Antirealism as a form of insanity
Twitter Tuesday #1
This is the inaugural post initiating a new series on this blog: Twitter Tuesdays. Twitter Tuesdays will consist of posts commenting on and in some cases responding to remarks made on Twitter.
For our first post, there was a discussion about my views some time ago following this remark from Emerson Green. The first few weeks of this series will consist of a comprehensive response to Emerson and to some of those who commented in response. Let’s go!
Part 1: Antirealism as a form of insanity
Emerson remarked that:
Six days of the week I think moral antirealism is a form of insanity, and one day out of the week I think we're just metaphysically reifying our social practices to cope
This is an interesting remark. I am curious why Emerson thinks moral antirealism is a form of insanity.
Antirealism is the position that there are no stance-independent moral facts. It’s a view which involves the denial of a distinct metaphysical claim: moral realism, the view which holds that there are stance-independent moral facts. Stance-independent moral facts are facts about what is morally right or wrong, morally good or bad, facts about what morally should or shouldn’t do (or are permissible, obligatory, and so on) that are true independent of the goals, standards, or values of the stances of any person or group.
Moral realism is a view held almost exclusively by philosophers and people who read their work, and hardly anyone else. Some of these philosophers will claim that most people are moral realists. The presumption that moral realism is “the commonsense view” or “the default” is widely held among at least some analytic philosophers. However, this notion is probably false.
The proportion of ordinary people who endorse moral realism or speak or think in a way that commits them to it is an empirical question. Available empirical evidence does not support the claim that most people are realists, or that moral realism is “the” commonsense view. There isn’t even much evidence that there is any singular commonsense view about moral realism, or many other philosophical views. Philosophers often seem to operate under the presumption that there is some shared commonsense view of the world that can be cashed out in terms of inchoate philosophical accounts that roughly accord with the categories and distinctions endorsed by contemporary analytic philosophers.
There is very little evidence to support this presumption; indeed, it appears to me to be based on deep and pervasive misconceptions about how human cognition works. In particular, everyday judgments and decision-making is largely improvisational, spontaneous, and constructed, emerging moment-to-moment in the relevant contexts in order to facilitate the proximal goals of the agents in question (see Chater, 2019). Our judgments don’t need, and probably don’t typically involve, any implicit commitment to principled philosophical positions at all.
In addition, I suspect many of those convinced moral realism is “intuitive,” “obvious,” or “self-evident,” mistakenly believe that what’s intuitive, obvious, or self-evident to them is intuitive, obvious, or self-evident to most other people as well. That, too, is an empirical claim. And it is not supported by available empirical data. Most people are probably not realists or antirealists, any more than most people endorse particular stances on quantum mechanics or mathematical platonism (Gill, 2009).
If you’re convinced something is as obvious (or should be obvious) to most people as it is to you, you might think if someone doesn’t believe it, they’re “insane.” If, for instance, there were people who sincerely insisted the sun did not exist, or that water wasn’t real, perhaps they would be insane. Note why this would seem insane: these people would appear to have beliefs utterly out of accord with publicly accessible information, and, in practice, such disbelief would likely lead to all manner of bizarre situations. If someone didn’t believe in water, how would they avoid dehydration? What would they report seeing if you brought them to a lake or a pool? How would they react to standing in the rain? Such a person would struggle to effectively navigate through the world, and might even struggle simply to survive as a result of their beliefs. Disbelieving in readily available objects or phenomena in the world might very well be a sign that the person may have some form of ,agnosia or a ridiculous philosophical position that we should not take seriously (and, we might expect, wouldn’t actually manifest on their behavior overly much).
Denying moral realism isn’t like this. There are no necessary consequences of moral realism at all. Most (if not all) consequences would be contingent on the psychological reactions of the person who stopped endorsing moral realism or the reactions of other people to their metaethical views, such as possible opprobrium from moral realists. Nothing about belief or disbelief moral realism requires one, for instance, to endorse or abandon any particular values or act or speak in any particular way aside from (if one is being honest and consistent) expressing certain abstract philosophical views (e.g., one might feel obliged to say “I am not a moral realist”).
Moral realism is a highly specific technical position invented by philosophers. There is very little theoretical basis for thinking that anyone outside of those who study philosophy are moral realists. And there’s nothing “insane” about denying the metaphysical thesis put forward by realists, given how practically insignificant the position actually is.¹ Moral antirealists aren’t so confused about morality that we find them torturing puppies or donating money to serial killers. Not endorsing moral realism is no more insane than denying the B-theory of time. Both positions consist of rejecting some highly abstract and theoretical position that has little or no practical implications for almost anything any actual person would ever do. Denying moral realism is not at all like believing the sun doesn’t exist or that water isn’t real.
