Watkins and the persistent (and probably false) claim that most people are moral realists
Twitter Tuesday #7
Image by DEZALB from pixabay.
This week’s Twitter Tuesday is a response to Benjamin Blake Speed Watkins. Watkins echoes a sentiment that persists among philosophers that ordinary people are disposed towards moral realism. I’ve criticized this claim numerous times on this blog, and I will continue to do so as long as (a) the evidence continues to cast doubt on these claims but (b) people keep making them.
Watkins made the following remark on 4/24/2023:
Most people have a rational intuition that our basic moral judgments can by mistaken [sic] and that moral progress is possible through moral reforms. Both of these rational intuitions involve a conception of objectivity that only moral realism can plausibly accomodate [sic].
One respondent replied, quite reasonably:
The first sentence is an empirical one that can be subjected to scientific inquiry. What's the evidence for it?
To which Watkins responded
The empirical evidence is our common moral discourse- whether it be about individual acts that we object to others doing (e.g. lying is wrong) or to political policies. For example, most Americans believe it is morally better that slavery was abolished than if it had continued.
This is not a very convincing response for several reasons.
First, “the empirical evidence is our common moral discourse” is, at best, ambiguous. Suppose I made the following claim:
All tigers like to eat salmon.
If you asked me what my empirical evidence for this was, I could make a claim similar to Watkins’s claim about people’s intuitions:
The empirical evidence is the dietary habits of tigers.
This claim is, at best, ambiguous. The dietary habits of tigers is precisely the kind of information one would need to document, hopefully in a scientifically credible way, if one wanted to provide evidence that all tigers like salmon. Yet it’s strange to say that the dietary habits of tigers is the evidence. The dietary habits are only evidence in principle, not evidence in practice. The empirical studies one conducts that systematically document those dietary habits is the evidence in practice…that is, the evidence that actually matters. Watkins has failed to disambiguate what kinds of information could as evidence in principle from the actual presentation of evidence.
To illustrate just how silly this is, consider the following example. Suppose a prosecutor asked what the evidence was that a particular defendant committed a murder, and the detective replies:
Detective: The murder weapon is the evidence.
Prosecutor: Great. You have the murder weapon? Where is it?
Detective: Oh. No. I don’t have the murder weapon. But the murder weapon is the evidence.
Prosecutor: …Sorry, what I mean is: can you please present evidence that the defendant committed the murder? I’m not going to agree to serve as a prosecutor if there’s no evidence.
Detective: Yes, of course. The evidence is the murder weapon.
Prosecutor: Okay, well, where is the murder weapon?
Detective: I don’t know.
Prosecutor: Alright, well, what is the murder weapon? We can go look for it.
Detective: I don’t know. We don’t know how the murder was committed and didn’t recover any weapons at the scene of the crime. But whatever weapon they used to commit the murder, that would be the evidence. So, my job here is done.
As you can see from this exchange, the detective appears to misunderstand what they’re being asked for. They are conflating two senses in which something could “be evidence”:
Evidence in principle: what kinds of data could serve as evidence if they were made available.
Evidence in practice: what kinds of data serve as evidence when they are actually available.
The detective is discussing (1), while the prosecutor is interested in (2).
Perhaps Watkins is not distinguishing evidence in principle from evidence in practice. However, it looks from Watkins’s subsequent remarks that Watkins is laying claim to (2) based on…well, I don’t know. No studies or data are presented, just claims about what most people supposedly think. This is, at best, speculative armchair psychology. That is, Watkins is making claims about what most people think, which is a psychological claim. Claims about what most people think are empirical questions, and such questions are difficult to answer even when one systematically gathers and evaluates data from thousands of people. Why? For many reasons, including considerable difficulties in ensuring the validity of one’s methods, and because findings in the populations you sample may or may not generalize to the rest of the world’s population. Yet Watkins doesn’t even have psychological data gathered in accordance with conventional scientific standards, e.g., carefully designing stimuli, surveying large numbers of people who represent populations of interest, and so on.
Note that Watkins says:
For example, most Americans believe it is morally better that slavery was abolished than if it had continued.
That’s probably true, but would that be evidence that people have the “rational intuition” that “our basic moral judgments can be mistaken” and that “moral progress is possible through moral reforms”?
Several elements of this remark are ambiguous or unclear. What’s a “rational intuition”? Is that a psychological state? If so, what kind of state is it, and how do we know people have these states, much less that most of them do? Second, the notion that “our basic moral judgments can be mistaken” is ambiguous. Does this mean “mistaken with respect to stance-independent moral standards,” or “mistaken in a way consistent with moral relativism or non-realist constructivist accounts”?
