1.1 The Generic Argument for Moral Error Theory
You are most probably an error theorist about our thought and talk about witches, or our ‘witch discourse’. You believe that there can only be witches if there are women with supernatural powers, you believe that no one can have supernatural powers, and so you believe that there are no witches (Smith 1994, p. 64). You also believe that anyone who utters a judgement like ‘Annie is a witch’ makes the systematic error of uttering a judgement that is truth-apt (i.e., can be true or false) but systematically fails to be true.
In this book, I argue that you should be an error theorist about our thought and talk about morality, or our ‘moral discourse’, too (Mackie 1977; Joyce 2001; Olson 2014). You may not yet, but you should believe that there can only be morally obligatory (or morally wrong, or morally right, or morally permissible) actions if there are categorical moral reasons to perform these actions, you should believe that there are no such reasons, and so you should believe that there are no morally obligatory (or morally wrong, or morally right, or morally permissible) actions. You should also believe that anyone who utters a judgement like ‘giving to the poor is morally obligatory’ makes the systematic error of uttering a judgement that is truth-apt but systematically fails to be true.
Categorical moral reasons are reasons for agents to perform a moral action regardless of whether performing that action satisfies one or more of their desires. For instance, Beatrice has a categorical moral reason to give to the poor regardless of whether doing this satisfies her desire to be praised by others (Mackie 1977, p. 35; Joyce 2001, p. 5; Dreier 1997, p. 84; Olson 2010, pp. 64–65). In this book, reasons are always normative reasons in the standard sense of considerations that count in favour of or against performing an action (Scanlon 1998, p. 17). For example, if you are thirsty then “the consideration that drinking would feel good to you” is a normative reason for you to drink (Scanlon 1998, p. 38). The intended contrast is with motivating reasons, which are considerations that motivate agents to perform an action but need not also count in favour of or against the performance of that action (Parfit 1997, p. 99). For instance, and ignoring non-standard cases in which believing false propositions is all-things-considered beneficial for you (for instance, because knowing the truth will drive you insane), you have a merely motivating reason if you accept the consideration that a proposition’s falsity counts in favour of believing it. Reasons can either be practical, in which case they are typically reasons to perform an action, or theoretical, in which case they are typically reasons to form a belief. I explain how I think about reasons in more detail in §4.4.
The term ‘error theory’ is sometimes used to denote just a local error theory about morality and sometimes error theory about normativity as such. The latter view has recently been defended by Bart Streumer (2017). Such a global error theory extends to, for example, prudential normativity, which includes reasons to eat healthy and to get enough sleep, and epistemic normativity, which includes the reason to believe all propositions for which you have conclusive evidence that they are true. In this book, error theory always denotes moral error theory unless specified otherwise.
The most generic argument for error theory runs as follows:
Moral discourse is our practice of uttering moral judgements in moral discussions with our peers about such issues as the permissibility or otherwise of euthanasia and abortion , of uttering moral judgements to teach our children what ought to be done and how to muster the motivation to act on these judgements, of adjusting our moral judgements in light of other people’s arguments, of reminding someone of a promise, of praising someone’s altruistic behaviour, et cetera. I use the terms judgement, claim, sentence, and utterance interchangeably, such that in my vocabulary, all judgements, claims, and sentences are uttered or otherwise communicated, for instance by being inscribed. The relation between moral judgements and fully private moral thoughts will not be at the forefront of my analysis in this book, though it stands to reason that if moral judgements are systematically untrue , then the moral thoughts that these judgements express are also systematically untrue (Väyrynen 2013, p. 44).
