Doing and Allowing

People ordinarily distinguish between doing and allowing. They distinguish between 'breaking' and 'allowing to break,' between 'raising' and 'allowing to rise,' between 'killing' and 'allowing to die.'  A question now arises as to how people make this distinction. How do people know, e.g., whether a given act counts as actually breaking something or merely allowing it to break? 

Fiery Cushman, Walter Sinnott-Armstrong and I have a new paper on this question.  As you may by know have guessed, our thesis is that people draw the distinction in part by looking to the moral properties of the act in question.

Probably the best way to give you a sense for the idea here is just to describe one of the studies we conducted.  In our first study, all subjects were given a story about a person who ends up in the hospital.  The person is being kept alive by life-support systems, but then the doctor turns off these systems specifically for the purpose of making sure the person dies.  With the life-support systems now absent, the person's life soon ends.  Subjects are then asked whether it would be more accurate to say that the doctor 'ended' the person's life or that he 'allowed it to end.' 

Now comes the tricky part.  Subjects were randomly assigned either to the 'morally bad' condition or to the 'morally ambiguous' condition. Subjects in the morally bad condition were told that the doctor removed the life-support system because he despised the patient and did not want to use valuable resources on him; subjects on the morally ambiguous condition were told that the doctor turned off the life-support systems because he honestly believed that the patient would be better off not having to go on suffering.  (This latter condition is morally ambiguous in that those subjects who are in favor of euthanasia should regard it as morally good while those opposed to euthanasia should regard it as morally bad.) 

As expected, subjects in the morally bad condition tended to say that the doctor 'ended' the patient's life, while those in the morally ambiguous showed a more complex pattern.  Specifically, subjects who said that in general they regard euthanasia as morally bad tended to say that the doctor 'ended' the patient's life, while those who said that in general they regard euthanasia as morally good tended to say that he 'allowed' the patient's life to end.  These results suggest that people's use of the doing/allowing distinction depends in some way on their moral judgments.

In our actual paper, we mostly just present these results without offering much of a explanation, but we are very curious about how exactly one might explain the effect found here.  It seems like it might be helpful to think in a more general way about what the distinction between doing and allowing is all about and then to figure out how that distinction might relate to moral considerations.  Any suggestions?

Comments

[Chris Hallquist]

I'm surprised by the euthanasia connection. When I've heard ordinary people get into these disputes, it seems common for them to make a distinction between euthanasia and with withdrawing life support (I live in the US, if that makes a difference).

[Scott Hughes]

Wow, this is one of the most interesting things I've read in a while. As the study showed, I believe there is a correlation between people's distinction between doing/allowing and their moral judgments. However, I do not think it is logical to conclude that "people's use of the doing/allowing distinction depends on their moral judgments." It could be the other way around. Their moral judgment could depend on their distinction between doing and allowing. For example, people who think euthanasia is immoral may think that because they think euthanasia is ending someone's life (a form of 'doing'); whereas people who think the act of euthanasia is not immoral may think that because they think it is just an act of allowing someone's life to end.

[Anibal]

The direction of causation is not so explicitely spotted as Hughes is pointing.

But the paper is a new great step demonstrating the pervasivennes of moral judgment when navigating our social lives.

We now know that moral props shaped our cogntive representations.

The further question is what mechanisms build up our moral appraisals and how we sort out the events and situations into different psychological salient stimuli either action type, consequences, value, peronal/impersonal... and the unapproachable holy grail of subjects´subjective mental life.