One couldn’t provide empirical evidence for a claim if that claim is ambiguous, since it would be unclear which specific operationalization of the claim one is referring to, and evidence for one operationalization isn’t necessarily evidence for another. If, for instance, Watkins means that people think people can be mistaken about moral issues in a realist sense, the evidence for this would be different from the evidence that people think people’s moral judgments could be mistaken in a way consistent with both realism and antirealism. Evidence for the former would be more difficult to obtain, since the claim is more narrow. Finally, the notion of moral progress is also ambiguous: progress in a relative or nonrelative way (if “non-relative progress” isn’t conceptually confused at the outset), or is this notion of progress consistent with either? As with “mistaken,” these remarks are too ambiguous to be subjected to empirical evaluation to begin with.
However, it does appear that Watkins has in mind moral realism, since the post was in response to this question:
What's the main reason some people insist that objective morality exists?
This exposes Watkins to two lines of objection. First, there has been empirical research on what nonphilosophers think about moral progress.. Pölzler, Zijlstra, and Dijkstra (2022) recently conducted a series of studies to address whether nonphilosophers believe in the possibility of objective moral progress. The results do not support the claim that people believe in the possibility of “objective,”(i.e., stance-independent) moral progress (though this is consistent with the possibility that people believe progress is possible relative to the standards of individuals or cultures, so it is not evidence against all conceptions of moral progress). Here is the abstract:
A prevalent assumption in metaethics is that people believe in moral objectivity. If this assumption were true then people should believe in the possibility of objective moral progress, objective moral knowledge, and objective moral error. We developed surveys to investigate whether these predictions hold. Our results suggest that, neither abstractly nor concretely, people dominantly believe in the possibility of objective moral progress, knowledge and error. They attribute less objectivity to these phenomena than in the case of science and no more, or only slightly more, than in the cases of social conventions and personal preferences. This finding was obtained for a regular sample as well as for a sample of people who are particularly likely to be reflective and informed (philosophers and philosophy students). Our paper hence contributes to recent empirical challenges to the thesis that people believe in moral objectivity.
This could be evidence against Watkins’s claims, but unfortunately those claims are so ambiguous that it’s hard to say. Watkins is a moral realist, though, so it’s entirely possible that these sorts of results would be some evidence against what Watkins has in mind by moral progress. Some indication comes from a further remark from Watkins:
Most people judge discrimination as bad, charity as good, lying as wrong, and obeying the law as right. These are not judgments about subjects attitudes [sic] nor are they psychological claims about how subjects in fact act. The objects of judgment are the acts themselves.
This is a bit hard to interpret, but it looks as though Watkins may be hinting at the judgments in question while presupposing some form of realism, since we’re told “These are not judgments about subjects [sic] attitudes.” If so, then I’ve provided at least one empirical study which would suggest that this is false. Furthermore, if Watkins believes that most people’s beliefs in moral progress are a belief in specifically objective moral progress, then the overall state of the empirical literature on folk metaethics does not support his claim in a broader respect, since the best available empirical studies on this question do not indicate that most people are moral realists.
In particular, Pölzler & Wright (2020) found that ~75% of the participants in their studies consistently favored moral antirealism across a range of paradigms, after implementing a variety of corrective measures that mitigated methodological problems with earlier research on how nonphilosophers think about metaethics.
The problem with Watkins’s claims are much worse than engaging in speculative armchair psychology, however. Note that although Pölzler and colleagues have provided evidence that most people aren’t moral realists and likewise don’t believe that objective moral progress is possible, these studies were conducted on a demographically narrow range of populations: college students in the United States, participants on Amazon’s Mechanical Turk, and surveys sent out to participants in the Netherlands, Germany, Austra, and the United States. Most of the participants in these studies are from Western, educated, industrialized, rich, and democratic (“WEIRD”) cultures (Henrich, Heine, & Norenzayan, 2010).
