The Paradox of Moral Focus

Liane Young and I have recently become interested in what I think is an intriguing feature of morality's influence on social and causal cognition. This research began by borrowing from some studies which suggested that people's moral judgments can influence their judgments of force and freedom (if you don't know it, here's a funny 3 minute and 29 second overview). Liane and I took these cases and changed them just a little bit, to see if  people's moral jugments could actually lead them to giving directly contradictory judgments. 

I think the basic finding can be explained pretty simply. In one study, participants read a vignette about a hospital where a chief of surgery orders a subordinate doctor (against his will) to prescribe a drug called Accuphine. And, in some cases, the patient dies as a result of having taken the drug. Then, after reading the vignette, particpiants were asked one of two related questions:

• Did the chief of surgery force the doctor to prescribe Accuphine?
• Was the doctor forced to prescribe Accuphine?

While most participants agreed with the first claim, they actually disagreed with the second :   But this pattern of responses seems obviously inconsistent: participants both agreed that the chief forced the doctor, and disagreed that the doctor was forced. So what might explain this paradoxical pattern of responses? (more after the break.)

Liane and I have come up with one idea about how you might try and explain this pattern of results. It could be that the differnces in the wording of the question (active vs. passive) lead people to focus on one moral agent rather than the other. This difference in moral focus might change people's judgments of force by first changing the focus of their moral judgment. For example, when participants were asked 'Was the doctor forced?', participants may have been more inclined to focus on the doctor and consider alternatives to the doctor's action (e.g., he should have refused). This would have then led them to judge that he wasn't forced. Yet, when the moral focus was shifted to the chief of surgery, participants would have been less inclined to consider the alternative actions available to the doctor. This may have led them to judge that the chief did, in fact, force the doctor. The main idea, then, is just that participants produced an inconsistent pattern of responses because they considered the possible alternative actions only for the agent who was in their moral focus.

So I was wondering what other experimental philosophers would make of these results, whether you could think of another (perhaps better) alternative explanation, and whether there are some further studies that could help us get to the bottom of it all. Thanks for your help!

Comments

[Greg]

I wonder what the results would've been if you framed the 'passive' version of your question as the true passive transformation of the active sentence, i.e., "Was the doctor forced to prescribe Accuphine by the chief of surgery?" What would your hypothesis predict about that case?

[Clayton Littlejohn]

I think what you say about moral focus is plausible, but it's not obvious to me that there's a change in judgment of the sort that indicates that the speakers say something contradictory when they say that so and so forces/deny that so and so was forced.

We might think that whether "x forces y" depends largely upon whether x exerted a pressure that sufficed for getting y to act. We might think that whether "y was forced" depends upon whether the pressures that were exerted on y got y off of the hook. Is it really contradictory to say (i) that so and so forced him to give the drug in the sense that x applied adequate pressure and then to say later that day (ii) that y wasn't forced to do anything since y could have freely resisted the pressure that x and others placed upon him?

I guess if someone overheard me saying both (i) and (ii), I could fill in the sort of details I did above to show that I wasn't flip-flopping, changing my mind, contradicting myself, etc... I take it that (i) and (ii) cannot be contradictory if adding in extra material can make them consistent.

[[I'm not sure if this is the same sort of thing or not, but there's an interesting discussion of "ought" in Geach's article (along the lines of) "What happened to deontic logic?" He discusses interesting examples where it seems right to say things like, "Someone should sock that guy in the jaw" or "Someone should get someone to sock him in the jaw" while adding that there's no one such that he/she ought to sock the guy in the jaw. He credits these cases to Anselm, but puts a spin on it. My version. The undercover officer needs to learn about some biker gang and can only do so if he joins. To join, you have to get beat up by the gang. So, he should get someone to sock him in the jaw. Still, it doesn't seem to follow that there's someone who should beat the guy up. And if it's not the case that someone should beat the guy up, there's a sense in which nobody should beat the guy up. Still, he should get someone to beat him up. I don't think there's a contradiction here and your case seems similar in some ways to the cases that interested Geach.]]

