By now you’ve probably encountered one of the various “do you push the fat man in front of the subway train to stop it from running over five innocent people trapped on the tracks?” puzzles that moral philosophers have been using lately to distinguish the consequentialists (who care only about ends) from the deontologists (who care also about means). (I invoked some of these puzzles, in a slightly non-conventional way, in Chapters 16 and 17 of The Big Questions.)
Now come psychologists Daniel Bartels and David Pizarro to report that choosing what they call the “utilitarian” solution (by which they mean the consequentialist solution, though one might have more faith in their scholarship if they’d used the accurate word) is correlated with higher scores on measures of psychopathy, machiavellianism, and “life meaninglessness”. Those who would push the fat man are more likely to agree with statements like “I like to see fistfights” or “The best way to handle people is to tell them what they want to hear” or “When you think about it, life is not worth the effort of getting up in the morning”. In other words, they are what we tend to think of as an unpleasant bunch.
But you see, here’s the thing: Bartels and Pizarro did the wrong test. Because in every one of the dozen or so moral dilemmas presented to the subjects, they asked: Would you shoot the crewmember? or would you execute one of the hostages? or would you flip off the switch to the ventilator? — which doesn’t get to the moral issues at all. The moral issue is this: In your opinion, should you shoot the crewmember or execute the hostge or flip the switch? The researchers never asked those questions, so we have no idea what the subjects’s opinions were.
Tell me a story about 20 children trapped in a burning building and ask me if I think I should rush in to save them. My answer is yes. Now ask me if I would. If I’m honest, I am very unsure I would do the right thing.
So what do we learn from this research? Here’s one interpretation: Maybe most people agree that you should push the fat man, and most people realize that they themselves would shrink from this moral duty out of squeamishness (just as most people agree that you should save the children from the fire, and most realize they might shirk this moral duty). Therefore they answer “No, I would not push the fat man.” But it’s only the psychopaths and Machiavellians who exaggerate their own moral virtues by claiming that they would rise to this particular occasion.
Do I know that that’s correct? No, but I’m betting it’s far closer to the truth than the researchers’ own favorite interpretation, which is that consequentialism is a manifestation of psychopathy (or something like that).
Of course even if consequentialism is a manifestation of psychopathy, the right conclusion might be that the world would be better off with more psychopaths. But I don’t think this study tells that, because I don’t think it tells us anything at all about who’s a consequentialist. It tells us only about who’s willing to exaggerate his or her own moral uprightness, and nothing at all about their morals.
I had the same thought reading a Dennis Prager column a long time ago:
“For 30 years, I have asked high school seniors throughout America which they would save first, their dog or a stranger. In every instance (except some religious schools), one third have voted to save their dog, one third for the stranger, and one third just didn’t know.”
If he wants to prove that religious schools provide better moral instruction, he’s asking the wrong question — he should ask “Who *should* you save first?” It’s possible that some people realize they *should* save the stranger, but they know that they *would* save their dog, so they answer honestly. Whereas maybe the religious students have been drilled to give the answer he wants to hear.
Since utilitarianism broadly construed is the most popular version of consequentialism (and since in this case the conclusion that a utilitarian would reach is the same as a generic consequentialist would reach) and since they appear to be publishing in response to, or at least prompted by, one or more papers referring to utilitarianism specifically, I am not sure it’s a fair criticism of them that they didn’t use your preferred word, since the solution is both consequentialist and utilitarian.
Maybe people who like fistfights would push the fat man off the bridge simply because they think it’s fun to push fat men off bridges.
I suspect that some deontologists are deontologists rather than consequentialists partially due to their sense of “squeamishness” which prevents them from imagining themselves acting like good consequentialists, even if deep down they think they *should*. So deontologists (like the study’s authors) also may be guilty of mixing questions of what they should do with questions of what they would do.
A critique I’ve heard of deontologists is that they are unduely concerned with “keeping their own hands clean” at the expense of great suffering of others (e.g. they don’t push the fat man because they don’t want to seem to have been “involved” in anyone’s death). But perhaps they are just more willing to give expression to this concern, although they unfortunately mix it up with the question of what is actually morally right.
In the case of psycopaths, it is not totally implausible that these people would lack the concern with “keeping their own hands clean” that others tend to have (and thus that the authors have got it partially right) — psycopaths generally aren’t as concerned with how they appear to the outside world as the rest of us.