Because if depending of how we regard euthanasia (or other situation) as morally bad or morally good we are predisposed to diferentiate between doing or allowing (perhaps is the other way around but the same issue arise in any case)...the really big question in ethics is then why people form very disparate moral appraisalls or see different moral properties to same situations (intentions are unobservable properties only directly and rightly percieved when philosphers create experiments)

[Fiery Cushman]

Thanks for these comments! A few thoughts in reply. Chris, you point out that there is a distinction between euthanasia and the removal of life-saving treatment. Perhaps the error is on our part when we wrote up the study for describing the case as "euthanasia". But what really counts for the interpretation of the results, of course, is the language we used in the survey. In Experiment 1, we asked people "In general, is it morally wrong for a doctor to remove treatment from a patient who wishes to die?" So the correlation we report is really between their judgment of a specific act of removing lifesaving treatment as doing vs. allowing, and their judgment of the moral status of removing lifesaving treatment in general. We referred to this as a a case of euthanasia in the paper, and as you say that wasn't as precise as it could have been, but the relevant correlation in stated directly in terms of removing life-saving treatments.

Anibal and Scott, each of you had questions about the causal inference we drew. Scott, your criticism of that causal inference in Experiment 1 makes a lot of sense, if we focus on the result that subjects' attitudes towards abortion correlated with their moral judgment of the doctor in the "ambiguous" case. But I'm not sure that it explains the difference in judgments of doing vs. allowing for individuals assigned to the "morally bad" condition versus the "morally ambiguous" condition.

As for the correlative analysis, the stronger result for our interpretation is in Experiment 2, where *first* we ask people for their general attitudes towards abortion, and *then* we ask them to judge whether Sarah made the fetus die or allowed the fetus to die. We could imagine three causal stories. First: judgments of Sarah caused their self-reported attitude towards abortion. But this is clearly a non-starter because they had never heard of Sarah when they reported their attitudes towards abortion. Second: attitudes towards abortion caused judgments of doing vs. allowing. This is the hypothesis we favor. Third, there is some third variable that both causes people to be pro-choice and causes them to view abortions as allowing events (but not because of their moral attitude). It's hard to imagine what sort of variable that would be, but we're very open to suggestions!

Anibal, I really liked your last point. One way of thinking about the impact of this work on how we think about moral disagreement is that, on the naive assumption that we make moral judgments by reasoning from concepts like intention, cause, action and omission, we might come to agree morally so long as we come to see eye-to-eye on the facts. But one of the alarming things about some of the new research going on in experimental philosophy is that our moral commitments cause us to see the very facts differently! This seems to make certain moral disagreements rather more intractable, doesn't it?

[Anibal]

Fiery,
so, paraphrasing Prehn et al. (2008)Individual differences in moral judgment competence influence neural correlates of socio-normative judgments, "Soc Cogn Affect Neurosci", Vol. 3, 1.
conclusion section:

"the question is not only wich processes are involved in moral judgment bu also how well a decision maker is able to integrate these different processes (e.g. emotional responses with rational reasoning processes) sensitive to the context of the particular social situation he or she faces..."

In this sense, your perosnal work that is cited in this article (Cushman, Young and Hauser 2006) and the new wave in the moral studies broadly construed (encompasing philosophy, psychology and neuroscience) what is highligthing is the role of individual differences in moral competence.

If this turn out to be correct, yes, this seems to make certain moral judgments really intractable!!

[John A.]

It seems to me that when there is a correlation between A and B it is reasonable to ask which influenced (caused) the other. It seems that the evidence could support either 'direction' of causation and I would think this would be problematic. Anyway, there appears to me to be some additional questions that need to be asked before any conclusions regarding the (causal) relationship between doing/allowing and moral judgments, etc. can be answered. For example 1) how were the subjects trained in ethics and ethical awareness (what sort of inventory of a subject’s basic moral beliefs was taken), 2) what metaphysical concepts regarding a teleological framework are in play within the subjects conceptual schemas, 3) have any subjects actually been involved in a situation where a person who was significant in their lives had life support removed or not removed, 4) do the (imagined) patient's desires regarding the continuation of their lives affect subjects responses? What I am suggesting is that isolating a particular distinction from within a person's entire conceptual schema may be problematic if the entire schema is at play when a person analyzes a moral issue. Anyway, I plant to read the entire paper next week and maybe some of these issues have been addressed.

