It’s no secret I was excited by word of Jonathan Haidt’s latest publication. His work with Joshua Greene first caught my attention during a law school course in Health Law, where it made an appearance in my critique of the wisdom of repugnance promoted by Leon Kass. Their research introduced me to moral psychology, neuroethics, and experimental philosophy; thanks to their work, the political and legal ramifications of moral disgust and moral intuition has since become one of my primary research interests.
So today, I dived into Haidt’s new paper expecting something great. I got it. You can too, although it is locked behind a subscription.
- Jonathan Haidt. The New Synthesis in Moral Psychology Science 18 (May 2007) 998-1002.
Rather than covering Haidt’s review in detail, I’ll distill three insights from the paper that make a theoretical foundation for any further work in the fields of moral psychology and moral ideology.
- Moral reasoning is driven by affective responses like moral disgust. Gut reactions are immediate, and bias or distort subsequent processes of reasoning and justification.
- Humans have five classes of moral intuitions, which each have unique evolutionary origins in the behaviour of individuals within gossiping moral communities. These are intuitions about…
- protection from harm – altruism
- fairness – rights, reciprocity, and justice
- loyalty – ingroup and outgroup relationships
- authority – respect and obedience
- purity of body and spirit – sanctified, not carnal
- In Western societies, conservatives make use of all five intuitions, but liberals pay more attention to harm and fairness. (Coincidentally, most research so far also focuses on harm and fairness – is research being driven by liberals, or are these easier to study?)
Moral reasoning is often like the press secretary for a secretive administration—constantly generating the most persuasive arguments it can muster for policies whose true origins and goals are unknown.
This is a rather dense package, so I encourage you to read the paper to unpack it.
The one thing that should be said about Haidt’s theory, and his review, is that there is to date very little empirical data to back it up. In the review, you’ll notice he sites only a few recent peer reviewed articles on moral judgment, and instead of publishing data-driven articles testing the hypotheses of his social intuitionist model, he’s spent most of his time writing reviews and theory articles for edited books. It’s good that people are taking emotion seriously in moral psychology, but at the moment, most of Haidt’s claims are largely speculatively.
Chris, you make an excellent point, although I beg to differ regarding the question of currency. Old data can still be good data.
There are two categories of evidence being used in this type of discussion: hard science from fMRI laboratory work telling us about moral reasoning in the brain, and soft science involving cross-cultural surveys of moral intuitions. The latter is common in social psychology (see, for example, Paul Rozin’s work on disgust), but I agree it needs more rigorous follow-up.
I’ve noticed that Haidt relies on 3 of his articles to back up his claims about the 5 moral intuitions. 2 of them are forthcoming (I’m not a fan for forward-looking citations) and the other is:
Haidt, J., & Joseph, C. (2004). Intuitive Ethics: How Innately Prepared Intuitions Generate Culturally Variable Virtues. Daedalus, pp. 55-66, Special issue on human nature.
I suspect none of these would provide the evidence you wish. What sort of data would you find convincing?
Corey, that’s a good question. I tend to find the imaging studies fairly unconvincing, in part because the stimuli themselves are so limited, and in part because there are serious issues related to the interpretation of imaging data. I’m also skeptical of making general claims about moral intuitions and moral judgments from dilemmas, particularly dilemmas that are different in a wide variety of potentially important ways (e.g., the trolley and footbridge dilemmas). The biggest problem with dilemmas, though, is that they’re outrageous and require people to place themselves in highly unfamiliar situations.
Right now, I’m conducting a study looking at the role of agency in moral judgment. I’m using dilemmas that are based on work on taboo tradeoffs (e.g., Phil Tetlock’s stuff), along with fairly normal situations (driving a car) to look at moral decision making with varying degrees of personal agency. In the future, I’m planning on using both types of situations (dilemmas and everyday scenarios) to test the role of different affective and motivational states on moral judgment (and the interaction between these and agency). I’m also looking specifically at the effect of unfamiliar or unusual actions on moral decisions, in the hopes of shedding some light on the generalizability of much of the contemporary research in moral psychology (which relies almost exclusively on dilemmas). Naturally, I don’t think my own studies will provide all the answers, but well controlled laboratory studies that produce behavioral data (as opposed to equivocal imaging data) are the necessary first step, and to date I don’t think there’s been much of that (particularly with respect to the “controlled” part).
By the way, the Haidt and Josephs paper presents the data in a very incomplete way. The rest of it is supposed to be forthcoming, but while I was initially excited by Haidt’s work, I’m becoming more and more disenchanted with it as the claims continue to outpace the data.
Also, in a smilar vein, I think Hauser’s quote in this month’s Discover magazine, to the effect that at the moment there really isn’t any clear evidence for innate moral intuitions (Hauser being one of the most vocal proponents of theories of innate moral intuitions) pretty much sums up the field right now.