P1 is error theory’s “non-negotiable commitment ” claim (Joyce 2001, p. 3). To understand what this means, consider again error theory about witch discourse. Error theorists about witch discourse have to show that this discourse must be about women with supernatural powers in order to be discourse about witches and not something else entirely, such as spinsters. Similarly, moral error theorists have to show that a particular commitment, such as a commitment to categorical moral reasons, is essential to moral discourse for it to be discourse about morality and not something else entirely, such as prudential values. For suppose that error theorists do not have a good argument for P1. In that case, some moral success theorists, who think that moral discourse does not commit the systematic error that error theorists attribute to it and who think that moral discourse is successful as it is, can argue that even if there are no categorical moral reasons, moral discourse was never committed to their existence in the first place, and thus that at least some moral judgements are true (Sayre-McCord 1986). Moral success theorists who deny P1 often think that moral discourse is just committed to hypothetical reasons, which are reasons that an agent has because she has a certain desire that will be promoted if she acted on that reason, and which are much easier to defend philosophically (Railton 1986; Brink 1989).
P2 is error theory’s “substantive claim” (Joyce 2001, p. 5). Error theorists have to show that there are no categorical moral reasons. This is similar to the task that error theorists about witches have when they must show that there are no supernatural powers. For suppose that error theorists do not have a good argument for P2. In that case, some other moral success theorists can claim that although moral discourse is committed to the existence of categorical moral reasons, such reasons exist, and therefore that some moral judgements are true (Enoch 2011).
P3 is simply an instance of the more general and obviously correct thesis that any claim that carries a false non-negotiable claim is untrue (Streumer 2007, p. 255). For error theory about witch discourse, the judgement ‘there exists a witch’ is truth-apt and carries a non-negotiable commitment to the claim that there is at least one woman who has supernatural powers. But the last claim is false, and so the judgement ‘there exists a witch’ must also be untrue. It cannot be the case that the claim ‘there exists a witch’ is true but that the claim ‘there exists a woman with supernatural powers’ is not true and that the truth of the first claim requires the truth of the second. The same holds for moral discourse. The judgement ‘giving to the poor is morally obligatory’ is truth-apt, and it carries a non-negotiable commitment to the claim that there are categorical moral reasons; but that claim is false, and so the judgement ‘giving to the poor is morally obligatory’ is untrue. It cannot be the case that the claim ‘giving to the poor is morally obligatory’ is true but that the claim ‘there exists at least one categorical moral reason ’ is not true and that the truth of the first claim requires the truth of the second claim.
P1, P2, and P3 entail C1. Yet from the truth of C1 we cannot deduce the truth of C2 (Evans and Shah 2012, p. 85). That is, we cannot conclude from the fact that moral judgements are untrue that our entire moral discourse is in error. This inference is unwarranted, first, because the Generic Argument does not specify whether all or only some moral judgements are untrue. Perhaps the deontic part of morality with judgements like ‘X is morally obligatory’ and ‘X is morally right’ is systematically untrue, but what to say of judgements about moral goodness or judgements involving thick concepts like courageous and virtuous? If such non-deontic moral judgements are true, then it becomes at least questionable to conclude from the fact that deontic moral judgements are untrue that moral discourse as such is in error. A second reason why this inference is unwarranted is that even if error theory extends from deontic to many other kinds of moral judgements, it remains to be seen whether negated and modal moral judgements such as ‘X is not morally obligatory’ and ‘X might be morally obligatory’ are systematically untrue (Kalf 2013, 2017). For again, if these moral judgements are not systematically untrue, then it seems that we cannot reach the conclusion that our entire moral discourse is in error. P4, if true, bridges this gap.
Generic Argument for Error Theory is the most generic formulation of error theory because it contains the largest number of placeholder terms for which error theorists must find a more precise meaning. For instance, traditionally error theorists have said that moral judgements are untrue because they are “all false” (Mackie 1977, p. 35; Clark 2009, p. 200). Alternatively, we can say that they are untrue because they are “neither true nor false” (Kalf 2013, p. 925). Error theorists must find interpretations of the placeholder term ‘untrue’, and also for the other placeholder terms in the Generic Argument, because these terms require further unpacking before we have an argument for error theory. Another example of a placeholder term in Generic Argument is ‘claim N’. We can’t argue that moral error theory is true because we know that it is committed to ‘claim N’ and that ‘claim N’ is false. We first need to know what ...