[Weston]

Are you ruling out the possibility that the folk are simply confused, or careless, or are over-attributing responsibility out of moral disapproval run amok?

[Joshua Knobe]

Weston,

Yes, we are definitely worried about the idea that people's negative moral judgments are just running amok here, and we would be interested in any ideas people might have about how to get a better handle on what is going on here. One idea we had was to try changing the outcome to something good. For example, we could make it turn out that removing all life support systems somehow causes the patient to be miraculously healed and for his life to be saved. Then we could ask whether the doctor 'saved the patient's life' or 'allowed the patient's life to be saved.' My guess is that one would still see the same effect even if the outcome was changed in this way. That is, my guess is that people would be more inclined to say that the doctor actually 'saved the patient's life' in the condition where he does something morally *bad* than in the condition where he does something morally *good*.

It seems to me, however, that the obvious ways of thinking of this phenomenon in terms of 'moral disapproval gone amok' would not generate that prediction. Does that sound right to you?

John A.,

Of course, we are never going to be able to test everything at once, but if there is a specific worry about one or another of these elements, it would be really interesting to think about how we might go about trying to address it. We look forward to hearing your suggestions!

[John A.]

Hi Josh

I agree that we will not be able to test all beliefs at one time, but maybe there is a way to establish additional correlations that can lead to a causal explanation by beginning to map out the belief structure. For example, in your study of harming/helping the environment it was shown that people thought that if one harmed the environment it was an intentional result, but if the environment was helped that it was not intentional. Is there a correlation between the findings in this study and the one regarding doing/allowing? For example, do those who think that the agent is ‘doing’ something correlated to a belief regarding the ‘intention’ of the agent? If there is a correlation, then there maybe an additional belief (deeper in the conceptual schema) that is the causal explanation that explains this correlation.

Another thought is that there might be a correlation between one’s experiences of removing life support that impact one’s beliefs regarding both intentionality and doing/allowing. For example if a study was done between two groups, one which had personal experience with removing life support and the other not having this experience is there a difference in how these groups react to the doing/allowing distinction. Again, the results might point to a deeper belief (or set of beliefs) that has a causal role in how the perception of doing/allowing arises.

The same question applies to the extent that one has been formally introduced to ethical theory and practice and which theory the subjects seem to adopt. For example, do consequentialist have a different view of the doing/allowing distinction, then Kantians?

If these, and other questions, are asked and studied, it would seem that we could start to map out a moral terrain within conceptual schemas. Is there one underlying schema that all of us share, or is there a plurality of schemas, each having independent validity?

I should point out that I am using ‘causal’ to mean that it has explanatory force in helping us to understand why we think and act as we do.

[stephen]

Perhaps one link between the doing / allowing to happen distinction and the morally blameworthy / unblameworthy distinction can be seen in terms of counterfactuals. We are judged to allow x to happen rather than doing x, if x would have happened anyway, even if we were to do nothing to bring x about (although x might not happen, if were to do something to prevent it). So perhaps the Dr is said not to terminate the patient's life but to allow the patient to die, just in case the patient would have died anyway, or could have easily died anyway, even if the Dr were to do nothing to cause the patient's death. And that is exactly the case in which the Dr might be said not to bear moral responsibility for the death, because the death would have happened anyway. If the Dr had not caused the death, the patient's illness would have caused it. So the Dr is not to blame. Even though he pulled the plug, he ‘merely allowed’ a death to occur which would have occurred in any case.

That might go some way to explain why absolving the Dr of moral blame walks hand in hand with the idea of his merely allowing the death to happen.