Moral antirealists are not only in a good position to deny moral realism, they’re also in a good position to deny that moral realism is “the” commonsense view, or is widely held among nonphilosophers. Again, this is an empirical claim, and those who insist most people are realists rarely marshal empirical evidence to support this claim. Why don’t they marshal such evidence? Because there isn’t good evidence for this claim. Or worse, because they don’t even recognize this to be an empirical question in the first place. The latter is an especially ludicrous position: facts about how people think and what people mean when they say things are empirical considerations. You’re not going to get much purchase on them from the armchair.
I specialize in the psychology of metaethics. I study whether or not, and to what extent, nonphilosophers are moral realists or antirealists. There is a growing empirical literature on this question. This literature does not unambiguously support the view that a vast majority of people are moral realists.
The results are, at best, mixed. Some studies find a preponderance of realism, or at least claim to (e.g., Goodwin & Darley, 2008; Zijlstra, 2023). Others find a preponderance in favor of antirealism (Davis, 2021; Polzler & Wright, 2020). Others find very mixed results (e.g., Wright, Grandjean, & McWhite, 2013). Studies show realist responses rates are higher among children (e.g. Nichols & Folds-Bennett, 2003; Wainryb et al., 2004), decline in early adulthood (Beebe & Sackris, 2016), and then increase again over the lifespan.
Realist and antirealist response rates correlate with a variety of other variables (Collier-Spruel et al., 2019; Goodwin & Darley, 2008, 2010; 2012; Feltz & Cokely, 2008), e.g., relativism appears to correlate with tolerance for contrary moral standards and openness to experience, but, importantly, people exhibit sufficient individual differences to be subject to measurement. That is, if you create a scale that measures whether the participants endorse moral realism, you don’t get a perfect run of affirmations: you get a mix of responses, with many participants favoring antirealism. Collier-Spruel et al’.s (2019) scale, for instance, found that respondents fell generally around the middle of a distribution for scale items designed to assess support for moral relativism.
What we don’t see is any kind of overwhelming tendency for people to favor realism. The presumption that pretty much everyone is an intuitive realist is not supported by the data. I have been repeating this for well over a year, and I still see realists insisting everyone’s a realist. This notion is, at best, yet to be confirmed. At worst, it is inconsistent with the best available data, and may simply be a myth.
Does this mean most people are moral antirealists? No. My own stance on this literature is that people’s responses are all over the place because they are not interpreting the questions in line with researcher intent. As a result, my take is that virtually all studies on the metaethical views of nonphilosophers rely on invalid measures.
For comparison, most nonphilosophers would probably be unable to interpret questions about advanced topics in philosophy without substantial training: presenting a forced choice of “yes” or “no” to whether they think there is synthetic a priori knowledge won’t be illuminating because people aren’t going to know what that means, any more than they’re going to be able to disentangle questions about theoretical physics or biochemistry (some participants will be able to do so, but it’s likely a very small proportion, and most won’t be laypeople, just specialists who happen to be filling in a survey).
When researchers design studies, they typically use a forced choice paradigm, e.g., multiple choice, that requires them to select one of a limited range of responses that determinately stake out a stance on the matter. Imagine asking a bunch of random people on the street the following question:
Which theory of time is correct?
A-theory
B-theory
C-theory
That thing my uncle Frank said about how time is an illusion or whatever, lol
And suppose 52% of people chose A-theory. Would that be an indication most people endorse A-theory?
Maybe. What if you asked them why they chose it and they said “I have no idea. I don’t know what the hell you’re asking me and I just chose a response without really thinking about it.” Well, in that case, perhaps not. Just the same, studies evaluating people’s metaethical views typically require them to select response options that necessarily entail that they’ll favor either a realist or antirealist position. This guarantees you’ll have the impression that there’s some proportion of realists and antirealists in your participant pool. But what if people don’t interpret what researchers are asking them as intended? In that case, you have a whole bunch of people whose responses look like they’re favoring realism or antirealism, even if the measure isn’t valid. This, I believe, is what we’re dealing with with respect to these studies.
Why do I believe that? Because I spent years studying written responses to many of the main measures designed to evaluate whether people are realists or antirealists. These written responses were open-ended requests to explain why one answered the question as they did, to explain what they thought the question was asking, to explain what they thought various responses to a given question asked, and so on. Reliably, I found that most participants did not interpret these questions as researchers intended. I had two colleagues code some of the data, and they likewise found extremely low intended interpretation rates (one judged even fewer than me to be intended interpretations, and the other somewhat more, though our judgments were not wildly different). Across numerous paradigms and measures, I almost always found that most participants did not interpret these questions as intended.