Chris, I had to go back to the Discover interview with Hauser to check this, but I think you have taken his comment out of context.
He is asked by interviewer, Josie Glausiusz: “What is the evidence that infants already have a moral code ingrained in their brains?”
He replies: “I don’t think we’re ready to say…”
… and follows that with a discussion of two studies – one with infants, another with toddlers.
So, we don’t have much data about young children, above and beyond that they seem to be able to distinguish moral norms from non-moral preferences. This doesn’t mean, however, that there isn’t good evidence about moral thinking in adults. Earlier in the interview, he talks about trolley problem research when asked, “What is the evidence that we draw upon unconscious principles when making moral decisions?”
From this reading, I doubt Hauser would agree there is no evidence to support his view that we have an innate moral faculty.
I’m looking forward to seeing your behavioural work on ethical decisions taking place in familiar circumstances. What are the sorts of normal, everyday decisions which interest you? Promise-keeping? Petty theft? Reactions to begging? Cheating on exams? Using a blue box? Witnessing gender discrimination? Will you include policy decisions about such familiar controversies as gay marriage, abortion, euthanasia or global warming within the scope of ‘ordinary’ decisions?
More generally, what are the grounds for your suspicion that people treat ‘ordinary’ moral decisions any differently from ‘outrageous’ moral decisions? Since people consume so many dilemmas via fiction and news, I’m not sure they don’t form moral opinions about astonishing things on a regular basis.
A conversation with a clinical ethicist might expand notions of what counts as ‘normal’, because hospital wards are common locations for agonizing moral decisions. It might also offer some good data – one benefit of the medical context is that it is likely to be highly documented.
Best of luck with your research. Its fascinating stuff, and deserving of study.
Hey Corey, almost forgot about this little discussion, so I’m late getting back to you. I first developed suspicions about the problems of outrageous moral dilemmas (e.g., the footbridge problem) when I started asking for feedback on them. People were genuinely confused about how to represent them, and weren’t happy with the options — throwing the guy off vs. not throwing the guy off, with no option for, say, throwing yourself off, or remarking, as many subjects do, that a body is unlikely to stop a trolley, or wondering why the hell the people can’t get off the tracks in time. It became clear to me that in these situations, participants didn’t really know what sorts of frames, moral or otherwise, to bring to bear on the problem, and they felt constrained by the choices.
The second source of suspicion came from what I know about cognition generally. In other domains (i.e., non-moral domains), much of our automatic (which is to say, “intuitive” in Haidt’s sense) reasoning is case-based. That is, when we encounter a situation, we interpret them and react to them on the basis of pre-existing schematic representations of past instances. These representations tend to be pretty specific, either in the sense that they represent a particular instance or that they represent particular features of a class of instances. Said another way, they’re not abstract. So for moral dilemmas to activate the sorts of schemas that would elicit automatic/intuitive responses, they’d have to activate features of those schemas. The outrageous moral dilemmas only map onto experienced moral situations in an extremely abstract way (e.g., sacrificing something sacred for something else that’s sacred). As a result, it’s not clear to me that the post-hoc reasoning that Haidt theorizes (though he hasn’t provided any direct evidence that the reasoning is post-hoc) is likely not the result of subjects interpreting their responses without access to the intuitive processes that caused them, but instead are the result of subjects trying to figure out what the hell they’re supposed to say.
And the third source of my suspicion is more methodological. The contrast conditions (personal vs. impersonal dilemmas, say) tend to differ on a wide range of dimensions. The trolley and footbridge problems are so different, in fact, that interpreting differences in responses to them is damn near impossible. More mundane scenarios allow for more control, in general.
So that’s why I’ve started using them. Right now we’re using scenarios that are both mundane and not so mundane (driving, and avoiding pedestrians). But we’re currently developing scenarios that involve some of the dilemmas you mentioned, including cheating, lying, and theft, as well as finding money. We’re also heavily relying on the taboo tradeoff literature I mentioned earlier.
As for Hauser, you’re probably right. Given his recently published study on moral dilemmas (with, if I recall correctly, 50+ scenarios), and his ongoing research, I suspect he believes there is evidence for innate moral modules. But given my reservations with regard to highly unusual moral dilemmas, I’m not sure any of that evidence is worth much. How, for example, would outrageous moral dilemmas activate innate modules? Through abstract similarity? How the hell would such abstract representations evolve? This is a problem that social exchange theorists have been unable to solve for a few decades now, and I don’t see it being any less of one for the moral nativists.
[...] and I have poked at the idea of the psychology involved in every-day moral decisions. This is most certainly not one of [...]
[...] on his 5 foundation theory (mentioned here), Haidt argues liberals need to take into account the 3 extra moral ‘foundations’ which [...]