[Jonathan Phillips]

This is a really nice question. We didn't ask for agreement ratings with the 'true' passive form in this experiment, but we've done that in subsequent experiments. We get the exact same pattern of results. Since I think the main issue is simply which agent is getting the moral focus, I wouldn't predict any differences with the 'true' passive and the form we used here.

[jonathan weinberg]

I like your hypothesis, but I have a somewhat deflationary one to offer (which would be easy to eliminate if it's wrong, I think). Philosophers and linguists I know often have more fine-grained intuitions about various sorts of success terms than it seems ordinary usage always follows. E.g., folks sometimes say things like "A refuted B" when all they mean is that (what we would say as) "A offered an argument aimed at refuting B". So, I wonder whether some of your subjects are perhaps understanding "Did the chief of surgery force the doctor to prescribe Accuphine?" as basically synonymous with "Did the chief of surgery commit actions aimed at forcing the doctor to prescribe Accuphine?" If so, then for those subjects, saying "yes" to the chief question might make sense, even if they would be inclined to say "no" to the doctor quesion.

[Jonathan Livengood]

I'm hoping that your probe didn't have the typographical error in the post here: in question (1) you have "Accuphine" but in question (2) you have "Accupine" with no "h" ...

[Jonathan Phillips]

Clayton, thanks for your response. I think my initial reaction would be to argue that, strictly speaking, participants responses actually were inconsistent. This point isn't anything particular about force: if you agreed that you ate a cookie, but disagreed that a cookie was eaten by you, I would think your answers were clearly inconsistent as well.

But I think your suggestion may be pointing out something slightly different. One way to understand the details you 'filled in' would be to say that there are certain implications of saying someone was forced, e.g., whether the pressure exerted on him got him off the hook. And perhaps participants were sensitive to these implications when making force judgments. Is this basically what you had in mind?

[Jonathan Phillips]

Jonathan W., I take it that your point is that there is a performance and a success reading of A forced B. This is a nice possibility to point out. One really interesting thing to note is that in cases in which the 'forcee' was forced to do something morally neutral, participants gave a completely consistent pattern of responses. They agreed that A forced B, and also agreed that B was forced by A.

This seems to be important because, if a performance vs. success reading is to explain the results, then it means that participants would have to systematically rely on a performative reading for statements of the form: 'X forced Y to do p' and systematically relied on a success reading for statements of the form 'Y was forced to do p' but only do so in scenarios involving immoral actions, despite the fact that the forcing was successful in all the scenarios.

But, I really don't think this can simply be ruled out in principle. Did you have an easy way to test this in mind? If so, I'd definitely be interested to hear it!

[Jonathan Phillips]

Jonathan L., Thanks for catching this! I (thankfully) didn't make that typo when presenting participants with the actual probe. (And I've gone ahead and fixed it on the page as well.)

[Clayton Littlejohn]

Hi Jonathan P.,

"One way to understand the details you 'filled in' would be to say that there are certain implications of saying someone was forced, e.g., whether the pressure exerted on him got him off the hook. And perhaps participants were sensitive to these implications when making force judgments. Is this basically what you had in mind?"

That's certainly part of it. Once this is on the table, I worry about your hypothesis that the speakers contradict each other. I take it that if two speakers contradict each other, their assertions can't both be true. I don't see why that description is forced on us. You seem to be assuming this:
(*) In the contexts you described, C1 and C2, the speaker who utters "A forced B" in C1 asserts a proposition that is logically incompatible with the proposition expressed by the speaker who asserts "B wasn't forced" in C2.

I don't see what the justification is for saying that the differences in conversational contexts C1 and C2 don't effect a difference in what's said when speakers utter what they do in C1 and C2. One reason to worry about (*) was a worry I had earlier. When you reveal these patterns to the speakers, do they take themselves to be disagreeing? One reason to think that they don't is that there is a way of filling out the details in such a way that speakers in C1 and C2 can see that each are right in their own way. That's not typically a mark of disagreement (I don't think). But, you'd agree that if there's no disagreement between speakers in C1 and C2, there's weak evidence that the speakers assert logically incompatible propositions.