Given these findings, there are at least a couple possibilities: we could design better measures and then find out what people really think about metaethics, or we could propose that people have no stance at all on these issues. I favor the latter interpretation. My colleague and I discuss the difficulties with the former a bit in Bush and Moss (2020). The latter is called metaethical indeterminacy (see Gill, 2009). Metaethical indeterminacy can come in a few forms, and apply to a few different sets of considerations, and I don’t want to adjudicate all of that here. So I’ll just state what this form entails: it entails that ordinary moral judgment and discourse is not driven by an implicit commitment to realism or antirealism, that there is no fact of the matter about which view best accounts for how people are speaking or what people think, and that people have no conscious commitment to, or belief in, moral realism or antirealism in particular.
This does not mean no people endorse such views. Some people have considered these topics sufficiently to have a view on the matter, and you do get a subset, across almost every measure, of participants expressing what look like clearly intended interpretations, suggesting that their responses may very well reflect a genuine metaethical position. However, such instances are rare, and it would take further research to determine whether they really do have determinate metaethical stances.
These findings also don’t rule out the possibility that people would be more intuitively disposed towards one or the other position on reflection. However, there are significant problems with any research that attempts to gather such data: data about people’s reflective intuitive reactions is only valid insofar as we can ensure that such responses align with the measures being used to evaluate them, and the only way to ensure that would be, effectively, to train participants in the appropriate terms, concepts, and distinctions…at which point, those people are no longer nonphilosophers, and we’d be unable to distinguish whether their judgments reflect preexisting philosophical tendencies, or were created by the act of thinking in philosophical terms. And, because thinking in philosophical terms is not theory-neutral, it is always possible that the proportion that favor realism or antirealism is biased by the mode of instruction they were subjected to. There is at least some evidence that experimenter bias can emerge in experimental philosophy (see Strickland & Suben, 2012).
I call the tendency for participants to develop a position as a result of engagement with research stimuli spontaneous theorizing, and it would be difficult (and perhaps not possible in practice) to distinguish whether participant responses following training reflected a discovery of their pretheoretical philosophical intuitions, or the invention of post-theoretical intuitions. Personally, I don’t believe this is possible because I don’t believe people have pretheroetical intuitions about metaethics. In short: I don’t think we can measure whether people were realists or antirealists all along because I don’t think people ever were realists or antirealists all along. People exist in a sort of philosophical superposition that is collapsed only by engaging in philosophy. This is why I think that, effectively, only people who study philosophy (whether professional or not) actually have positions on realism or antirealism, just as only people who study quantum mechanics can meaningfully endorse the Copenhagen, Many Worlds, or Bohm interpretations. Why don’t I believe people have pretheoretical stances on metaethics? There are a few reasons, but I’ll focus on the main ones.
One reason is an appeal to abduction: I believe this is the best explanation of empirical literature, including my own research.
A second involves an appeal to broader theoretical considerations. I endorse Chater’s (2019) view that ,the mind is flat: there is very little going on beneath the surface, and most of our explanations of what we say and do are improvised and constructed narratives that are often as fictional as the lives of characters in novels. Moral judgments don’t need to be, and simply aren’t, driven by implicit commitments to philosophical theories. To the extent that such theoretical constructs are imposed on people, they have no claim to psychological reality. One might retreat to claiming that such explanations provide the best account of ordinary moral discourse, with or without a claim to psychological reality (see e.g., Sinnott-Armstrong, 2009). At this point, what are these even theories of? They’re not theories of what people intend to communicate, or take themselves to mean, or what they think, or believe, or feel. Such accounts treats words and sentences as though those words and sentences have some type of intrinsic meaning independent of what people are using those words and sentences to do, what they think about them, and so on. This is a bizarre (but I suspect common) view of language, and I don’t buy it, either. I endorse Mortensen and Chater’s (2022) account of language in The Language Game. I think Chomsky’s views on language are wrong, and that emerging reconceptions of language emerging out of Chomsky’s shadow cast serious doubt on much of the view of language that dominated the 20th century.
Unfortunately, I suspect a dogmatic adherence to strange views about language is embedded itself in much of contemporary analytic philosophy. Language is a behavior, words and sentences have no fixed meaning, and we use language to convey our intent in a way that is contextual, local, improvised, and functions to achieve proximal goals. What we say, and by extension, how we think, is not underwritten by a rigid and fixed adherence to words and phrases having distinct, determinate meanings. Efforts to figure out the precise application conditions for attributions of “knowledge,” were doomed from the very start for a very simple reason: there are no such application conditions. There is no JTB+some secret semantic sauce that would let us know precisely which judgments are or aren’t instances of “knowledge.” Knowledge, like any other word, is used in practice in ways that are highly variable and context dependent. Efforts to prune the pragmatics away from this term, or from our moral claims, and so on, to examine the pure semantics are futile: there is no such thing. It’s pragmatics all the way down. Words don’t have meanings apart from how people intend to use them. Just the same, moral claims like:
Murder is wrong.
