This week I close out my now ludicrously long response to Emerson Green’s initial thread by addressing an exchange between alobar and Dominik. That exchange begins with this remark from alobar:
I think part of the work he’s doing in moral psychology is looking to see whether or not laypeople do in fact have the concept of stance-independent reasons or external moral facts.. perhaps he will find most do not, and will be vindicated
My research does not directly address those questions. I primarily conduct meta-studies that evaluate the validity of measures used to assess whether people endorse moral realism or antirealism, but these paradigms are not constructed in a way where one could readily infer whether participants have the concept of stance-independent reasons or external moral facts. That is, the studies aren’t designed to directly assess whether they have the concepts, but to assess which views they hold using fairly shallow measures that tell us little about how the participants themselves conceptualize their response to the paradigms in question. For instance, the most common paradigm is the disagreement paradigm. Here’s how that might look:
Alex and Sam disagree about lying for personal gain. Alex thinks it is morally wrong to lie for personal gain. Sam thinks it is not wrong. Can they both be correct, or does at least one of them have to be mistaken?
They can both be correct
At least one must be mistaken
Choosing (b) is interpreted as endorsing moral realism. However, this doesn’t provide any direct evidence about what concepts they have, since we have no direct evidence, merely in virtue of their response to the question, of why they chose (b). At present, there is no empirical evidence that directly addresses whether people have the concepts in question, though the empirical account I defend would preclude them from having those concepts. Note, however, that such a case is indirect, abductive, and not based on any direct attempt to assess whether people have the concepts in question.
Dominik responds to this with the following remark:
I am pretty familiar with the literature on moral psychology and afaik most of the research confirms the idea that realism is the common sense position. but I don't even think that matters all that much: the fact that SMART people have that intuition is what counts
Dominik later clarifies, regarding the latter remark:
Yeah I just mean't intelligent people, and particularly I meant philosophers (as you might know the vast majority of philosophers are realists)
Most empirical literature exploring whether nonphilosophers are moral realists or antirealists does not confirm that moral realism is “the common sense position.” If Dominik is aware of some body of literature I missed, or is referring to the same literature I reviewed but has a different interpretation of the results, I’d like to hear about it. However, I specialize in this topic and I never came across any such body of research over the course of many years. I am confident such studies simply don’t exist. It wouldn’t surprise me that much if there is an isolated study here or there that purports to show some sample reported high levels of realism, but it’s very unlikely there is any significant body of literature with well-validated measures and high generalizability, which is what we’d need if we wanted to say realism was the commonsense position, and not (at most) a position common among some familiar populations (e.g., WEIRD populations). Given how little cross-cultural research has been done, it’s absurdly unlikely that a body of literature like this exists and I somehow never encountered it.
In case you’re wondering why I’m so confident the literature does not support this claim, the reason for that is that I just completed my dissertation on the psychology of metaethics, where my emphasis was specifically on the methods of studies used to assess whether nonphilosophers are moral realists and antirealists.
I don’t think the fact that I've spent so much time studying this topic automatically makes me correct, nor would I suggest anyone presume I’m correct merely because I specialize in this topic. Specialists can be, and often are wrong about what they specialize in. Nobody should take my declaration about the state of the literature at face value. If you want to know why I don’t think empirical evidence supports the conclusion that most people are realists, you could read my dissertation, or some of my shorter work on the topic with my colleague David Moss (see Bush & Moss, 2020) or I’d be happy to discuss it.
When you specialize in a topic this narrow, there’s rarely anyone who is sufficiently familiar with the literature to discuss it. If anyone has a different interpretation on the literature, it would be a profound understatement to say that I'd be interested in hearing their perspective. If Dominik thinks the studies I discuss in my dissertation support the claim that moral realism is “the common sense position,” I’d like to know why. If there’s some other empirical literature I’m not aware of that suggests this, I’d like to know about that, too.
The remark that smart people’s intuitions are what count is a strange one. Surveys of philosophers don’t provide us with a representative sample of what smart people think. They provide us with a sample of what people who chose to respond to the PhilPapers survey think, most of whom are analytic philosophers. While analytic philosophers tend to be smart, that does not mean that how they think about moral realism generalizes to how all smart people think (or would) think about moral realism. To my knowledge, there is no empirical research on what smart people in particular think about moral realism. One might point to the results of the PhilPapers survey as evidence of what smart people think, but this is a highly questionable inference.