Nah. Once you’ve tried both, you learn that the thrill of pushing tubbo off the edge is over pretty soon. A good fistfight, however, provides minutes of entertainment — plus long-lasting grudges. All in all, a better return for your efforts.
Damn, did I make a typo again?
Bennett – I think Prager would argue that someone who would do something that they should is more moral.
*ie more moral than someone who wouldn’t do something that they know they should do.
Ok, perhaps the study doesn’t demonstrate what the authors intended, but the authors found a correlation nonetheless. What significance can we glean from the fact that people who like to see fistfights are also more willing to exaggerate their “moral uprightness” (that is, their willingness to publically embrace an unpopular view in the interest of achieving desired outcomes)?
What conclusion should we draw from the fact that most people have an aversion to pushing the fat guy off the bridge, even if it would save lives?
A. The fact that most people share this intuitive sense demonstrates that there’s something innate and adaptive about this attitude. In the choice between our powers of rationality and the powers of evolution, we should humbly acknowledge that our intuitions are the product of millennia of natural selection and embrace them — even if we can’t articulate why.
B. The fact that most people share this intuitive sense demonstrates that human sensibilities evolved in a context that differs in meaningful ways from our modern world. To paraphrase Einstein, modernity has changed everything except our way of thinking. In the choice between our powers of rationality and the powers of evolution, we need to transcend intuition and embrace reason.
Psycopath says “At last, I get to push someone off a bridge! Just what I’ve always wanted!”
You can see why they came up with their conclusion.
This effect would remain if you asked “should”.
Therefore you should also ask “X is on trial for pushing a man off a bridge. His defence is that he saved 20 people. Do you convict? What is the penalty? Do you pull the lever?”
The psycopaths who will happily push a man off a bridge will also happily jump at the chance to condemn someone else for doing so.
Mike H: This is a great idea for a research project.
That distinction between “would” and “should” is pretty important–as Landsburg suggests, a lot of people will think they couldn’t bring themselves to do what they think they really should do, and reasonably so. Moreover, these questions often presuppose a false dichotomy: either one is obligated to push the fat man or one is obligated not to. I hear the results are a lot more varied and interesting when respondents are given options like, “It is morally acceptable either to push the fat man or not,” or “It is best to push the fat man, but not morally required.” A lot more people turn out to be permissive consequentialists than strict consequentialists.
Meanwhile, as for correlations between consequentialist answers and “higher scores on measures of psychopathy, machiavellianism, and ‘life meaninglessness’,” it would be interesting to see how these correlations compare to other findings. According to Josh Green’s experiments, using an MRI to observe people answering these questions, those who give a knee-jerk answer based on an emotional reaction tend to oppose pushing the fat man, whereas those who think about the situation before answering tend to favor pushing the fat man. And I’ve heard elsewhere (probably Jesse Prinz) that people tend to give more consequentialist responses right after experiencing something funny, while they are more deontological after experiencing, say, an uplifting story or image.
If psychopathy turns out to be positively correlated with thinking about a situation before leaping to a judgment, or with having a good sense of humor, then, as Landsburg suggests, so much the better for the psychopaths.
Is it up your alley, or more of a freakonomics thing? I’d love to see the results anyway…
One thing that bugs me about this whole thing is that I don’t see a truly substantive distinction between “consequences” (or “ends”) and mere “means”? Everything that happens in this universe is an event. What is the justification for grouping some events together and calling them “consequences” to the exclusion of other events, which are relegated to being “means”?
Couldn’t a person claim to be a consequentialist but still refuse to push the fat guy because the “consequence” he views as overwhelmingly important is the # of fat people pushed onto train tracks each year? Often, consequentialists are utilitarians — they try to maximize utility. Why can’t a consequentialist come along and say that the “consequences” he cares about are those things most other people regards as “means?” After all, a consequentialist should be allowed to pick and choose any *events* that take place in the universe and call those the “consequences” he cares about, shouldn’t he?
So then a deontologist who regards honesty as the most important thing could equivalently be called a consequentialist, where the consequence he is interested in is the ratio of # of truthhoods to falsehoods told.
In short, just about any type of moral philosophy boils down to some kind of optimization problem – couldn’t the thing being optimized be regarded as a “consequence,” no matter what it is?