…literally have no meaning outside a context of usage. Questions about whether this sentence is propositional or not are literally nonsense. This isn’t even a real moral claim. It’s a toy claim. It’s no more a real moral claim than hot wheels toys are real cars. See Figure 1.
Figure 1. Note. Not an actual car. Probably not a good model for understanding how actual cars work. Unless you’re a philosopher. In which case you won’t even need a toy car. You can study cars by imagining how cars work. Analytic philosophy is awesome. It’s like magic! Only much less cool or useful.
It makes no more sense to study ordinary moral language by analyzing fake sentences than it does to teach mechanics by presenting people with toy cars. If you want to study ordinary language, you’re going to have to actually study ordinary language, which, unfortunately for anyone who enjoys the armchair, occurs in ordinary contexts. Not in your imagination.
“Murder is wrong” would only obtain meaning by virtue of usage. Someone would have to be saying this in some real-world context. Then, and only then, would there be some fact about what is meant, and that fact would be determined by the function of the sentence in that context, which would only tell us what it meant in that context, and would not logically necessitate what “murder is wrong” meant in some other context, because it could be used for some other purpose in another context. Facts about how consistent, stable, and similar other people’s usages of a structurally identical sentence, or how the same person used the same structure in other contexts, cannot be presumed from the armchair, but again turns on empirical facts which one cannot determine by analyzing toy sentences.
Yet every major metaethical position has, at its foundation, a semantic thesis about what “moral claims” mean:
Realism/error theory: moral claims are truth-apt and describe stance-independent moral facts
Noncognitivism: moral claims are not truth-apt, but instead convey exclusively nonpropositional content
Stance-dependent theories: Moral claims are truth-apt and describe stance-dependent moral facts
Every one of these accounts, in its standard form, is not even wrong. There are no such things as “moral claims” of the sort over which these “theories” quantify. “Moral claims” of the sort these accounts deal with are a fiction, a theorist’s construction, a philosophical Pinocchio that exists only in the minds of philosophers.
Actual moral claims are events. They are instances circumscribed by some reasonable operationalization of what does and doesn’t count as a “moral claim,” that picks out a subset of utterances, written phrases, and so on that occur in the real world. Suppose, for instance, one wanted to study fruits. Presumably, one would go out and gather instances of fruits: bananas, apples, oranges, papayas, and so on. One would avoid things that don’t count as fruit: tables, wombats, tires, and so on. One would then examine the characteristics of the fruits one gathered. It would make no sense to imagine fruits and then examine the properties you impute onto these imagined fruits. How much vitamin C do you imagine is in watermelon? The question is ridiculous, and would tell you very little about the nutritional content of watermelons. If you want to know what’s in a watermelon, you need a watermelon.
Indeed, one can ask: what does an imaginary watermelon taste like? One can imagine the task, but that is not what I am asking. I am asking what an imaginary watermelon actually tastes like. This question is absurd, for the obvious reason that imaginary watermelons have no actual taste. The philosopher’s invention, the toy moral sentence, is just like this: it can no more have properties like being propositional or being true than an imaginary watermelon can taste like something. One can, of course, imagine that the watermelon tastes a certain way, but this is a matter of projection, or stipulation. Just the same, the philosopher’s toy moral claims only have those properties the philosopher projects onto them or imagines they have. The philosopher may do their best to project or imagine what they think nonphilosophers would probably mean in some unspecified context, but note that (a) if they’re not doing this, it’s not clear what they’re doing or whether or how the claims in question could have any properties in principle that aren’t projected or stipulated, (b) if they’re doing this, it is very bad psychology: how do philosophers know whether their judgments about what other people mean are accurate unless they go out and check?
Again, it is important to stress that real moral claims are events. If you wanted to study events you’d likewise identify instances that count as events. Suppose you wanted to study fist fights or cooking practices or bird behavior. You’d need to figure out what you’re going to count as a fist fight, an instance of cooking, or a bird. Then you’d study the appropriately circumscribed phenomena. Moral claims are just like this. They are events that occur in the real world. If you want to know what people mean when they make moral claims, you need to study people making those claims. And what does it mean to study what that person means? How would you operationalize that? You have a few options: studying what the person takes themselves to mean, studying the psychological processes associated with those claims, studying how those claims function in their linguistic community, and so on. What you’d be doing is, in an obvious sense, empirical research. It’s cognitive science, linguistics, psychology, anthropology, and so on.
This is not the type of research philosophers studying “moral claims” opted to study. Most philosophers don’t study actual instances of moral utterances, and almost never in any systematic, scientifically respectable way. Instead, they invent sentences that seem kind of like the sorts of things people might say, and then examine the properties of these toy sentences.