There are a bunch of factors the respondents to the PhilPapers survey share in common that make them highly unrepresentative of other smart people along numerous dimensions. Most are from the anglophone world and most are analytic philosophers. Available data suggests that a majority of analytic philosophers endorse moral realism, but it’s unclear whether we’d find the same results among smart people in general. A host of factors could account for the 62% rate of realism among analytic philosophers who responded to the survey, including self-selection effects, distinctive features of WEIRD populations engaging in philosophy, and the lack of independence in the methods used to arrive at their conclusions. We can’t even confirm that studying analytic philosophy causes people to be realists.
If most or all nonphilosophers were moral realists, and only 62% of analytic philosophers were moral realists, that would hardly be evidence that specializing in analytic philosophy makes you a moral realist; it might even suggest the opposite, since it could be that analytic philosophy causes many people to reject realism in favor of antirealism. It could be, for instance, that rates of moral realism begin at a very high baseline, due to a strong bias in favor of realism, but that analytic philosophy partially corrects for this bias.
Regarding philosophers in particular, the 2020 PhilPapers survey found that 62% of philosophers endorsed moral realism. That is not a “vast majority.” That’s a solid but slim majority.
The study is also not a survey of “philosophers” but of respondents to that specific survey. Demographically, about half of the respondents were from the United States, and an overwhelming majority operated within the analytic tradition. It is far closer to a survey of analytic philosophers working in the Anglophone world, in particular. As I address at length, what most analytic philosophers think is not strong evidence. See my series of posts on what I call the PhilPapers Fallacy.
Next, Alobar says:
I see, yeah if he’s just cherry-picking that’s not cool.
I’m not. I address studies on the psychology of metaethics to a comically extensive degree in my dissertation. See chapters 2 and 3, along with supplements 2, and 3, in my dissertation. I spend over 500 pages addressing other studies, including studies that purport to find that people lean towards realism (e.g., Goodwin & Darley, 2008, Zijlstra, 2023, etc.). If I don’t address a study that supposedly shows most people are realists there, let me know, and I’d be happy to discuss it. Prior to this, I also published an article with my collaborator, David Moss, that includes a reanalysis of Goodwin and Darley’s (2008) findings.
However, a few things to note: I don’t think most people are antirealists. The reason I cite Pölzler and Wright (2020) study so often isn’t because it happens to support my views, for one very simple reason: it doesn’t support my views. My dissertation is a sustained defense of metaethical indeterminacy, not the claim that most people are moral antirealists. I don’t think most people are moral antirealists. And I dedicate an enormous portion of my dissertation specifically to criticizing this study, which I think has very serious methodological problems of its own. As such, any claim that I’m appealing to the paper because it supports my views would be ludicrous: it doesn’t! I reject the findings in the article. It just happens to be the best article on the topic. What I have endeavored to emphasize is not that this study is correct, but that the preponderance of the evidence, taken as a whole, simply does not support the claim that most people are moral realists or that moral realism is “the” commonsense view.
The reason I regularly cite this is because it is the most methodologically robust of any of the studies that have been done so far. Pölzler and Wright have been at the forefront of identifying and correcting for methodological issues in earlier research. For instance, Pölzler (2018) wrote a paper specifically cataloging methodological issues in earlier research, and the Pölzler and Wright (2020) paper appeals directly to these concerns and does its best to minimize their influence. Have a look at the papers for yourself. Both include an extensive catalog of methodological problems with earlier research, and the latter makes extensive efforts to correct for those problems, including:
A broader range of response options that reflect a wider array of metaethical positions respondents could take
Comprehension checks designed to assess understanding of relevant concepts
Familiarization with some relevant metaethical concepts (such as truth aptness)
More extensive and carefully constructed instructions to minimize bias or conflations
More extensive and carefully constructed response options
A wide variety of distinct paradigms which were conducted within-subjects and thus provide much stronger corroborating data about any particular participant’s metaethical views
…that probably doesn’t even exhaust the methodological advantages of their studies over other studies. But I encourage you to read their studies and compare them to earlier ones.
At any rate, I critique their methods and results as well, and do not interpret their findings as compelling evidence that most people are moral antirealists. For instance, I have frequently pointed out that their studies were conducted on college students and MTurk workers. College students are more likely to give antirealist responses (see Beebe & Sackris, 2016) and MTurk workers are more likely to be atheists and agnostic (Burnham, Le, & Piedmont, 2018; Levay, Freese, & Druckman, 2016), who are in turn less likely to be moral realists (Collier-Spruel et al., 2019; Goodwin & Darley, 2008). As such, I do not think their findings are of a representative sample of the US or the world population.
However, my most serious criticism of their approach points to the extensive training and additional work put into ensuring participants understood the stimuli. While I think this is a laudable effort to correct for the shortcomings of previous research, I think it trades one problem for another.