Considerable empirical evidence suggests that WEIRD cultures are psychologically different from non-WEIRD cultures along various lines, including those related to moral judgments (see e.g., Atari et al., 2022; Barrett et al., 2016; Fessler et al., 2015; Graham et al., 2016). This is alongside evidence that there are some ways in which moral judgments are consistent across various demographic characteristics (see e.g., Banerjee, Huebner, & Hauser, 2010). As such, we are simply not in a position to generalize from studies conducted among WEIRD populations to the moral psychology of non-WEIRD populations. Indeed, there are such severe problems with generalizability that it would be extremely difficult to appeal to any set of studies to make confident claims about how everyone, everywhere thinks about any issue (see e.g., Yarkoni, 2020). In short, making confident inferences about what “most people think,” even when we have carefully and rigorously gathered data from thousands of people from our general cultural milieu, is difficult. It is difficult despite the use of carefully constructed studies that have been specifically designed to mitigate the interpretative difficulties associated with conducting research on metaethics among nonspecialist populations. Yet Watkins is making sweeping empirical claims about how everyone thinks with absolutely no data whatsoever. Watkins says things like:
The empirical evidence is our common discourse
Whose common discourse? English speakers? Americans? All humans? How does Watkins know how non-English speakers speak or think when it comes to morality? Does Watkins know how people spoke and thought about morality in ancient Mesopotamia? How about in 10,000 BCE? Sure, Watkins may know a few languages, or be familiar with a few cultures, but we aren’t justified in generalizing from how members of one or a handful of cultures think to how everyone thinks.
This is one of the more troubling aspects of analytic philosophy: its proponents seem to either not know or not care that the way they think is not necessarily how everyone else thinks, nor is it necessarily how they would think under different historical conditions.
While analytic philosophers are welcome to claim, or to think, that their methods yield insights and conclusions that transcend culture, history, and time, and to imagine that if we met aliens, they’d converge on the same philosophical puzzles, and reach similar philosophical conclusions, because we might imagine that e.g., all rational agents, employing a priori reasoning, would tend to think the same way. Only there’s that same problem I keep pointing out: that’s an empirical hypothesis.
Perhaps we could gather inductive evidence that independent thinkers would think the same way as analytic philosophers, but at present, we don’t have much evidence to support such assumptions. To the extent that the field of analytic philosophy, and more generally a priori reasoning relies on the presumption of convergence among other rational agents, the field is making an audacious assumption that culture, history, and education haven’t caused them to think in highly idiosyncratic ways.
If Watkins’s claim is about everyone’s discourse, then it is a very strong claim. How does Watkins know what everyone’s discourse is like? And why even think we have a “common discourse” when it comes to morality? What is this based on? Some languages appear not to have straightforward cognate terms for Western conceptions of “morality,” at all, see e.g. Berniūnas (2020). We aren’t in a position to know whether all societies even have “moral discourse” as such, or if they do, whether it is anything like our own. Non-western cultures, even those which have been historically impacted by contact with Western culture, conceive of morality differently than Western societies (see Berniūnas, Dranseika, & Sousa, 2016), and even different subcultures within WEIRD societies conceptualize the moral domain differently from one another and from mainstream analytic philosophers (see e.g., Levine et al., 2021).
In addition, there is little evidence of any principled way to distinguish moral from nonmoral norms (Stich, 2018). There is little evidence that a distinctive capacity for moral cognition evolved (Machery & Mallon, 2010). There is little evidence of uncontroversial account of the distinctive characteristics of putative “moral judgments,” (Sinnott-Armstrong & Wheatley, 2012; 2014). And there is at least some reason to suspect the very concept of morality, and not just the terms we use to refer to it, is an historical invention that arose in some cultures, but not others (Machery, 2018). This is not to say these questions are unanswerable. Perhaps there is a shared conception of morality, and perhaps moral cognition is a product of natural selection. But neither of these claims has been decisively, or even moderately well-established by available empirical evidence.
Yet analytic philosophers routinely make sweeping claims about how everyone, everywhere, throughout history thinks when it’s not that the evidence fails to support such claims, but that the kinds of studies that could sustain such claims don’t exist at all. Recall that Huemer made the following claim:
It appears that the overwhelming majority of people throughout history were moral realists
It appears this way? Based on what? Does Huemer present robust findings from archaeology, anthropology, history, and comparative linguistics that provides sufficient evidence to warrant such a claim? No. Huemer didn’t present any evidence at all. It’s one thing to claim that it appears to you that there are objective moral facts. But nobody has the phenomenology of the internal mental lives of people living in the Hittite empire in 1500 BCE. We don’t even know if ancient societies had moral concepts at all, much less that they were moral realists. The notion that the overwhelming majority of people throughout history were moral realists is preposterous. Joyce (2006), for instance claims that:
“[m]orality (by which I mean the tendency to make moral judgments) exists in all human societies we have ever heard of.” (p. 134, as quoted in Machery & Mallon, 2010, p. 30)
As Machery and Mallon (2010) point out, in response to Joyce:
To evaluate this argument properly, it is important to keep in mind that the claim that morality is present in every culture is not just the assertion that one finds norms in every culture. Rather, it asserts that in every culture, one finds norms that possess the properties that distinguish moral norms from other kinds of norms according to the characterization used by a given evolutionary researcher. Thus, when Joyce asserts that morality is present in every culture, he is claiming that in every known culture, one finds norms that have most of the seven properties he uses to single out moral norms from other kinds of norms [...] (p. 30)
Available empirical evidence simply does not support such claims. We don’t know what proportion of people had moral concepts at all, much less how many were moral realists. And Huemer himself is on record suggesting that divine command theory can be construed as a form of antirealism; even if, for instance, most people were theists, historically, it does not necessarily follow that they were therefore moral realists.