If that’s the implied argument of the pro-euthanasiasts, then I think it has shortcomings. Mainly, it begs the question on intention and moral responsibility. It might be true that even if the Dr doesn’t pull the plug then the patient will die soon anyway. The patient might even die at exactly the same time and in exactly the same way that he would have died anyway. So the consequences of pulling and not pulling the plug are identical. Yet it could be said there is a world of difference between pulling a plug in order to cause someone’s death and leaving the plug alone in order to preserve his life, even granting that the consequences of both scenarios are identical. So the pro-euthanisasts argument above will work only if consequentialism is true. But it’s the truth of consequentialism that’s one of the issues at the heart of whether euthanasia is justified.

[John A.]

I was wondering would happen if the person on life support wanted the life support system turned off regarding the exmaple where the person turned it off because he wanted the person dead for the reasons you stipulated.

[stephen]

That's an interesting point.

One question here is: what intuition lies behind the blame attaching to the Dr in the case where he pulls the plug out of spite? Even if we knew the consequences for the patient would be identical - i.e., let's suppose that if the Dr didn't pull the plug out of spite then the patient would die anyway and just as quickly etc - even in that case I would guess (and I’m not sure whether the experimental results from the 'morally bad condition' as described in the o p back this up) that there would still be a tendency to blame the Dr. That's not a problem for a non-consequentialist, but it would seem to be a problem for the consequentialist, since the consequences of action and inaction are ex hypothesi identical.

So what's going on?

Again there is a difference in the counterfactual, though with a slight variation on the previous case. Perhaps the tendency to blame the Dr arises from the judgement that, even if the patient were by luck to survive the Dr's spiteful act, nevertheless the Dr *would have killed him if he could*. Even if the Dr's act were by luck to bring about some miraculous cure, that judgement would hold. (It's similar to the case where a gunman shoots at his innocent target, misses by accident, and instead hits a second would-be assassin, thus saving the target's life. The gunman's act is still perhaps blameworthy: it's attempted murder.) I think you have to go at least some of the way with non-consequentialism to sympathise with that judgement. But if we stick to consequentialism then we perhaps put ourselves in the experimental minority, since the majority seem inclined to blame the Dr in the case of spite but less inclined in the case of euthansia, regardless of at least similar consequences.

(Of course, the majority may be mistaken...)

I’d just like to add that I think the x-phi project is very worth while. People often argue on the basis of intuition and I think it’s a useful exercise to test whether the intuitions that we sometimes perhaps talk ourselves into having are actually held at all widely. I’d be fascinated to see x-phi applied to metaphysics – I mean, to what extent is the Common Western Metaphysic as outlined by van Inwagen actually common in the West? – we might be surprised… or not.. thanks for the blog

[John A.]

Joshua, Fiery, and Walter

The example from your paper: Dr. Bennett is an emergency-room physician. An unconscious homeless man is brought in, and his identity is unknown. His organ systems have shut down and a nurse has hooked him up to a respirator. Without the respirator he would die. With the respirator and some attention from Dr. Bennett he would live for a week or two, but he would never regain consciousness and could not live longer than two weeks.

Dr. Bennett thinks to himself,
Morally ambiguous case: “This poor man deserves to die with dignity. He shouldn't spend his last days hooked up to such a horrible machine. The best thing to do would be to disconnect him from the machine.”

Morally bad case: “This bum deserves to die. He shouldn't sit here soaking up my valuable time and resources. The best thing to do would be to disconnect him from the machine.”

