Folk Meta-Ethics

There's a really interesting paper by Geoffrey Goodwin and John Darley in press at the journal Cognition on the subject of lay meta-ethics, and ethical objectivism specifically. That is, the paper explores the question, "How do lay individuals think about the objectivity of their ethical beliefs?" (from the abstract). The paper contains a ton of data, and I couldn't possibly do it justice in a blog post, but unfortunately, there's no free version online (if you have a subscription, you can read the paper here). So you'll have to do with my incomplete discussion of it.

The paper is interesting because it is the first that I know of to take a direct look at how people view their own moral beliefs. There's plenty of literature on how people make ethical decisions (I've discussed some of it on this blog), of course, but it's also important to understand their meta-ethical beliefs, because those will have an impact on ethical disagreements. And to date, our understanding of those meta-ethical beliefs is pretty much non-existent. So Goodwin and Darley conducted a series of what are essentially exploratory studies, focusing on objectivism -- the belief that ethical beliefs have some factual, extra-mental basis.

Their first study, which I think is the most interesting, involved giving 58 Princeton undergrads twenty-six statements and asking them to indicate how much they agreed with them (on a 1 to 6 scale, with 6 being the highest level of agreement). The statements were of four different types:

• Ethical statements: E.g., "Robbing a bank in order to pay for an expensive holiday is a morally bad action," "Anonymously donating a significant portion of one's income to charity is a morally good action."
• Factual statements: E.g., "Homo sapiens evolved from more primitive primate species," "Mars is the smallest planet in the solar system."
• Convention statements: E.g., "Wearing pajamas and bath robe to a seminar meeting is wrong behavior," "Talking loudly and constantly to the person next to you during a lecture is permissible action."
• Taste statements: E.g., "Shakespeare was a better writer than is Down Brown," "CNN provides better news coverage than does FoxNews."

After each statement, they were also asked whether they considered the statement to be a true statement, a false statement, or an opinion or attitude. This question was one of two measures of their level of objectivism.

After reading and rating all 26 statements, participants were given two ethical statements, and one each of the other three types of statements, and told that a previous participant in the experiment disagreed with them with regard to each. They were then asked to circle an explanation for the disagreement. Either the previous participant was "surely mistaken," it was possible that both or neither are mistaken, the current participant is mistaken, or "other." Indicating that one or the other is mistaken would be evidence of objectivism, because if a statement is objectively true (or false), then disagreeing (or agreeing) with it is a mistake, whereas if it's not objectively true (or false), then disagreement is simply the result of differences in opinion.

Finally, at the end of the study, participants were asked to indicate how they justify their moral beliefs. They could choose from the following justifications: "ordained by a supreme god," "every good person on earth, regardless of culture, holds these beleifs," "a society could not survive without its citizens holding these beliefs," "their truth is self-evident." Each participant could pick as many of the different justifications as he or she wanted.

Despite some of Goodwin and Darley's summary descriptions, the results of this first study are anything but straightforward. For the factual statements, between 90 and 98% of the participants indicated that each was either true or false, except in the case of the evolution statement (included in the examples above), in which case 28% indicated that it was a matter of opinion (4% said it was false!). For the statements of taste, between 78 and 98% of the participants indicated that they were matters of opinion, depending on the statement. So with these two types of statements we essentially have the two extremes of objectivity -- factual statements are the most objective, while statements of taste are the least objective, though it's important to note that in some cases, up to 21% of the participants believed the statements of taste were objectively true.

For the statements of convention, participants' responses seems to fall between the levels for facts and taste. The mean proportion of objectivist (true or false, not opinion) responses, across the convention statements, was 55%, which is right between the means for factual statements (90%) and statements of taste (8%). The range of true or false vs. opinion goes from 12% to 82%, indicating that participants believe some conventions (running a red light, e.g.) are objectively true, while others are mostly seen as matters of opinion (e.g., the pajamas statement used as an example above) . It's not clear from the data alone what factors influence whether people consider a particular convention to be objective though.

When it comes to ethics, though, things are rather messy. To give you an idea of what the data for ethical statements looks like, I created this graph using Goodwin and Darley's Table 1 (p. 9):

The largest proportion of participants who indicated that a statement was true or false (true, in this case) was 68%, for a statement about shooting a gun in a public place. For all participants, the objectivist answers (true or false, and not opinion) ranged from 6% to the public shooting's 68% across the different statements. The mean percentage of objectivist responses for ethical statements was 38.4, but since the proportions clearly aren't normally distributed (they almost look bimodal, but I suspect it's even messier than that), the mean doesn't, in fact, mean very much. What's most interesting, I think, is that the most controversial moral issues (and they are undoubtedly moral issues), like abortion, euthanasia, and stem cell research, get the highest proportions of non-objectivist (i.e., opinion) responses: 92%, 90%, and 94% respectively. It's interesting to note that these three also had the lowest average agreement scores (4.12, 4.36, 4.58 respectively, on a scale of 1-6) of all the ethics statements as well, consistent with them being controversial. Though Goodwin and Darley don't mention it, it seems that prior exposure to disagreements on ethical issues could make people feel that their beliefs about those issues are less objective.