When asked for evidence, I’ve sometimes had people point out that this or that society had codes of conduct, or rules, and so on. But this only establishes that those societies had norms. “Norms” aren’t the same thing as “moral norms,” unless one thinks “moral” isn’t doing any extra work (which, if one wants to go that route, is going to open a whole new avenue of problems). Machery and Mallon (2010) likewise criticize appeals to the existence of codes of conduct as evidence of the universality of moral thought:
But why should we believe that moral norms are present in all known cultures? Researchers often bluntly assert this claim (e.g. Dwyer, 2006: 237) or illustrate it with ancient or exotic codes of norms. Thus Joyce refers to the norms in the Egyptian Book of the Dead and in the Mesopotamian epic of Gilgamesh (2006: 134—135). We find this casual use of the anthropological and historical literature problematic. The problem is not so much that these researchers have not examined a large number of cultures and historical periods to substantiate the claim that moral norms are ancient and pancultural. Rather, the problem is that because they fail to clearly distinguish norms from moral norms, the evidence they allude to merely supports the claim that norms are universal, but not the much more controversial claim that moral norms are universal (see Stich, 2008 for a similar point). (pp. 30-31)
It’s really difficult for psychologists to figure out how college students in Midwest universities think. Yet philosophers operate under the conceit that, merely by thinking about (without systematically analyzing recurrent patterns via rigorous data collection methods) what they think people mean when those people speak English, that they know how everyone, everywhere, throughout all of time and space thinks. This is no kind way to put it: this requires either an embarrassing ignorance of cross-cultural psychology, a breathtaking level of conceit, or both.
Think about the standard methods philosophers employ to examine what “ordinary people” think when they e.g., make moral claims.
Do they interview ordinary people?
No.
Do they conduct carefully-designed studies to systematically elicit those people’s judgments under controlled conditions to determine what they mean when they engage in moral judgment?
No.
Do they employ corpus data to examine large swaths of actual human speech?
No.
Do they conduct observational research?
No.
They don’t systematically gather any information about actual human behavior or speech. Instead, they sit in a chair, dream up awkward, stilted, decontextualized sentences that are a kind of crude facsimile of how normal people talk, and then they analyze those sentences. In other words, philosophers work out what nonphilosophers think (a psychological claim) not by examining what those people actually say or do, but by making up sentences that are sort of like what those people might say, and then thinking about what they think those people would think, were they to say those things. In short: they attempt to examine the meaning of actual sentences without engaging with any actual sentences, but by exclusively examining the meaning of toy sentences: hypothetical facsimiles of the sorts of sentences someone might say.
For comparison, imagine if a biologist interested in animal physiology wanted to examine tiger anatomy. Rather than collecting specimens of actual tigers: observing them walk and stalk and hunt and sleep, and collecting specimens of tigers: corpses, bones, etc., then reconstructing their anatomy on the basis of a direct examination of their actual anatomy, the biologists instead drew, from memory what they think a tiger looks like, and then created an artificial tiger skeleton from memory based on the bones they saw in a museum a couple months ago, and then examined these. Then, not only did they conclude that they now know what tiger physiology is like: they claim to know what all cats everywhere are like, including non-tigers. This would be laughed at as a ridiculous waste of time by anyone who actually wanted to understand tiger physiology. So why are we taking philosophers seriously when they make claims about how everyone thinks when they simply aren’t using the methods necessary to answer such questions?
There’s a simple answer to this: we shouldn’t.
If you want to know what most people think, you’re going to need to do empirical research. Philosophers who don’t employ or study empirical methods are simply not using the kinds of tools that could answer questions about what “most people” think.
Philosophers should either (a) stop making these claims or (b) employ or at least appeal to other people’s work that employs the proper methods to assess these questions.