@Jonathon, I agree, at least provisionally, that the distinction between “means” and “ends” could be a false dichotomy. I disagree that any type of moral philosophy boils down to an optimization problem, if by this you mean it becomes consequentialist.
After all, in the moral philosophy presented in the Bible (I mean, including the New Testament, if that is not clear) individuals are defined as moral or not based principally on the status of their relationship with the divine being. Their actions and attitudes (and the consequences of these) are a result of (and a reflection of) that relationship, not the other way round.
If you can cast this (not flippantly) as consequentialist, or even as an optimisation problem, I’d be interested.
@Jonathan- I’d have thought any value system can be described by a Utility function- even impredicative ones or ones which deny mind-body supervenience provided the domain is constrained by something like the axiom of choice. Similarly, virtue ethics of any sort can be turned into a constrained optimization problem provided we have omniscience.
The Bible question, raised by Mike H, is interesting because it shows we can say something similar about Consequentialism provided the range is unconstrained and can add states of the Universe not found in the domain. The example of Pinchas is canonical. He kills Zimri and Kosbi. This is a crime. But God is pleased with him and raises him to the status of Kohain. But, the Rabbis say this ‘halacha’ (rule?) is ‘vein morin kein’ (not to be taught- indeed, the rule, if known, prohibits the very action it enjoins if unknown!) This appears to strike at the foundations of Consequentialism. One way out is Occassionalism or Augustinian predestination plain and simple. Another, taken by the Hasids relies on metempsychosis- ‘ibbur’. They say that Pinchas, having killed a Prince of Israel, became terrified of the vengeance sure to befall him and so his soul fled him in fright. It was replaced by the souls of the two sons of Aaron (who were already Kohains) whom the Lord had previously destroyed. Thus, though Pinchas before and after the killing of Zimri is indiscernably identical in Leibniz’s sense, he isn’t actually. Additional states have been added to the Universe and so Consequentialism can get traction once again.
The subject of the post is very interesting but I’m afraid I don’t have access to the paper cited. I just don’t see how any experimental result they got could support the conclusion they give. As Prof. Landsburg points out, swapping ‘would’ for ‘should’ is necessary to find out about norms.
Mike H – First consider the case of a moral philosophy that regards one’s moral standing as primarily a function of the status of his relationship with a mortal being, e.g., his wife. But how is a “relationship” evaluated? It must be evaluated based on the nature of the physical *events* that define that relationship, e.g. the way in which neurons fire in the brain of the husband and wife when they are together, the amount of dopamine that is generated by the relationship, etc. Is there any data, other that physical data, that could be compiled to describe this relationship? If no, then the consequentialist could simply describe the consequences he’s interested in in terms of that data.
When we move to the supernatural realm, we can’t rely on “physical” events only, since theists claim that relationships with god transcend the physical. But still there must be *some* way of actually describing the way that the relationship with god is appraised, analogous to the way a relationship in the physical world is appraised – e.g. perhaps it involves things like angels and demons and spirits – why can’t the consequentialist just explicitly say what those things should do (i.e. what the “consequences” should be) for a relationship to be “good” and thus for him to be happy with the world?
@Jonathan- ‘But how is a “relationship” evaluated? It must be evaluated based on the nature of the physical *events* that define that relationship,’- Why must it be so evaluated? If a description of physical events necessarily includes reference to possible states of the system, and if, moreover, on the evolving multiverse hypothesis, ‘impossible’ universes too must be considered- then no Occam’s razor or Canon of Parsimony militates for a theory of strong supervenience of intentional & phenomenal states on physicalism.
You write ‘why can’t the consequentialist just explicitly say what those things should do (i.e. what the “consequences” should be) for a relationship to be “good” and thus for him to be happy with the world?’ The problem here lies in the nature of intentionality. Happiness may have strong supervenience- i.e. such and such physical state equals happiness- but happiness WITH THE WORLD is intentional in a radical way. Essentially you’re saying there’s a set of brain states with the property ‘happy’ and some proper subset with the property ‘happy with the world’. However, for a virtue ethics which says ‘if I am happy and happy with the world then I am damned so long as I believe the happiness of even one being in the world is at hazard. Instead of being happy, I should be working my ass off to fend off the risks to that beings happiness’.
In Buddhism, we have the Boddhisatva who postpones his ‘happiness with the world’ till everybody gets to be ‘happy with the world’. But , so long as one Boddhisatva exists, no Boddhisatva can be happy with the world. Buddhism gets round this by a doctrine of ‘kshanikavada’- momentariness- but a Consequentialist Physicalism cant.
