment or also on the latter.

The comments to this entry are closed.

Books about Agency

3QD Prize 2014: Marcus Arvan