I want to provide a few more examples to try to illustrate just how bizarre this is. The process they engage in makes about as much sense as trying to study actual tiger behavior by drawing a picture of a tiger or writing a story about one and then studying that. If pressed, some “tigerologists” insist they’re not even making empirical claims about tigers-in-the-world, they’re studying the nature of tigers in the abstract, independent of the qualities or properties of any actual tiger. That would be a ridiculous notion in the case of tigers. Why is it any less ridiculous in the case of words or sentences? They are just as much a part of the natural world as tigers.
Suppose, instead, the philosopher embraces their practice as empirical. If so, they can do their best to envision tigers as they really are. But you’re bound to get some things wrong, and put yourself at great risk of projecting your own biases, preferences, and attitudes onto the nature of this imaginary tiger. Why think philosophers are any less susceptible to such errors when reflecting on the meanings of terms and phrases? After all, those terms and phrases are often central to the philosopher’s goals, hopes, dreams, aspirations, interests, and career.
And their own experiences are going to reflect a single, highly idiosyncratic, culturally parochial conception of the way words and terms are used, often reflective of only how they are used in one or very few largely culturally overlapping linguistic communities, almost exclusively composed of English speakers speaking specific mainstream dialects of modern English. There are over 6,500 languages in the world, and thousands and thousands of linguistic communities spread across time and around the world. The notion that how you use terms and phrases reflects how everyone, everywhere, through all time and that you’re entitled to presume this is the case in the total absence of anything even approximating decent empirical support is so ludicrous it beggars belief. That is insane.
One way to put this is that philosophers are attempting to use their minds to understand reality, but they have an ineliminable conflict of interest that compromises their ability to think about philosophical topics in a way that doesn't bias their views in the direction of preferred conclusions. Given the ephemerality of their subject matter, compared to the real thing, it is far easier for one’s imaginings to become warped and distorted by one’s biases. For comparison: it is much easier to draw the stripes on a tiger incorrectly than it is to look at the tiger and mistakenly believe its stripes match one’s preconceptions. In short: reflection over abstract topics is far more malleable in light of one’s biases than the world around us. Your motivation to believe lava isn’t hot isn’t going to save you if you fall into a volcano. Part of the reason for this is that when false beliefs have practical consequences, we have a tendency to correct those beliefs or face the consequences. Believing there’s no tiger or volcano or snake means you die. Believing moral realism is true or false has far fewer immediate and concrete consequences.
In the case of toy sentences, they don’t have any properties waiting to be discovered. Rather, insofar as they have properties, those are properties that philosophers themselves project onto them or simply stipulate that these made-up sentences have. Put another way, insofar as “murder is wrong” can have properties, it has whatever properties you imagine. You could say it’s a proposition, but why stop there? Why not say it’s blue or insist that it gyres and gimbles in the wabe? If one’s goal is to understand actual moral utterances, why settle for fictions? And if your goal isn’t to study actual moral utterances by way of these toy sentences, then what relevance does whatever meaning you believe you’ve discovered in your analysis have for anything in the real world? In other words, if you aren’t generalizing from your analysis of the meaning of toy sentences like “murder is wrong,” to actual instances of people asserting similar remarks, then your reflections suffer a terminal case of zero generalizability: one’s analysis may fail to generalize to any actual sentences.
Imagine, for instance, that your analysis of toy sentences led you to believe cognitivism were true. But now suppose we went out and did an empirical study and we found out that 100% of the time, when ordinary people made moral claims, they spoke and thought as noncognitivists. Would this be irrelevant? What is your account even an account of at that point? Because it’s not describing anything people actually think or intend to say. This really is lose-lose for anyone who denies their analysis of the claims in question are empirical. If they’re not, it’s unclear what their subject matter is, nor is it clear whether it has any relevance to what people say, do, and think in the real world (at least not without conducting, yet again, empirical research). If they are, then philosophers are using the wrong methods, and the results of what studies have been done don’t generally support the view that nonphilosophers tend to be realists.
In short: toy sentences aren’t out there in the world to be discovered in the way actual moral utterances are. They are inventions of philosophers. The moment one is dealing in fictions, one’s imagination is free to impute any properties onto the imaginary object. The problem is that philosophers appeal to intuition, introspection, and other inscrutable psychological processes in a way that feels from the inside, given its phenomenology, like an act of discovery. It is not. It is an act of confabulation. Toy sentences could fail to mean anything. The toy sentence “murder is wrong” is simply a string of letters that has no meaning in and of itself, for the simple reason that words and sentences don’t have meaning outside a context of usage. One has to impute the shadow of a context onto them for them to have any conceivable meaning at all. This may involve, for instance, imagining in vague outline a typical context of usage. That is, we imagine the most prototypical, ordinary, everyday, or usual instances of the usage of the terms or utterances in question (at least one’s we are familiar with).