I don’t know of any way where you could have strong supervenience of the sort you describe and still deal with impredicative values like the above type of virtue ethics.
The interesting thing about these dilemas is that responses change with no apparent change in the outcome. Thus nearly everyone thinks you are justified to flip a switch to put the train on a track with one rather than 5 victims, but far fewer think you are justified to push the fat man in the way. The outcomes are the same, but peoples’ responses are different. Therefore peoples’ approach to moral issues is not consequentialist because they change according to the means rather that the outcome.
Thus the study mentioned should have chosen whether a person was cosequentialist or not based on whether their position changed from a switch to a fat man. Whether “would” or “should” was asked would not matter, as both would reveal a change.
We have two origins for morality, instinct and reason. Instinct works through the mediation of our emotion, often via empathy. The emotional response is stronger for direct actions and for those close to us. I find it entirely consistent that those with a cosequentialist approach have higher scores for psycopathy, since psycopathy implies lower empathy and lower emotional response to others. If one accepts the consequentialist’s contention that their approach is more moral, then psycopathy would lead to more moral outcome in this situation, where the consequences for ones-self are negligable either way. However, psycopathy also leads one to make choices that most would consider immoral when the personal consequences are greater.
In a slightly similar situation, depressed people are much more realistic in estimating their relative competence and abilities. However this greater realism dos not do them much good. We succeed more if we are slightly deluded. Possibly society succeeds more if individuals’ morals are not entirely consistent.
Interesting study that may sheds some light
http://experimentalphilosophy.typepad.com/experimental_philosophy/2005/09/the_trolley_pro.html
– with the “flip a switch” scenario, 34% thought they personally were obliged to flip it, 47% thought an innocent bystander was. 78% thought they would be justified in flipping the switch, 94% thought an innocent bystander would be justified.
Further evidence that economic libertarians are correlated with a high score on the “dark” triad personality traits (D3): Machiavellianism (i.e. tendencies to deceit), narcissism (overinflated sense of self-worth), and psychopathy (lack of guilt and remorse).
http://experimentalphilosophy.typepad.com/experimental_philosophy/2011/08/update-somewhat-better-news-for-conservatives-but-worse-news-for-economic-libertarians-death-penalty.html
They conclude:
” We observed systematic positive correlations at a .01 level of significance between three “dark and anti-social” personality traits and (A) the libertarian judgment that government should never intervene in free markets except to punish or prevent force or fraud, (B) conservative views on the death penalty, and (C) “neo-conservative” positions on war”
I think the correlation could be a one-way thing. Those with D3 traits may tend to these views, but those with these views do not necessarily have D3 traits.
Utilitarianism / consequentialism may stem from low empathy, but it can also arise from greater empathy than normal, i.e. you empathise with everyone rather than those more immediately conected with us. Thus most people would not push the fat man off the bridge because they empathise strongly with him. The psycopath would, because he dos not care about the fat man. The highly empathic would, because they empathise just as strongly with the 5 victims.
Harold:
Utilitarianism / consequentialism may stem from low empathy, but it can also arise from greater empathy than normal, i.e. you empathise with everyone rather than those more immediately conected with us. Thus most people would not push the fat man off the bridge because they empathise strongly with him. The psycopath would, because he dos not care about the fat man. The highly empathic would, because they empathise just as strongly with the 5 victims.
Yes, I absolutely agree with this.
This reminds me of a study a few months back conducted by a Nobel prize winning physicist. The study was published in either Science or Nature and was shown to be so poorly conducted that the results were of no value. It had to do with determining how to get physics students to learn more efficiently, but the researcher forgot to randomize treatment etc.
Can any one recommend any good works on experimental design?
Frankly I hate all these kinds of quizzes masquerading as serious studies. As Steve points out a simple change in the question will change the answers. So will changing details in the framing. So you really should make a quiz with every such variant. And you cannot.
(Here’s an example. The fat man is nake and sweaty. You must lay hands on him … Or the fat man just gave a child a dollar. Or …)
Plus extrapolating from the reponses NOW — even cross culturally — is a dubious way to get results you maintain are FOREVER.
And of course the revealed/expressed thing too.
Steve asked for a list of things we will look back on and shake our heads over. I nominate this kind of social science.