I'm not convinced that this response is right, btw, but the claim that these speakers are asserting logically incompatible propositions seemed to be an interpretation of the evidence that is underdetermined by the evidence. It seems to me that there are p.f. plausible ways of interpreting the evidence on which the speakers don't contradict and are both right in what they say.

You remarked, "I think my initial reaction would be to argue that, strictly speaking, participants responses actually were inconsistent."

I don't think that's an unreasonable initial reaction, but it does seem worth checking that reaction against further tests. I don't mean further surveys, but the sorts of tests that you'll find phils of language and linguists using to work out whether these terms exhibit any sort of context sensitivity.

[Jonathan Phillips]

Clayton,

I think your suggestion about context sensitivity is a really nice one. I've actually been thinking seriously about possibilities along these same lines myself. Specifically, it seems like one way we might allow for participants answers to be consistent would be to take 'force' as having an essentially modal component. (This is a view that is implicitly adopted by most philosophers who have talked about force, e.g., Aquinas talks about the necessity of coercion, saying that ‘When someone is forced by some agent…he is not able to do the contrary.')

If it is true that force has a modal component (or specifically an a-quantificational component like 'must') then it shouldn't be much of a stretch at all to think that 'force' would exhibit sensitivity to context. Of course, it would still be left to figure out exactly how differences in C1 and C2 (i.e., the shift in moral focus) would lead participants to make such opposing judgments. (And moreover why this differences doesn't arise when the forcee's action is not immoral.) But it definitely seems like a fruitful possibility to consider.

Another helpful issue, which you raised, was whether participants would take the two judgments to be inconsistent. I'm not sure how informative this really is, but we tried a version of this study within-subject, and once again got the effect. However, the effect was driven by only about third of participants who strongly showed the difference.

[jonathan weinberg]

That you didn't find it with the morally neutral materials seems pretty compelling here, in suggesting that it isn't that ambiguity at play. But if you did want to test for the performance reading anyway, perhaps just adding "successfully" would disambiguate? I.e., "Did the chief of surgery force the doctor to prescribe Accuphine?"

[Tim Dean]

I'd agree with commenters that respondents might not see it as a strict contradiction - or perhaps some of those who agreed to both did see it as a contradiction.

The questions asked have an ambiguity that could conjure up different senses. One reading of the doctor being forced is that the doctor was forced by the head of medicine. The other is that the doctor was forced to administer the drug.

The former is a statement about the act of the head of medicine, and you'd expect that to be more consistent with the responses to the question about the head of medicine forcing the doctor.

The latter is a statement about the doctor, and raises questions about moral obligation to follow orders over one's other convictions. Those with less moral respect for authority might end up giving you this apparent contradiction.

A few more iterations of the experiment with different wording and different scenarios could tease out these points.

I'd also be interested in variation. You're getting around 30% and 50% agreement. That's a lot of variability. How many of those who *disagreed* to one *agreed* with the other? Can you find a pattern there? Perhaps correlating it with attitudes towards moral authority?

[Bryony Pierce]

What if we take 'did x force y' to be synonymous with 'did x present y with any other options' and 'was y forced' to be synonymous with 'did y have any other options'? In a morally neutral case, there would be no need to include improbable or unreasonable courses of action (most other options will simply be ruled out because they are irrelevant/inappropriate/would have unacceptable consequences), but in a case where death could result, an obvious alternative option presents itself - to refuse, despite the anticipated consequences. So it's about how hard the person is expected to try to find other options, in view of the possible consequences, with moral decisions leading to greater pressure to seek alternatives. If participants think people can be forced to varying extents that would make it easier for judgements to shift between seeing it as forcing and not forcing, too.

[Jonathan Phillips]

Tim,

This is an interesting thought. It seems like you're suggesting that if we worded our questions differently, but still focused on the doctor we could have gotten a similar pattern of responses. So, what if we used these two questions instead:

- Was the doctor forced to prescribe the medicine?
- Was the doctor forced by the chief of surgery?

It seems like if we get the same pattern of responses as before, then this would be evidence against the best explanation being one of moral focus. I think interpreting this data would be a little tricky, but I guess there is only one way to find out what participants would say...