Of course, this will lead us to be subject to the availability heuristic, which will in turn be modified and canalized, for philosophers, down idiosyncratic path-dependencies distinctive to their education and experience. Ever notice how philosophers reliably draw on the same sorts of situations?: Trolley problems, torturing babies, and so on. The kinds of moral issues philosophers focus on and think about are not representative of everyday moral considerations. Philosophical reflection suffers an analogous shortcoming as psychological research, which often suffers poor generalization because we model stimuli as fixed effects when we should instead model them as random effects (Judd, Westfall, & Kenny, 2012).
Roughly, the problem is that philosophers will choose some narrow set of stimuli: two or three questions, one or two examples, and so on, and then extrapolate from how people think about these to how people think about an entire domain of discourse. For instance, imagine presuming that one could understand the entirety of human morality by exclusively examining how they think about trolley problems. That would be absurd. When philosophers think about moral issues, those moral issues that happen to be popular in philosophical papers and discussion will tend to be more salient, even if those issues aren’t representative of everyday moral judgment. How often have you had to decide whether to stop runaway trains from running people over?
Consider a “prototypical” instance of a moral claim that may appear in a philosophical text (despite the fact that few people regularly render judgments about murder). Philosophers will ask us to imagine a person saying:
Murder is wrong.
…in such cases we are probably not imagining that this is a password to enter a seedy underground bar. Nor would we imagine that they’re an instructor having the following exchange:
Instructor: Alex, how did George Washington die?
Student: He was murdered by a British soldier.
Instructor: No, ‘murder’ is wrong. Try again.
This is almost certainly not what people have in mind. Rather, they might instead imagine someone in an exchange like this:
Alex: What should we do with the captives? We can’t bring them back to the base. And we can’t just kill them.
Sam: No, of course not, obviously murder is wrong. We have to find another way.
In other words, we may imagine some case where a person is expressing a stance towards murder. Notice how strained even this example is. People don’t ordinarily need to say that murder is wrong. An interesting feature of ordinary moral language is that people are probably less likely to make moral claims about moral issues everyone agrees about than more controversial ones. As a result, our actual moral beliefs and attitudes are probably not closely correlated with our speech patterns: it may be that controversial moral issues are overrepresented, since people are more likely to argue about them.
In any case, we might suppose that there’s some fact about what this person would mean that reflects the standard, conventional, primary usage of the utterance “murder is wrong”: they could be expressing a claim about stance-independent moral facts (realism), a claim about their personal standards or the standards of their culture (relativism), an attitude of disapproval (noncognitivism), and so on. However, again, this is an imaginary case. There is no fact of the matter about what Sam means. Sam isn’t a real person. At best, consideration of such cases may generate empirical hypotheses about what an actual person would actually mean in actual contexts were they to say “Murder is wrong.”
But if you wanted to answer that question, generating armchair hypotheses isn’t going to provide any decisive resolution to the matter. One can speculate all one likes about the structure of such sentences, and consider, from the armchair, how you think about such sentences, but the moment you extract remarks like these from their everyday contexts and consider what they mean in artificial circumstances, you are no longer dealing with an analysis of an actual moral utterance. You are dealing with an analysis of a toy sentence: an artificial facsimile of the real thing. For instance, you may find that when you reflect on what you would mean in this context, you would take yourself to be asserting a claim about stance-independent moral facts. Even supposing that were true, does it follow that any actual person who says “murder is wrong” is therefore likely to also mean the same thing? Maybe, if we have good reason to think other people mean the same thing as you do. But why should we suppose that’s the case? One might think it’s obvious why we should think this: generally members of a linguistic community use words and phrases to mean the same sorts of things. There are a few problems with this:
First, this may be true: people do use words and phrases to mean the same sorts of things. But it does not follow that they use utterances like “murder is wrong” in ways that are rigid, fixed, and insensitive to context. Instead, it may be that people generally use such phrases in similarly improvised and context-sensitive ways. It could be that Alex, Sam, and the philosopher all use the phrase to mean different things in different contexts, and would even happen to use the phrase in similar ways in similar contexts. The presumption that people are doing more or less the same thing with language is reasonable. But it does not follow that people use phrases in rigid and fixed ways that are insensitive to context, or that there is any context-free semantic core of meaning behind our shared use of language. As such, if you wanted to know whether people used language the same way, similar usage may itself still be variable and context-dependent, and it may still be the case that the only way to determine shared usage is to engage in empirical research.
Note that it would be one thing to presume that if you meant something in a specific context that others would mean the same thing in that context. But this is not what philosophers are doing when they analyze toy moral claims. Instead, they are presuming that what they “mean” by a decontextualized sentence that has no context at all reflects what they and others mean in and across actual contexts of usage. It’s unclear why we should suppose we are justified in generalizing from the meaning of decontextualized toy sentences to the meaning of actual sentences. For comparison, even if you and most other people produced very similar images of tigers when you imagined them, this would not justify conclusions about the actual appearance of tigers.
