From The Myth of Morality. Cambridge: Cambridge University Press, 2001.
Joyce argues for an “error theory” regarding moral discourse. An error theory about a given discourse merely claims that sentences in that discourse tend to be used assertorically (i.e., they are used as if they are true), and that by and large such sentences are actually untrue (which is not the same as ‘false’: a sentence can be asserted which is neither true nor false; such sentences are called “Strawsonian” in the literature and include, e.g., ‘The present king of France is wise.”). Joyce thinks this because moral statements are said to provide categorical reasons for engaging them; but, such reasons do not exist. So, there are no real moral claims.
Joyce clarifies that by ‘categorical reason,’ he means one that is in accord with practical rationality. If it were merely in accord with an institution – a set of rules – then there would be categorical reasons for moral acts (assuming morality is a sort of institution). So the categorical reason must be one which transcends any particular set of rules.
Joyce presents what he calls the ‘moral rationalist’s dilemma’: in order to give an adequate account of a categorical reason, it must not be possible for anyone to ask “So What?” of the reason. In other words, it cannot be questioned that the reason applies to the person in question. But if one develops an account that is truly universal, then individual persons can justifiably ask how the reason applies to them. As an example, Joyce discusses Roderick Firth’s account of an ‘idealized observer’: a person who is fully rational and dispassionate. What this observer does, so the argument goes, so too we ought to do. But Joyce contends that any person can wonder why what the observer does pertains to them. Why should one care what a completely dispassionate person would do?
Any attempt to avoid this problem leads to the other half of the dilemma: when one tries to tie normativity (and thus the categorical reason of a moral act) to desire, then it is true that no one will be alienated from their normative reasons. But, the result is relativism: what one ‘ought’ to do will be based upon one’s desires; so, there will be alternative, conflicting accounts of moral oughtness. Again, Joyce procures an example: Michael Smith’s theory of ‘non-Humean instrumentalism.’ On this view, a person X ought to do what X+ (a fully rational, epistemically successful version of X) would prescribe X should do. This resolves the problem of alienation: no one will be able to rationally dispute the reasons offered by such idealized selves, and thus such reasons will be categorical. But, each person will have different desires, and thus different idealized selves. In other words, X+ and Y+ will probably not prescribe the exact same acts, despite their being fully rational. So moral ‘oughts’ will be relative to the desires of idealized selves. Thus, it seems there are no categorical reasons one way or another, and thus no moral reasons. Moral discourse is founded upon an error.
Joyce then tackles what Smith thinks is good evidence that normativity (and thus moral ‘oughts’) is non-relative: the fact that there is as much moral agreement as disagreement, that moral disagreement of the past has been successfully resolved via rational means, and that moral disagreement is merely the result of one side or the other failing to act rationally. Joyce thinks all of these reasons, even if true, would be inadequate to show that normativity is non-relative. For example, even if there were complete agreement on norms, that convergence could be a result of any number of things besides rational resolution. It could be the case that cultural pressure is forcing hegemony of norms on the world. Or, it could be some other process like the convergence of taste in food. But Smith is dedicated to the view that such convergence is rationally driven. So, his view is far from vindicated.
Further, Joyce does not think that the thesis that moral disagreement occurs solely because of irrationality on one party or another is sufficiently justified. Must it be the case that any moral failing is equivalent to a rational failing? It does not seem so, or in any case, the view that it is needs support in order to be justified. So it seems Smith’s attempt to justify non-relative normativity (and thus non-relative morals) flounders.
Primary Argument: There are no real moral claims
1. In order for an act to be ’moral,’ it must necessarily provide a reason
for carrying it out.
2. No actions necessarily provide reasons for carrying them out.
S1. For an action to necessarily provide a reason for carrying it
out, that reason must adhere to practical rationality for all persons.
S2. In order for a reason to adhere to practical rationality for all
persons, it cannot rationally alienate anyone, nor can it collapse into relativism.
S3. Therefore, for an action to necessarily provide a reason for
carrying it out, it cannot alienate anyone, nor can it collapse into relativism. [S1,S2]
S4. The most plausible accounts of such a reason are ideal observer-based or desire-based.
S5. Ideal observer-based accounts alienate persons.
S6. Desire-based accounts collapse into relativism.
S7. Therefore, the most plausible accounts of such a reason either
alienate persons or collapse into relativism. [S4-S6]
S8. Therefore, no actions necessarily provide reasons for adhering
to them. [S7,S3]
3. Therefore, there are no moral acts.
Primary Argument: There are no moral reasons
1. M → N
2. ¬N…Therefore, ¬M
3. ¬M (2, 1,MT)
Where M=There are moral reasons and N=Moral reasons are necessarily
provided by moral acts.
Sub-Argument: No actions necessarily provide reasons for carrying them
1. (∀x)[Ax → (¬Ux ∨ ¬Rx)]
2. (¬Ai ∧ ¬Ad) → ¬N
3. Ui ∧ Rd…Therefore, ¬P
4. Ai → (¬Ui ∨ ¬Ri) (1, UI)
5. Ui (3, SIMP)
6. Ui ∨ Ri (5,ADD)
7. ¬Ai (6, 4,MT)
8. Ad → (¬Ud ∨ ¬Rd) (1, UI)
9. Rd (3, SIMP)
10. Rd ∨ Ud (9,ADD)
11. Ud ∨ Rd (10,COMM)
12. ¬Ad (11, 8,MT)
13. ¬Ai ∧ ¬Ad (7, 12,CONJ)
14. ¬N (13, 2,MP)
Where Ax=x is an adequate account of categorical reasons, Ux=x is
universalized in a way which alienates persons, Rx=x collapses into
relativism, N=There are necessary reasons for certain acts, i=An account of
categorical reasons based on idealized observers, and d=An account of
categorical reasons based on desires.