Goodwin and Darley argue that these results (and results from the subsequent studies, which looks pretty much like this data) show that people treat ethical beliefs as less objective than factual ones, but more objective than beliefs concerning taste, but I don't think the data actually show that. Instead, it looks like some people think a few particular ethical beliefs are objectively true (or false), like facts, while others treat those same beliefs more like matters of taste. Other ethical beliefs, however, are treated by just about everyone as being as objective as matters of taste, not facts. Whether people are objectivists about ethical beliefs then seems to depend heavily on which people we're talking about and the content of the particular ethical belief.

Which brings us to the data indicating who the objectivists (for some ethical beliefs) are. First and foremost, they are the people who "grounded their ethical beliefs in the notion of a divine being" (p. 11), but to a lesser extent, people who gave the pragmatic justification (the one about society, above) were also more objective than those who didn't. The same was true of those who believed that "every good person" shares their ethical beliefs. So religion, pragmatism, and self-identity seem to lead to objectivism about certain ethical beliefs.

Goodwin and Darley's third experiment (the second was just a replication of the first using slightly different questions) explored the religion-objectivism connection in more depth. Looking at religious and political beliefs, they found that religion, but not politics, was associated with greater ethical objectivism. Participants were asked to rate the how much religion justified their ethical beliefs, and the higher the rating, the more objectivist their responses.

In sum, then, there don't appear to be very many (if there are any) ethical beliefs for which everyone believes there is an objective justification. Instead, for some beliefs, a majority of people believe they have an objective justification, but for others, very few believe that to be the case. It would seem, then, that while some people are more objectivist than others, few if any are objectivists about all ethical beliefs. Religion is the most reliable predictor of objectivism, though even it can't turn people into objectivists with regard to all ethical beliefs. Pragmatism and "self-identity" also lead to some level of ethical objectivism.

Of course, the fact that they only used Princeton undergrads limits the generalizability of these results, but it seems like a good first look at an important issue. Goodwin and Darley focus on the fact that the average level of objectivism for ethical statements is between fact and taste (and about the same as convention), but I don't think that's the least bit interesting really, particularly since "average" here is pretty meaningless. What I find most fascinating is that nobody appears to be consistently objectivist, and that very few of the participants were objectivists about controversial moral beliefs. That last part is particularly important since those are the beliefs that result in contentious debates in politics and society in general. So I hope that future research looks at when people are likely to be objectivists and when they're not, why that is, and what the implications of objectivism or subjectivism are for ethical debates.

You may be wondering why I haven't talked about how people answered the question about disagreements. And I should say something about those answers, because Goodwin and Darley used them to come up with their interpretation of the results. However, it's difficult to say what people's responses to the disagreement questions were, because Goodwin and Darley don't present that data. Instead, they compute a "three-point scale of objectivism" by combining the true-false/opinion answer with their answer in response to being told that previous participants disagreed with their ethical beliefs. A score of 3 was assigned if participants indicated that either the previous participant or they themselves were mistaken and said that the statement was either true or false; they got a two if they either indicated that no one need be mistaken or that the statement was a matter of opinion, but not both; and a score of 1 if they indicated both that no one had to be mistaken and that the statement was a matter of opinion. So a 3 means most objective and a 1 means least objective.

According to this scale, participants treated ethical statements as more objective (m = 2.56) than either conventional (2.00) or taste statements (1.34), but less so than factual statements (2.91). But they don't break this down by particular statements, so it's hard to get a real sense of the pattern. Furthermore, while Goodwin and Darley do tell us that exactly half of the responses (out of 100 total) to the ethical statements yielded objectivism scale scores of 3, this can only be the case if they chose ethical statements with relatively high amounts of true-false responses, rather than opinion responses, because over all, participants indicated that ethical statements were true or false only 39% of the time. The appendix confirms this, because in it they list the percentage of participants who received the different ethical statements in the "disagreement" phase of the experiment. There we find that cheating on an exam (84%), robbery (60%), and shooting in public (36%) are the most frequently used statements. These are also the questions that got the highest percentage of true or false responses (see the graph above). Statements that got true or false responses under 50% of the time (5 of the 9 statements!) were used only 12% of the time in the disagreement phase, and those with under 40% (4 of 9) were used only 2% of the time. In other words, we don't have a good sense of the real distribution of scores on this scale, because they didn't choose a representative sample of the statements (the abortion and stem cell questions were never used!). Put simply, the 3-point scale is worthless, so we should assess people's level of objectivism using the true-false vs. opinion data only, and that data doesn't support Goodwin and Darley's interpretation.

Goodwin, G.P., & Darley, J.M. (In Press). The psychology of meta-ethics: Exploring objectivism. Cognition.