Second, philosophers are trained to think about such phrases in extremely idiosyncratic ways. It may be that they have adopted a way of thinking that is so at odds with ordinary thought that they’re actually worse judges of what people think, in virtue of a commitment to deep misconceptions about how language and meaning work. Thus, even if philosophers had good insight into how they think, it may or may not generalize to how other people think. It probably doesn’t even for members of their own linguistic community.
Matters are even worse than this, however. As I argued earlier, all languages function the same, and not all people think about these issues in the same way across languages. Most philosophers think, write, and discuss mainstream analytic metaethics exclusively in English. Why should we think their linguistic analysis of the English language in particular can generalize to the other 6,500+ languages in the world, and the linguistic communities distinctive to each of those languages? Even if we could generalize, that itself would require empirical evidence that would warrant such generalizations.
In other words, we’d first need good evidence that the way people are disposed to think, on reflection, about moral utterances in English generalizes to how native speakers of Madang languages, or Pirahã, or or Proto-Indo-European, and so on. Do we have that evidence?
No. Not even close. Indeed, we don’t even have good evidence that all of these languages lexicalize moral language in a way that comports with Western conceptions of morality (see e.g., Berniūnas, 2020; Dranseika, Berniūnas, & Silius, 2018).
In short, the notion that we can extrapolate from a tiny, culturally parochial, and psychologically idiosyncratic (see Henrich, Heine, & Norenzayan, 2010) group of thinkers operating exclusively in one particular language to how everyone, everywhere, throughout all of human history speaks and thinks is not at all supported by our growing understanding of the role of linguistic variation and human cognition.
Indeed, researchers have argued that “,Overreliance on English hinders cognitive science” (Blasi et al., 2022), while sampling bias has led to serious problems of generalizability in applied linguistics, which often involves excessive focus on English (see e.g., Andring & Godfroid, 2020). Consider the highlights from the first of these articles:
The cognitive sciences have been dominated by English-speaking researchers studying other English speakers. We review studies examining language and cognition, contrasting English to other languages, by focusing on differences in modality, form-meaning mappings, vocabulary, morphosyntax, and usage rules. Critically, the language one speaks or signs can have downstream effects on ostensibly nonlinguistic cognitive domains, ranging from memory, to social cognition, perception, decision-making, and more. The over-reliance on English in the cognitive sciences has led to an underestimation of the centrality of language to cognition at large. To live up to its mission of understanding the representational and computational capacities of the human mind, cognitive science needs to broaden the linguistic diversity represented in its participants and researchers.
Mainstream analytic philosophical emphasis on analysis of moral utterances specifically in the English language involves an embarrassingly and demonstrably foolish approach to addressing these questions.
The field of mainstream analytic philosophy, insofar as its proponents purport to access the deepest insights into the nature of the universe by exclusively analyzing modern English sentences is arguably far more a form of insanity than being a moral antirealist. It is, at the very least, a terrible method, and those who persist in its use are negligent so long as they fail to comport their practices with growing evidence in cognitive science and comparative linguistics, or, insofar as they insist in eschewing such findings as relevant to their practice, persisting in the use of even more questionable methods.
I want to address a few follow-up remarks from Emerson as well:
If not insanity, then disingenuous radical skepticism
Disingenuous? That’s a strange term to describe antirealists. I generally take this term to mean something like “insincere” or pretending to be candid. If this is what Emerson means, this is Emerson suggesting that antirealists are somehow deceitful or insincere?
If so, it’s weird to describe antirealism as a form of insincerity. This remark seems to be corroborated by a subsequent remark:
No. I’m the only honest one
This contrast of disingenuous with honest, even if made in jest, suggests that the intended interpretation of “disingenuous” in some way indicates a form of dishonesty.
Given that I’m most familiar with Emerson as a counter-apologist, I find it especially troubling that Emerson would opt for this characterization of antirealism, when accusations of disingenuity are common among Christian apologists who insist atheists are somehow insincere in their beliefs or attitudes, e.g., that atheists don’t really deny the existence of God, that we all know God in our hearts.
Psychologizing one’s opposition in this way, especially without justification for doing so (which I doubt Emerson has), is an easy way to dismiss views without engaging them: convince yourself that those who disagree with you are merely pretending, and you’ve instantly eliminated genuine opposition, at least to your own satisfaction.
I suspect (or at least hope) Emerson has a fairly negative attitude towards convenient, unfounded accusations that those who disagree with you are being disingenuous. Yet here we have Emerson seemingly doing the same with respect to antirealists as Christian apologists do to atheists.
As for the remark about “radical skepticism”: I simply deny there’s anything radical about antirealism. Antirealists can and should not only deny moral realism, but deny there’s anything radical about doing so, any more than atheism is a form of “radical skepticism.” I deny most people are realists. I deny realism is “the” commonsense view. And I deny, as a result, that my view is in any way unusual or reflective of a highly unconventional way of thinking.
