Lately, I’ve been reading Richard Joyce’s The Myth of Morality, and it’s impressive. As the title suggests, Joyce argues in favor of an error theory concerning moral discourse: he claims that rational inescapability is an ineliminable component of it, that this feature is unsatisfied by the world we find ourselves in, and that therefore moral discourse is hopelessly flawed. From there, he moves on to discuss why this error should be so widespread, finding an answer in natural selection. Lastly, Joyce speculates that moral discourse might, despite its flaws, nevertheless be worthwhile continuing on with – he advocates a fictionalist stance that purports to preserve some of the benefits of moral discourse (such as combating akrasia), without committing to belief in moral propositions. I’ll be blogging my way through the book, in this and upcoming posts.
Setting the Stage
The Myth of Morality opens with a discussion of the Polynesian concept of ‘tapu’, which is to be shown an appropriate subject of error theory. Tapu, we are told, “centrally implicates a kind of uncleanliness or pollution that may reside in objects, may pass to humans through contact, may be then transmitted to others like a contagion, and which may be canceled through certain ritual activities, usually involving washing“(p.1). But whilst primitive Polynesians have certainly thought this concept to have application, the modern Western mind denies this: the former assert some sentences of the form “Φ is tapu”, whereas the latter think that these sentences are false1. That is, the modern Westerner adopts an error theory concerning tapu-discourse. Yet perhaps another option is open to the Westerner here. The fact that nothing satisfies the Polynesian concept of tapu need not be fatal to the discourse, for, if there is something in the vicinity which comes close to fitting the bill (Western moral forbiddeness?), one might instead redefine ‘tapu’ such that the concept does have application. This raises the question of just when redefinition might be the appropriate response, as opposed to adopting the error theoretical stance. Joyce’s answer to this problem involves distinguishing essential from dispensible facets of a concept. For example, though the Polynesians might say much about tapu, there are presumably some propositions central to the concept for the sake of which it has gainful employment. Such propositions are ‘non-negotiable’, and may be identified by a test of translation. Here is Joyce applying the test to the discredited concept of ‘phlogiston':
“Imagine that we were to encounter a population speaking a quite different language to our own, most of which we have translated and tested to our satisfaction, and we find that they have a concept that appears rather like our concept of phlogiston (say, it plays a central role in explaining combustion and calcification) – call their term “schmogiston” – but we also find that they don’t endorse one of the three propositions about schmogiston. If that would be sufficient for us to decide not to translate “schmogiston” into “phlogiston,” then the proposition in question must be a non-negotiable part of our concept phlogiston. It may not be that any one proposition is non-negotiable: perhaps we would be content with the translation if any two of the “schmogiston”-propositions were dissented from, but if the speakers dissented from all three (i.e., they said “No” to “Is schmogiston released during combustion?”, “Is schmogiston stored in bodies?”, and “Is soot made up of schmogiston?”) then we would resist the translation – we would conclude that they weren’t talking about phlogiston at all. In such a case we might call the disjunction of the three propositions “non-negotiable.”“~ Joyce, p.3.
Of course, such a test is difficult to carry out in practice. But the point here is not so much to lay the groundwork for empirical investigation as to provide a criterion in thought for the core commitments of a concept. With this in mind, we can more easily decide what attitude to take to moral discourse: if at least the non-negotiable propositions implied by moral concepts are true, then some suitably revised moral realism is in order; if not, then an error-theoretical stance is appropriate. To sum up the method, then..
“An error theory, as we have seen, involves two steps of argumentation. First, it involves ascertaining just what a term means. I have tried to explicate this in terms of “non-negotiability,” which in turn I understood in terms of a translation test (but there may be other, and better, ways of understanding the notion). So, in artificially simple terms, the first step gives us something roughly of the form “For any x, Fx if and only if Px and Qx and Rx.” We can call this step conceptual. The second step is to ascertain whether the following is true: “There exists an x, such that Px and Qx and Rx.” If not, then there is nothing that satisfies “. . . is F.” Call this step ontological or substantive. The concept of phlogiston – with its commitment to a stuff that is stored in bodies and released during combustion – and the concept of tapu – with its commitment to a kind of contagious pollution – do not pass the test.” ~ Joyce, p.5.
After some discussion on the nature and shortcomings of non-cognitivism – which he takes to founder on the fact that moral language is used assertorically (i.e. to express belief) – Joyce examines one fleshing out of an error-theory, which he ultimately takes to be unpersuasive. Instead of following him in this, I’ll jump directly to Joyce’s own positively contrasting argument in the next post.
1 Why ‘some’, rather than ‘all’? Because sentences which use the terms of the erroneous discourse to denounce it would be true (“Nothing is tapu”), as could be sentences ascribing belief (“The Poynesians believe in tapu”), and so too could be sentences expressing hypotheticals.