It's possible to believe normative error theory
Kane B is a philosopher who posts videos on YouTube about a variety of philosophical topics. I consider it one of the best channels on YouTube, and encourage you to check it out. You can find it here.
Kane B produced a video that focuses on Streumer’s argument that (a) normative error theory is true, but (b) we can’t believe normative error theory. You can see that video here.
I think it’s a mistake to focus on normative error theory rather than normative antirealism. There is no problem with believing that there are no stance-independent normative facts. If Streumer’s use of “normative property” means the same thing as “stance-independent normative fact,” or is sufficiently similar, then I see no problem with believing that there are no such things.
The normative antirealist can deny both (a) that there are stance-independent normative facts and (b) that we can't believe this if we deny there are normative reasons for us believing this.
Normative antirealists can simply reject any form of normative error theory that requires the antirealist take on board some of the assumptions baked into the dispute between normative realists and antirealists. I see this same issue time and again. A dispute between positions X and anti-X will be framed in such a way that the anti-X position is expected to accept some of the same background assumptions, methodological assumptions, metaphilosophical presuppositions and so on, such that the anti-X position ends up facing some kind of serious problem in virtue not of rejecting X, but rejecting these background assumptions. This is then taken to create some kind of problem for the anti-X position. But a thoroughgoing anti-X position can both reject X and reject some or all of the relevant background assumptions and presuppositions in which the dispute is framed. This is my approach as a normative antirealist. For instance, I don’t agree with Streumer’s account of the conditions necessary for a belief, and I doubt I share Streumer’s views about “reasons” and their role in belief. As a result, I may not be able to have the sorts of beliefs Streumer describes, but I don’t particularly care, because that isn’t a kind of belief I’m troubled with. I have a different conception of beliefs, and I don’t think those are threatened by Streumer’s arguments or are in some way not possible for someone who endorses error theory.
Also, so long as we take “beliefs” to involve assent to the truth of propositions, I take the conditions necessary for beliefs are empirical questions about our psychology; one cannot conclude a priori that if certain conditions aren’t met we “can’t” believe something: it could be the case that our psychology just doesn’t conform to such a priori considerations, and that people can believe things without meeting the conditions Streumer outlines. I would think it would make more sense to say that such beliefs are unjustified or in some other way defective, but it seems bizarre to say we “can’t” have such beliefs - how would Streumer know that without empirical evidence? Unless we simply stipulate that by “belief” we just mean something that meets certain stipulated conditions, or we attempt to show that on analytic grounds “belief” just means something that meets those conditions. I doubt efforts to do this would prove successful in the case of beliefs, though.
Streumer seems to be following the traditional framework of depicting normative antirealist views in a way that constrains them to particular semantic theses that are uniform and determinate. What I mean by this is the following. Typically, *all* normative judgments in a given normative domain are supposed to be either (a) systematically false (b) relative (c) express nonpropositional claims. This leaves the antirealist with three options: error theory, relativism, or noncognitivism, respectively. Huemer explicitly argues that these are the only possible positions available to antirealists in his book, Ethical Intuitionism.
The idea here seems to be that there is such a thing as a “normative judgment” or a “normative claim.” And such claims have specific properties. Namely, they either have the property of referring to stance-independent facts, or the property of being relative because they index or refer to one or another of different standards, or they have the property of expressing some nonpropositional content. Much as we might suppose that all hydrogen atoms have some set of properties, and that it is these properties that make them hydrogen atoms (rather than, say, gold atoms or carbon atoms), so too are normative judgments like this: they have some specific set of properties.
The problem, however, is that philosophers are reifying ordinary thought and discourse in such a way that they imagine that some set of terms people may use, in this case, normative language, must all necessarily refer to the same thing. After all, any competent person talking about the “sun,” is, if not joking or speaking elliptically or metaphorically or whatever, presumably referring to the same thing (i.e., the sun). So, too, must we suppose that anyone saying something like “you shouldn’t do that” or “that’s bad,” must be using “shouldn’t” and “bad” to mean the same thing as one another.
There are a few problems with this framing of metaethical disputes. First, such accounts often leave unspecified what account of meaning is presupposed. Are these accounts supposing that normative language has some kind of meaning independent of what people intend to communicate when using normative language? That is, do such accounts presuppose that words and terms mean things independent of the communicative intent of speakers? If so, I reject this conception of meaning, in which case I reject the underlying presuppositions behind the framing of the different options available to antirealists.
In other words, I don’t think words mean things, I think people mean things. Words are just a way for people to express what they mean. The words themselves don’t literally mean anything independent of usage. As a result, I don’t think normative language has any meaning outside context of usage, and to the extent that we could say the words mean things, this is a nonliteral way of saying that the speakers using these words mean things. But what do the speakers using these words mean? That is an empirical question. And there is no reason to suppose, on a priori grounds, that everyone using normative language must necessarily share the same set of metaethical presuppositions.
As such, the question of whether normative judgments are beliefs that ascribe normative properties makes no sense. Which normative judgments are we talking about? The claim seems to reify normative judgments as though they are a specific thing with a specific, shared meaning that is true independent of the potential variation in the communicative intent of speakers, a notion that I don’t buy into. In other words, the claim that normative judgments are beliefs that ascrib normative properties appears to presuppose a conception of normative judgments that I reject: there is no such thing as a “normative judgment” of this kind; there are “normative judgments” in the real world, but those judgments are picked out by some stipulative operationalization that captures a range of discourse and mental states that may, in some cases, involve beliefs about normative properties, and in other cases they may not. I reject the presumption that there is any such thing as “normative judgments” that necessarily share the same set of metaenormative characteristics. Rather, I think people use certain kinds of terms, such as “should,” “ought,” “good,” and so on, and when they use these terms, they intend to communicate various sorts of things, which
(a) don’t necessarily have to share the same set of metanormative presuppositions
(b) in many if not most cases don’t require any metanormative presupposes at all, and are literally indeterminate with respect to which metanormative presuppositions are implicit in such remarks, for the simple reason that there simply aren’t any such presuppositions implicit in these remarks and
(c) to the extent that such presuppositions are present, they vary from one context of usage to another, such that to the extent that there is determinate implicit metanormative presuppositions, they are variable.