If anything, I think realism is an outlier as far as human beliefs go. Antirealism is, too, but it’s simply a rejection of such a view, so there’s nothing especially radical or weird about denying an unusual view for which there is little good reason to think it is true. It’s no more a form of radical skepticism to deny moral realism than it is radically skeptical to deny Jesus as my personal savior. If that sounds really implausible to you, because you think most people are moral realists, see my other posts: there is, at present, little empirical evidence to support the claim that most people are moral realists.
In any case, the next remark is from alobar, who said:
I’d like to hear you have a chat with Lance Bush. His takes on antirealism are annoyingly hard to ignore
I’d be happy to have a conversation with Emerson about moral realism. Emerson, feel free to get in touch, and alobar is likewise welcome to get in touch to have a discussion on my channel or offline.
Notes
1. Some realists may insist that realism is practically significant, but that’s something we’re free to deny, and I deny it. If there were moral facts, I simply wouldn’t care. Realists are welcome to insist I would care, or to even stipulate that such facts are by definition motivating or something I’d care about. In that case, that’s also an empirical claim, and they’re welcome to demonstrate that I do or would care about the facts in question. I won’t hold my breath. They’ll probably just deny it’s an empirical question, as though facts about human motivation could be settled from the armchair. They can’t, and one can stipulate that the facts they are referring to have all sorts of special powers to impel action in myself or others. If the facts in question have the power to make people think or do anything, they fall within the scope of empirical inquiry.
References
Andringa, S., & Godfroid, A. (2020). Sampling bias and the problem of generalizability in applied linguistics. Annual Review of Applied Linguistics, 40, 134-142.
Beebe, J. R., & Sackris, D. (2016). Moral objectivism across the lifespan. Philosophical Psychology, 29(6), 912-929.
Berniūnas, R. (2020). Mongolian yos surtakhuun and WEIRD “morality”. Journal of Cultural Cognitive Science, 4(1), 59-71.
Blasi, D. E., Henrich, J., Adamou, E., Kemmerer, D., & Majid, A. (2022). Over-reliance on English hinders cognitive science. Trends in cognitive sciences.
Bush, L. S., & Moss, D. (2020). Misunderstanding metaethics: Difficulties measuring folk objectivism and relativism. Diametros 17(64): 6-21.
Chater, N. (2019). The mind is flat: The remarkable shallowness of the improvising brain. New Haven, CT: Yale University Press.
Christiansen, M. H., & Chater, N. (2022). The language game: How improvisation created language and changed the world. New York, NK: Basic Books.
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.
Davis, T. (2021). Beyond objectivism: New methods for studying metaethical intuitions. Philosophical Psychology, 34(1), 125-153.
Dranseika, V., Berniūnas, R., & Silius, V. (2018). Immorality and bu daode, unculturedness and bu wenming. Journal of Cultural Cognitive Science, 2, 71-84.
Feltz, A., cokely, E. T. (2008). The fragmented folk: More evidence of stable individual differences in moral judgments and folk intuitions [paper presentation]. In B.C. Love, K. McRae & V.M. Sloutsky (Eds.), Proceedings of the 30th Annual Conference of the Cognitive Science Society (pp. 1771-1776). Austin, TX: Cognitive Science Society.
Gill, M. B. (2009). Indeterminacy and variability in meta-ethics. Philosophical studies, 145, 215-234.
Goodwin, G. P., & Darley, J. M. (2008). The psychology of meta-ethics: Exploring objectivism. Cognition, 106(3), 1339-1366.
Goodwin, G. P., & Darley, J. M. (2010). The perceived objectivity of ethical beliefs: Psychological findings and implications for public policy. Review of Philosophy and Psychology, 1(2), 161-188.
Goodwin, G. P., & Darley, J. M. (2012). Why are some moral beliefs perceived to be more objective than others?. Journal of Experimental Social Psychology, 48(1), 250-256.
Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world?. Behavioral and brain sciences, 33(2-3), 61-83.
Judd, C. M., Westfall, J., & Kenny, D. A. (2012). Treating stimuli as a random factor in social psychology: a new and comprehensive solution to a pervasive but largely ignored problem. Journal of personality and social psychology, 103(1), 54-69.
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., & Wright, J. C. (2020). Anti-realist pluralism: A new approach to folk metaethics. Review of Philosophy and Psychology, 11, 53-82.
Sinclair, N. (2012). Moral realism, face-values and presumptions.
Sinnott-Armstrong, W. (2009). Mixed-up meta-ethics. Philosophical Issues 19(1), 235-256.
Strickland, B., & Suben, A. (2012). Experimenter philosophy: The problem of experimenter bias in experimental philosophy. Review of Philosophy and Psychology, 3, 457-467.
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.