The main problem is with extensive training participants can change psychologically in such a way that they no longer represent the population they are intended to represent, i.e., ordinary people without much philosophical training. This can occur whenever participation in the study itself causes participants to form views they didn’t previously hold, a phenomenon I call spontaneous theorizing.
Note that the purpose of studies on folk metaethics is to determine whether ordinary people are moral realists or antirealists. “Ordinary people” is key here, as ordinary people are those without sufficient training in analytic philosophy. Yet Pölzler and Wright’s (2020) approach relies on exposing participants to instructions, training, and comprehensive response options. This exposes their findings to a dilemma. Either these efforts are successful, in which case it is no longer clear that the participants are “nonphilosophers” but are more of an intermediate, or novice philosopher, or they aren’t, in which case participants still, as I argue in Bush and Moss (2020) and in my dissertation, aren’t interpreting stimuli as intended and thus the measures are invalid. As such, these studies don’t generalize well to the populations they are intended to represent or the measures remain invalid. The former wouldn’t permit inferences about the broader population since participants no longer represent that population, while the latter would prohibit any confident inferences about what (if any) metaethical position ordinary people endorse simply because participants aren’t interpreting the stimuli as researchers intend. Unfortunately, if your goal is to understand how nonphilosophers think about complex philosophical topics, you may be damned if you do, and damned if you don’t: adequate assurance that they interpret stimuli as intended would change people such that they no longer are nonphilosophers, while inadequate training (or no training at all) can’t circumvent reasonable suspicions (and, in this case, empirical evidence) that the measures aren’t valid. For a more comprehensive discussion of these points, see sections S3.10 and S3.11 in my dissertation (note that my critique of their studies is about 100 pages. I am not slacking in criticizing even studies I like).
This leads me to my own interpretation of the data: metaethical indeterminacy, a proposal initially put forward by Gill (2009).
Roughly, my view is that ordinary people have no determinate metaethical positions on realism or antirealism, and ordinary moral language no more commits people to, or requires thinking in terms of, any particular metaethical view, than ordinary causal claims require commitments to how to interpret quantum mechanics. Yet studies typically present participants with forced choices between a range of categories and distinctions imposed on them by the survey design. If you ask people whether they endorse this or that theory, they’ll tend to pick one, even if that theory does not reflect how that person thought prior to participating in the study. Just the same, it is difficult to determinate in any given instance whether participants held a realist or antirealist position prior to exposure to substantial training, or whether participating in the study caused them to either (a) express a view they don’t hold, as a result of the forced choice paradigms they’re presented with (that is, the study requires them to choose from a list of options, in this case, all realist or antirealist positions or (b) participating in the study causes people to reflect on and form a position they didn’t previously hold.
This account defuses one concern Dominik or others might raise. One could point to earlier research, such as Goodwin and Darley (2008) or Wright, Grandjean, and McWhite (2013), which found evidence of metaethical pluralism. First, note that Goodwin and Darley (2008) describe their findings in a way that can easily mislead people. In their abstract, they say:
Experiment 1 showed that individuals tend to regard ethical statements as clearly more objective than social conventions and tastes, and almost as objective as scientific facts. Yet, there was considerable variation in objectivism, both across different ethical statements, and across individuals.
These findings could be (and have been) cited as evidence that people are moral realists. There are a handful of serious problems with these results, however. First, Experiment 1 created a composite score that combined two paradigms. However, the first of these paradigms is categorically invalid as a measure of realism/antirealism. Participants were asked, for each moral statement, whether it was a:
True statement.
False statement.
Opinion or attitude.
The most serious problem with these response options is that they could, at best, only allow us to distinguish cognitivism and noncognitivism. Antirealists do not necessarily deny that moral statements can be true or false. Moral relativists think that moral statements can be true or false, and error theorists think that moral statements are false. Only noncognitivists would say that moral statements are neither true nor false. There are other problems with these response options, as well: (a) opinions and attitudes aren’t mutually exclusive with true and false statements; moral claims can convey both, (b) “opinion” can be used to express a proposition. Indeed, their own instructions use “opinion” in a seemingly propositional way in the very act of informing participants about what they’ll be doing:
The first part of the experiment asks you to rate your agreement with 26 statements (on a scale from 1 to 6), and to indicate your opinion about the status of each statement – whether it is true, false, or an opinion. (Goodwin & Darley, 2008, p. 1343, emphasis mine)
Response option (3) is also somewhat (due to the use of ‘or’ instead of ‘and’) double-barrelled. You may want to express that it’s an opinion but not an attitude, or vice versa, but there is no way of unambiguously expressing this. More generally, it’s simply not clear how ordinary people interpret terms like “opinion” or “attitude,” or whether they interpret them the same way as one another or as researchers intend. For that matter, technically, there isn’t even good evidence ordinary people endorse the correspondence theory of truth (Barnard & Ulatowski, 2013; Reuter & Brun, 2022). If they don’t, it’s not clear whether they could be moral realists or antirealists in the conventional sense (which further reinforces my proposal that ordinary people simply don’t have determinate philosophical views about metaethics or other issues more generally).