With respect to whether most people are moral realists: there is no good empirical evidence this is true. There isn’t even good evidence of a distinctive evolved capacity for moral cognition. At present, we are many steps removed from anyone having sufficient empirical resources to make sweeping claims about some sort of shared, universal moral discourse. Positing the existence of one seems plausibly attributable to a post-Christian cultural mythos more than any kind of empirical evidence (which, for what it’s worth, increasingly suggests that it’s not the case that all or most people are moral realists).
It is important to emphasize that the studies conducted by Pölzler and colleagues may have methodological flaws or may be mistaken, but they are still instances of evidence in practice. Watkins has not presented any such evidence.
Empirical research on how people think about metaethics is not only difficult, but is very difficult. I have spent nearly a decade conducting such research, and have painstakingly documented the empirical difficulties associated with it methodologically (which you can read about in Bush & Moss, 2020 and Bush, 2023). Even when we do our best to phrase questions as carefully as possible and survey thousands of people, it’s still difficult to determine whether they endorse moral realism, or believe in moral progress. The notion that Watkins could know what most people think from the armchair is ridiculous.
References
Atari, M., Haidt, J., Graham, J., Koleva, S., Stevens, S. T., & Dehghani, M. (2022). Morality beyond the WEIRD: How the nomological network of morality varies across cultures.
Banerjee, K., Huebner, B., & Hauser, M. (2010). Intuitive moral judgments are robust across variation in gender, education, politics and religion: A large-scale web-based study. Journal of Cognition and Culture, 10(3-4), 253-281.
Barrett, H. C., Bolyanatz, A., Crittenden, A. N., Fessler, D. M., Fitzpatrick, S., Gurven, M., ... & Laurence, S. (2016). Small-scale societies exhibit fundamental variation in the role of intentions in moral judgment. Proceedings of the National Academy of Sciences, 113(17), 4688-4693.
Berniūnas, R. (2020). Mongolian yos surtakhuun and WEIRD “morality”. Journal of Cultural Cognitive Science, 4(1), 59-71.
Berniūnas, R., Dranseika, V., & Sousa, P. (2016). Are there different moral domains? Evidence from Mongolia. Asian Journal of Social Psychology, 19(3), 275-282.
Bush, L. S., & Moss, D. (2020). Misunderstanding metaethics: Difficulties measuring folk objectivism and relativism.
Bush, L.S. (in press). Schrödinger’s Categories: The indeterminacy of folk metaethics. [Doctoral dissertation, Cornell University].
Fessler, D. M., Barrett, H. C., Kanovsky, M., Stich, S., Holbrook, C., Henrich, J., ... & Laurence, S. (2015). Moral parochialism and contextual contingency across seven societies. Proceedings of the Royal Society B: Biological Sciences, 282(1813), 20150907.
Graham, J., Meindl, P., Beall, E., Johnson, K. M., & Zhang, L. (2016). Cultural differences in moral judgment and behavior, across and within societies. Current Opinion in Psychology, 8, 125-130.
Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world?. Behavioral and brain sciences, 33(2-3), 61-83.
Joyce, R. (2006). The evolution of morality. Cambridge, MA: MIT Press.
Levine, S., Rottman, J., Davis, T., O'Neill, E., Stich, S., & Machery, E. (2021). Religious affiliation and conceptions of the moral domain. Social Cognition, 39(1), 139-165.
Machery, E., & Mallon, R. (2010). Evolution of morality. In J. M. Doris & The Moral Psychology Research Group (Eds.). The moral psychology handbook (pp. 3-47). New York, NY: Oxford University Press.
Machery, E. (2018). Morality: A historical invention. In K. J. Gray & J. Graham (Eds.), Atlas of moral psychology (pp. 259-265). New York, NY: The Guilford Press.
Pölzler, T., & Wright, J. C. (2020). Anti-realist pluralism: A new approach to folk metaethics. Review of Philosophy and Psychology, 11, 53-82.
Pölzler, T., Zijlstra, L., & Dijkstra, J. (2022). Moral progress, knowledge and error: Do people believe in moral objectivity?. Philosophical Psychology, 1-37.
Sinnott-Armstrong, W., & Wheatley, T. (2012). The disunity of morality and why it matters to philosophy. The Monist, 95(3), 355-377.
Sinnott-Armstrong, W., & Wheatley, T. (2014). Are moral judgments unified?. Philosophical Psychology, 27(4), 451-474.
Stich, S. (2018). The moral domain. In K. Gray & J. Graham (Eds.), Atlas of moral psychology (pp. 547- 555). New York, NY: Guilford Press.
Yarkoni, T. (2022). The generalizability crisis. Behavioral and Brain Sciences, 45, e1.