In short, I endorse indeterminacy and variability with respect to the metanormative presuppositions that figure into ordinary use of normative terms and language. There simply is no categorical, conceptual, a priori answer as to whether or not “normative judgments are beliefs that ascribe properties” - such a remark either presupposes a false conception of language and meaning, or to the extent that it does not and instead comports with my conception of meaning, such a remark is empirical, and doesn’t admit of a single uniform and determinate answer. It’s a bit like claiming that “fruits are bananas.” Some are, and some are not.
This is important, because if some normative judgments are beliefs that ascribe normative properties, and some aren’t, this frees us, as antirealists, from lacking the linguistic and conceptual tools to both deny normative realism and do so in a way that isn’t self-defeating or that entails that we literally can’t believe that there are no stance-independent normative facts. This is because, once we acknowledge that “normative judgments,” can come in both realist and antirealist forms, rather than in exclusively realist forms, we are now free to appeal to the antirealist conceptions of normative judgments to reject realist conceptions of normative judgments, without thereby succumbing to or requiring appeal to the conceptual resources that would only be available to us on the realist conception of normative judgments.
What I mean is this. The antirealist can reject noncognitivism and relativism. This, traditionally, would leave them no choice but to endorse “error theory.” But “error theory,” in its classical construal, relies on the false notion that *all* “normative judgments,” must, of conceptual necessity, consist of “beliefs that ascribe normative properties” - and these normative properties are precisely the kinds of normative properties the error theorist doesn’t believe in. At the same time, such error theorists are traditionally committed to the view that one may only have a reason to believe something if there are normative properties that would entail that one has a reason to believe it, and, as a result, the error theorist has no reason to believe the error theory.
Yet the “error theorist,” does not have to think that *all* normative judgments are beliefs that ascribe normative properties. Instead, they can endorse an error theory that partitions off only those normative judgments that are beliefs that ascribe normative properties of a particular kind, claim that there no properties of that kind, and thereby claim that all normative judgments that carry the presupposition that there are normative properties of that kind are false. This still leaves the “error theorist” with the conceptual and terminological resources to maintain that they can make normative judgments according to which antirealism about this particular conception of normative judgments is true without having to appeal to the type of normative judgments entailed by the account that they’re rejecting.
For comparison, suppose most of the population believed that all normative facts were facts about God’s commands. Suppose also that there is no God. Such statements would thereby be false. One could make the same argument Streumer has in this case: normative judgments are beliefs about God’s commands. God doesn’t exist. Therefore, all normative judgments are false. You could then argue that one can only believe normative judgments are false if there is a normative fact that would give you reason to believe this. In this case, such a fact would have to be a decree from God. And yet since God doesn’t exist, there is no fact, and thus you have no reason to believe there are no normative facts. If we supposed that such facts are necessary for belief, it is literally impossible to believe in this error theory.
Yet all of this relies on the notion that *all* normative judgments were judgments about God’s commands. Why couldn’t someone instead think the following: normative judgments are judgments about what one ought to do or not do, what’s good or bad, and so on. Some people maintain that such judgments are judgments about God’s commands, while I believe they are something else. These other people’s normative claims are all false, because God does not exist, but my normativer claims are not necessarily false, because I’m not referring to God’s commands. I am referring to something else.
One might say in this case that I am “changing the subject,” and that the meaning of normative claims is fixed or determined in some way by ordinary usage, such that if everyone else uses normative language to refer to God’s commands, then that’s just what normative language means. Okay, let’s suppose that’s the case, and I agree that I can’t make “normative judgments” that aren’t judgments about God’s commands. Fine. I can still deliberate, weigh different considerations, consider my goals and values, and act in accordance with what will achieve my interests.
If I am barred from access to normative language because we’ve agreed that such language will be used to refer to God’s commands, I still have access to all the decision-making, and action-guiding considerations that I would have were people to use normative language in some other way. I can simply declare such language “zormative,” and use my “zormative judgments” to make “zormative” claims about what I “zould” and “zouldn’t” do, what’s “zood,” what’s “zad,” and so on. And this paranormative analog to genuinely normative thought and language can allow me to say, think, and do all the things I’d like to say, think, and do. If this leaves me without the ability to “believe” that there are no normative facts, fine: I can “zelieve” it, which involves assent to the truth of the proposition that there are no normative facts. The fact that I can’t “believe” it is little more than a trivial and uninteresting fact about what words wouldn’t be available to me provided I agree via some kind of intersubjective agreement not to use certain terms in certain ways. What I do not accept, however, is that I must necessarily lack a psychological state, a “belief,” that could go by what amounts to the same description *other than* the labels we use to refer to it.
I am dashing these remarks off on the fly, and so there may be some errors and difficulties to work out, but what I am getting at is that Streumer is following the mainstream analytic philosophical approach to determining what’s true about the world in a way that seems to treat language and metaphysics as having some kind of strange symbiotic relationship, whereby facts about ordinary usage somehow dictate or constrain not merely what I can say, but what is literally true about the world, as though our language creates reality. There’s something profoundly strange about the way analytic philosophers think about the relation between language and metaphysics, and I think the only way to adequately assess claims like Struemer’s is to dig more deeply into the underlying metaphilosophical presuppositions behind such ideas.