Even their second measure, which was an early form of the disagreement paradigm, has numerous problems, such as the use of epistemic language. The main realist response option is:
The other person is surely mistaken.
The use of ‘surely’ needlessly conflates the response option with epistemic considerations. And, for what it’s worth, my own findings suggest ordinary people routinely conflate epistemic and metaphysical considerations; when the response options themselves don’t even disentangle the two, it’s hard to know whether someone who selects such a response is expressing a metaethical stance, an epistemic stance, or something else. I’m not the only one to raise concerns about Goodwin and Darley’s results. As Pölzler (2017) observes, as I do, As the measures Goodwin and Darley use would be more appropriate for distinguishing cognitivists from noncognitivists:
The results of Goodwin and Darley’s experiments have often been claimed to suggest a tendency towards realism. Reinterpreted according to my above suggestions, however, this clearly is not the case. As many as 62% of the responses of Experiment 1 and 47% of the responses of Experiment 2 — in total more than half of the responses — belonged to the non-cognitivist options R3 (Experiment 1) and R2 (Experiment 2) (see 2008, pp. 1347, 1351; see Fig. 2). And the prevalence of intuitions in favor of anti-realism in general was likely even considerably higher than that. After all, subjects’ cognitivist responses can reflect anti-realist (in particular, error theoretic or subjectivist) commitments rather than realist ones as well.
These remarks are accompanied by a graph:
As this reanalysis suggests, it’s not even clear Goodwin and Darley’s own data supports the notion that people are more inclined towards realism. Note that this is the same Pölzler who designed new studies, a decade after these results were published, that were explicitly intended to circumvent the limitations of these findings. It’s not cherry picking to focus on newer, better studies. There’s a reason I cite Pölzler and Wright’s findings rather than Goodwin and Darley’s results.
Pölzler (2017) goes on to raise numerous objections to the conventional interpretations provided for a handful of other studies which found that most participants endorsed moral realism, though it’s worth noting that two of these studies were conducted among children, I have raised significant methodological objections to them myself. I critique the disagreement paradigm in Chapter 2 of my dissertation, which addresses Wainryb et al. (2004) and Nichols (2004), and I critique Nichols & Folds-Bennett (2003) in Supplement 3, section S3.4.
Dominik’s suggestion that I “ignore” the “many studies which back intuitive realism” is simply not true. I’ve made a point of explicitly and thoroughly addressing studies that purport to do so. This includes two chapters of the dissertation, which center on critiques of all major metaethics paradigms, along with Appendix E and Supplement 2 and 3, which total to more than five hundred pages of work exclusively dedicated to addressing these studies, as well as one published article (Bush & Moss, 2020) and more to come.
Lastly, Dominik says:
And Bush doesn't just say that they are simply wrong... but rather that their ideas are probably fundamentally incoherent. The idea that the majority of metaethicists are just conceptually confused is preposterous
Note that I only emphasize that non-naturalist conceptions of moral realism are unintelligible. And the PhilPapers 2020 survey results don’t show that most philosophers are non-naturalist realists. The 62% of moral realists includes anyone who considers themselves a realist, including naturalists. The survey also included another question about metaethics, which found the following:
26.6% non-naturalism
31.6% naturalism
20.8% constructivism
10.6% expressivism
5.3% error theory
17.9% other
26.6% is not a majority of philosophers, it isn’t even a majority of moral realists! So no, I don’t think the majority of metaethicists are necessarily conceptually confused, or at least I’m not suggesting most are conceptually confused when I claim that non-naturalist realism is unintelligible, since most philosophers don’t endorse non-naturalist moral realism. Only about a quarter do.
