Lewis on Quasi-Realism
In "Quasi-Realism is Fictionalism" ((Lewis 2005)), Lewis seems to suggest that Blackburn's quasi-realism about moral discourse is a kind of fictionalism. The suggestion is bizarre. Has Lewis made silly mistake? (Spoiler: No.)
Let's compare what quasi-realism and fictionalism say about moral discourse.
Blackburn's quasi-realism (as presented, e.g., in (Blackburn 1984, ch.6) and (Blackburn 1993)) is a brand of expressivism. According to Blackburn, moral statements like (1) don't serve to describe special facts, but to express moral attitudes.
The exact nature of moral attitudes won't matter, except that they are not beliefs.
Fictionalism is harder to pin down. Different authors give different definitions; Lewis gives none. But we get a sense of what he has in mind. According to Lewis, a fictionalist is disposed to utter sentences of a certain type even though she doesn't believe that they are true, understanding them as tacitly "prefixed" or "prefaced" by a disclaimer which cancels the commitment to truth. Lewis cites (Joyce 2001) as an example of moral fictionalism. Joyce suggests that we should keep uttering things like (1), but clarify – when pressed in the philosophy seminar – that these utterances are only pretend-assertions, not real assertions: that we only make-believe what we say.
These two views about (1) are obviously different. The quasi-realist does not think that (1) is, strictly speaking, false, but that we may nonetheless utter it with an understanding that we don't really believe what it says. (Blackburn 2005) repeats this point at length, in response to Lewis, but the point should have been thoroughly clear from Blackburn's other writings. How could Lewis have missed it?
So far, we've only looked at the title of Lewis's essay ("Quasi-Realism is Fictionalism"). Blackburn's response barely engages with the content. Does Lewis have an argument for his surprising claim?
If we skim the paper for such an argument, we find on p.319 what seems to be the central argument:
Blackburn's quasi-realism is just this kind of moral fictionalism. […] One of Blackburn's avowed aims is to earn the right to say what the 'moral realist' does: that means either being or make-believedly being a realist. Another of his avowed aims is to avoid the realist's errors: that means not being a realist. Taking these aims together, he aims to make-believedly be a moral realist.
As (Jenkins 2006) points out, this argument seems to rest on the false assumption that realism and fictionalism are the only options: either (1) expresses belief in a mind-independent moral fact, or it expresses make-believing such a fact. Expressivism denies both.
What a silly mistake!
Well, let's stop skimming and have a closer look at what Lewis actually says. I think it's clear that Lewis isn't talking about statements like (1) when he says that the quasi-realist wants to "say what the 'moral realist' does".
Lewis puts 'moral realism' in scare quotes because he uses it in a technical sense. This is explained on p.315f.:
Let us […] reserve the name 'moral realism' for a moral theory that is committed to [a distinctive error]: that there are properties […] such that we can detect them; and such that when we do detect them, that inevitably evokes in us pro- or con-attitudes towards the things that we have detected to have these properties.
So here is what the 'moral realist' says:
When Lewis says that Blackburn's avowed aim is to earn the right to say what the 'moral realist' does, I think he obviously meant things like (2), not things like (1), which are not at all distinctive of 'moral realism'.
As the first paragraph of the paper makes clear, Lewis is interested in a well-known puzzle: if the quasi-realist really echoes everything the realist says, if "he even echoes all the realist says about moral psychology and metaethics" (p.314), how is the position different from realism?
This puzzle does not arise for old-fashioned expressivism (as in (Ayer 1936), for example). Old-fashioned expressivism is easy to tell apart from realism by that fact that it declares, for example, (3) and (4) to be false.
The aim of Blackburn's quasi-realist program is to extend the expressivist semantics so as to vindicate our apparently realist moral discourse, including statements like (3) and (4). Lewis's paper begins with the supposition that this program succeeds: that the expressivist semantics for (1) can be extended to (3) and (4) and beyond, up to the point where the quasi-realist has "earned the right to echo everything the moral realist says" (p.314).
One might wonder whether Blackburn really wants to vindicate statements like (2). In his response to Lewis, Blackburn distinguishes between our practice itself and speculative philosophical theorizing about our practice. (2) looks like a piece of philosophical theory, rather than something that's integral to our ordinary practice. The aim of quasi-realism, he explains (on pp.331f. of (Blackburn 2005)), is not to vindicate erroneous philosophical theories.
But the example Blackburn gives of an erroneous philosophical theory is not like (2). And there is certainly pressure towards vindicating (2). Blackburn explicitly does want to vindicate statements like (5) and (6) and (7) and (8).
Isn't (2) just a logical consequence of statements like these? Aren't we allowed to draw the inference?
In any case, if 'moral realism' subscribes to (2) and quasi-realism does not then there is no puzzle. We could easily tell apart the two views by what they say about (2). The premise of Lewis's paper is that realism and quasi-realism can't be told apart in such a simple way.
By hypothesis, then, the quasi-realist is prepared to accept and utter (2). But doesn't the quasi-realist also want to deny (2)? To make the point even more obvious, consider (9).
It is only a short step from (5) and (6) and (2) and uncontroversial facts about Hume, to (9). So 'anti-realism' endorses (9), and quasi-realism – ex hypothesi – follows along. But wouldn't Blackburn want to deny (9)? At the level of belief: doesn't Blackburn believe that Hume's projectivist account is essentially right?