Here are a few specific issues:
1) Why do you use the word ‘horrible’ to describe the machine in the Morally Ambiguous Case? Does the use of this word influence how the case is viewed? If the persons life could be saved and the quality of life restored would it still be a ‘horrible’ machine?’ Would the outcome have been different if that (or any) adjective had been left out?
2) If it were decided not to hook him up to the machine would the outcomes have been different? Why not ask if the patient should be hooked up in the 1st place and see if there is a distinction between doing/allowing regarding that aspect of the situation. If we do not hook him up are we doing something to end his life or allowing him to die?
3) You use the word ‘bum.’ What would be the outcome if we did not know this and restated the Morally Bad Case to be that the doctor wanted to take him off the patient simply because he did not want to waste the resources without reference to the type of person the patient was? Why is it a ‘morally bad’ case in the first place; because the doctor despises the type of person the patient is or because the doctor does not want to waste resources?
4) Assuming that there is something important in your findings (and I think there is, but this is not the issue), is this study sufficient to warrant any conclusion regarding the correlation between doing/allowing and moral judgments? What would happen if the situations were changed and tested against different types of patients; i.e., a child, a parent, a rapist, etc. I am assuming there should be a consistency of results between different cases if anything epistemically important regarding the role that doing/allowing plays in moral evaluations. This would include cases of removing life support outside of ER’s to include patients like the one discussed by Rachels.

[Joshua Knobe]

Thanks for all these helpful comments! Unfortunately, I won't be able to address all of them, but I did want to say something about one of John A.'s suggestions.

John points out that people's intuitions would probably have been different if we had given them a story in which the agent does not actually perform any action at all. Thus, suppose that instead of disabling the patient's life-support system, the doctor had simply chosen not to attach that system in the first place. In such a case, it seems that subjects would probably be more willing to say that the agent was 'allowing' someone to die.

So now we have two factors that make people more willing to classify a case as 'allowing.' They tend to classify cases as allowing (a) when the agent does what is morally right and (b) when the agent does not do anything at all. The question now is whether it might be possible to develop a single, unified theory that explains both of these effects.

My sense is that it is. Suppose we introduce the idea that people have a notion of *the way things naturally go*. Then suppose that people are more inclined to say that an agent merely 'allowed' someone to die when that person's death was part of the way things would naturally go. Both effects would then follow automatically if one assumes that people think that, other things being equal, it is more 'natural' to (a) do what is morally right and (b) not do anything at all.

[John A.]

Joshua
You write: "My sense is that it is. Suppose we introduce the idea that people have a notion of *the way things naturally go*. Then suppose that people are more inclined to say that an agent merely 'allowed' someone to die when that person's death was part of the way things would naturally go. Both effects would then follow automatically if one assumes that people think that, other things being equal, it is more 'natural' to (a) do what is morally right and (b) not do anything at all."

If this is the case and doing the right thing is in terms of what would happen (consequences)regardless of the motivation (intention) of the doctor, then would not one expect people to see the morally bad case as an example of allowing? It seems that the doctor in the morally bad case is doing the right thing but for the wrong reason if turning off (or not turning on) the life support system is the correct thing to do.

Do you think there may be a correlation between doing/allowing and deontology/consequentialism?

I guess what I am getting at is that there may be some value in relating examples of A and B to A* and B* and A** and B**, etc. to see if correlations do in fact hold across a number of similar, but significantly different, scenarios. Sometines is it not the correct thing (or at least understandable) to say that a person "deserves to die." Being a bum may not normally fit this evaluation, but what about a Hannibal Lecter? I think it would be very interesting to see if those who thought it was a matter of 'doing' in your morally bad case would still think it so in the H. Lector type case. And, if they did, would their explanations be the same. Intuitively, I am inclined to think that they would (sstill) think that the 'doing' in case A is wrong but that the 'doing' in case A* is the correct thing to do.

This might be an interesting experiment: Case A: HL is brought in and hooked up etc. (just like your morally bad scenario) compared to Case 2 where HL is executed by lethal injection. Both would be cases of doing, but would one be seen as better (morally correct)then the other? If so, what does this show about the role that the doing/allowing distinction plays in moral evaluation?

By the by, I think your Sarah Case does a good job of backing your thesis, but here there is an interesting confluence of metaphysical beliefs affecting moral beliefs. I do not see this confluence in the life support cases.

[Terry Tomkow]

I have a (rather long) post on this work up at http://tomkow.typepad.com/tomkowcom/2008/05/oh-agent-starli.html