Incidentally, that’s about the same amount as antirealism, at 26.1%. There are only about as many non-naturalist realists as there are antirealists. I don’t know what naturalists and constructivists think of non-naturalism, but it wouldn’t be that surprising if at least some also think a belief in non-naturalist realism is due to conceptual mistakes. And note that neither naturalists nor constructivists necessarily endorse any of the spooky forms of normativity non-naturalists invoke. It’s entirely possible most people think the sorts of normative facts non-naturalist realists believe in at best simply don’t exist. Indeed, some naturalist accounts appear to do away with this kind of normativity entirely. In that respect, naturalist accounts often have more in common with my views than with non-naturalists.
Since the main target of my critique is non-naturalism, it’s somewhat misleading for people to appeal to the 62% of realists as evidence that “realism” is correct: naturalist and non-naturalist accounts are very different, and I, at least, don’t reject naturalist accounts as conceptually confused or unintelligible for the reasons I reject non-naturalist realism.
On the other hand, if most metaethicists think the concepts in question are intelligible, even if they don’t endorse them, then it’s true that I think most philosophers are conceptually confused. However, Dominik offers no argument or explanation for why my position is “preposterous.” It’s simply asserted. I’ve put a fair bit of effort into explaining why I disagree with Dominik’s remarks. I hope Dominik will consider returning the favor, getting in touch, and offering an explanation as to why my views are preposterous.
If so, I hope there is more to it than pointing out that lots of smart people disagree with me. Truth isn’t a numbers game, and philosophical disputes aren’t determined by who has the most smart people on their side. Imagine if an evil positivist obtained the Infinity gauntlet and snapped everyone who disagreed with me out of existence. Now, 100% of people endorse my views. Would that make me any more likely to be correct? No, of course not. What matters is why realists endorse realism, and why I endorse antirealism. Not how many of us there are.
More generally, I’d be up for having a discussion on these topics. I’m not overly thrilled at being on the receiving end of accusations of arrogance, suggestions that I’m “pretending,”, and what seems to me to be a generally derisive attitude towards me. But maybe we could get past that and have a productive discussion. I’m willing to give it a try.
References
Barnard, R., & Ulatowski, J. (2013). Truth, correspondence, and gender. Review of Philosophy and Psychology, 4, 621-638.
Beebe, J. R., & Sackris, D. (2016). Moral objectivism across the lifespan. Philosophical Psychology, 29(6), 912-929.
Burnham, M. J., Le, Y. K., & Piedmont, R. L. (2018). Who is Mturk? Personal characteristics and sample consistency of these online workers. Mental Health, Religion & Culture, 21(9-10), 934-944.
Bush, L. S., & Moss, D. (2020). Misunderstanding metaethics: Difficulties measuring folk objectivism and relativism. Diametros 17(64): 6-21.
Collier‐Spruel, L., Hawkins, A., Jayawickreme, E., Fleeson, W., & Furr, R. M. (2019). Relativism or tolerance? Defining, assessing, connecting, and distinguishing two moral personality features with prominent roles in modern societies. Journal of personality, 87(6), 1170-1188.
Goodwin, G. P., & Darley, J. M. (2008). The psychology of meta-ethics: Exploring objectivism. Cognition, 106(3), 1339-1366.
Gill, M. B. (2009). Indeterminacy and variability in meta-ethics. Philosophical studies, 145, 215-234.
Levay, K. E., Freese, J., & Druckman, J. N. (2016). The demographic and political composition of Mechanical Turk samples. Sage Open, 6(1), 2158244016636433.
Nichols, S. (2004). After objectivity: An empirical study of moral judgment. Philosophical Psychology, 17(1), 3-26.
Nichols, S., & Folds-Bennett, T. (2003). Are children moral objectivists? Children's judgments about moral and response-dependent properties. Cognition, 90(2), B23-B32.
Pölzler, T. (2017). Revisiting folk moral realism. Review of Philosophy and Psychology, 8, 455-476.
Pölzler, T., & Wright, J. C. (2020). Anti-realist pluralism: A new approach to folk metaethics. Review of Philosophy and Psychology, 11, 53-82.
Reuter, K., & Brun, G. (2022). Empirical studies on truth and the project of re‐engineering truth. Pacific philosophical quarterly, 103(3), 493-517.
Wainryb, C., Shaw, L. A., Langley, M., Cottam, K., & Lewis, R. (2004). Children's thinking about diversity of belief in the early school years: Judgments of relativism, tolerance, and disagreeing persons. Child development, 75(3), 687-703.
Wright, J. C., Grandjean, P. T., & McWhite, C. B. (2013). The meta-ethical grounding of our moral beliefs: Evidence for meta-ethical pluralism. Philosophical Psychology, 26(3), 336-361.
Zijlstra, L. (2023). Are people implicitly moral objectivists?. Review of Philosophy and Psychology, 14(1), 229-247.