Again, one might respond that quasi-realism was never meant to be that far-reaching: the quasi-realist only wants to echo harmless statements like (3) and (4) and (5), not things like (2) and (9). And again, it will be hard to draw the line, and Lewis simply sets this possibility aside, because it doesn't raise an interesting puzzle. Let's set it aside as well.
Another thing the quasi-realist could do is accept (2) and (9) as true, and leave it at that. Projectivism, she might say, was the ladder on which we've climbed to quasi-realism, but once we have climbed the ladder, we had to kick it away. We have to disavow projectivism, and presumably expressivism, and naturalism.
This looks deeply unappealing. Is there any other way out?
There is: fictionalism!
The quasi-realist might utter (2) and (9), in a suitable context. But she might also clarify, when pressed in the philosophy classroom, that she doesn't genuinely believe (2) and (9).
That's why quasi-realism is fictionalism. The argument goes like this.
- 'Moral realism' is committed to statements like (2).
- If quasi-realism succeeds, it licenses uttering these statements.
- But the quasi-realist doesn't believe things like (2).
- So the quasi-realist only make-believes things like (2).
This is not unlike the argument quoted above. And it's not a stupid argument.
The crucial point is that (2), unlike (1), is a suitable object of genuine belief, even by the quasi-realist's lights. Quasi-realism is motivated, to a large extent, by the belief that (2) is false. If the quasi-realist's utterance of (2) were to express genuine belief, she would believe that (2) is both true and false.
(It wouldn't help much to declare, implausibly, that (2) expresses a moral attitude. The quasi-realist who doesn't want to kick away the ladder still wants to reject (2) when explaining and motivating her position.)
Here is Lewis's concluding comment, coming right after the argument:
[Quasi-realism] earns the right to agree with all the moral realist says in just the same way explicit fictionalism does, whether or not it goes on to earn that right twice over by offering its special semantics. (pp.319f.)
Lewis here acknowledges that the quasi-realist may already have a way to agree with the realist by means of their special (expressivist) semantics, and that this way is distinct from the fictionalist way. (Hence 'twice over'.) So Lewis clearly didn't think, as Blackburn and Jenkins and other commentators assume, that the quasi-realist interpretation of moral discourse amounts to a fictionalist interpretation. (Jenkins at least raises this as a puzzle, at the end of her paper.)
How does the quasi-realist's special semantics explain why it's OK to utter (2)? I don't know. To my knowledge, Blackburn has nowhere offered a sufficiently detailed semantics (or semantics+pragmatics) that would cover statements like (2).
In fact, one might suspect that the quasi-realist's special semantics only covers statements which express moral attitudes. Since (2) doesn't express a moral attitude, it would follow that the special semantics doesn't license uttering (2). So it wouldn't license echoing everything the realist says. Quasi-realism would have failed on its own terms, as Lewis understands them.
This means that quasi-realism may require fictionalism not only to remain consistent, but also to fulfil its ambition of echoing realism. The special semantics does that job for moral statements, but perhaps not for statements like (2). Here, fictionalism may be needed to fill the gap.
I think that's why Lewis says "whether or not it goes on to earn that right twice over by offering its special semantics", suggesting that it's an open question whether the special semantics alone is enough.
So. Is the quasi-realist account of first-order moral discourse a fictionalist account of that discourse? Of course not. But if the quasi-realist wants to echo more than the realist's first-order discourse, if she also wants to echo more theoretical statements that may seem to follow from our first-order discourse, then she arguably must endorse a kind of fictionalism.
That was Lewis's point.
"Isn't (2) just a logical consequence of statements like these? Aren't we allowed to draw the inference?"
In order for this sentence not to beg the question against Blackburn, it has to be taken in quite a specific form. (2) specifically has to follow from (5)-(8) under the logic and semantics that Blackburn provides for statements like (5)-(8). Certainly, if (5)-(8) are treated like sentences expressing beliefs, treated according to normal basically-classical logic, then (2) seems follow. But Blackburn explicitly doesn't want to treat (5)-(8) in this way; and this has an effect on what inferences we are allowed to draw.
I think – though it's been a long time since I've read Blackburn – that the semantics he gives for ethical statements actually results in a much weaker logic (or 'logic'). And without having any Blackburn books on hand to check, I wonder if that doesn't block the inference from (5)-(8) to (2). Certainly it seems like it *should* block this inference, if the semantics are to do the work Blackburn requires.
If you can't conclude (2) from (5)-(8) _within Blackburn's logic_, then this whole line of argument breaks down. When you ask Blackburn "Aren't we allowed to draw the inference?", he can respond "No, and I've explained why. You can disagree with me on this, but that just means disagreeing with my quasi-realist project; assuming quasi-realism, this inference is blocked, and I'm able to go on being not-a-fictionalist with no tension."
Obviously this answer requires that Blackburn's semantics for ethical statements actually block the inference. But this seems right: Blackburn's ability to distinguish quasi-realism from other positions depends crucially on his ability to give an account of moral inference that is powerful enough to allow us to do moral reasoning, but not so powerful that it allows us to reach metaphysically realist conclusions. If the logic is too weak, he collapses into an old-style expressivist; if it's too strong, he has to be either a realist or an error theorist. Blackburn says as much himself, albeit not in these exact words.