Problems with Moral Expressivism: the Frege-Geach Problem

If you're already familiar with expressivism, you can skip to the next paragraph. What is expressivism? Roughly, this:

• Expressivism is a view about moral judgments or moral statements, like "you ought to give more of your money to aid those living in poverty." According to expressivism, moral judgments aren't capable of being true or false in the way that various non-moral judgments are (e.g., "there is a squash in the fridge"). Moral statements don't aim to represent or describe the world. Rather, they aim to express an attitude or emotion. The judgment that "you ought to give more of your money to aid those living in poverty" is really just an expression of one's approval or positive attitude toward the act of giving more money to those in poverty. So the use of any moral predicates whatsoever -- "is good", "is bad, "is obligatory", "is a virtue", etc. -- don't refer to actual properties in the world. They signal attitudes. Expressivism is sometimes motivated by an interest in a lean, naturalistc ontology. Moral properties and facts seem strange and hard to reduce to purely naturalistic properties/facts. So the expressivist excises them from her ontology. She's then left with the task of explaining what moral discourse is about. The preceding description is her answer: moral discourse may have the appearance of truth-evaluable discourse, but it's really about conveying one's attitudes or emotions toward an action. 

One challenge for expressivism concerns the use of unasserted moral statements that feature in arguments. Consider the following modus ponens argument:

1. If throwing squash at Mike's head is wrong, then getting someone else to do it is wrong. 
2. Throwing squash at Mike's head is wrong. 
3. (Therefore) getting someone else to do it is wrong. 

Even if you don't accept the premises, it is easy to see that the argument is valid. According to expressivism, premise 2 expresses a negative attitude toward throwing squash at Mike's head. And since premise 1 is a moral judgment, it must also be expressing some attitude. But notice that someone could very well affirm the conditional in premise 1 without having a negative attitude toward throwing squash at Mike's head. This means that whatever the antecedent of the conditional contributes to the whole statement, it's contribution isn't to convey an attitude of disapproval toward throwing squash at Mike. But the antecedent of the conditional in premise 1 has to convey exactly what is conveyed by premise 2, and according to expressivism, premise 2 expresses an attitude of disapproval toward throwing squash at Mike's head. The validity of modus ponens requires that premise 2 and the antecedent in premise 1 be identical. To see this, note the form of the argument:

1. If P, then Q
2. P
3. (Therefore) Q

P conveys the same thing in premise 1 as in premise 2. Yet, if expressivism is true, it would seem that P in premise 1 need not convey the same thing as is conveyed by P in premise 2. Again, this is because, possibly, someone could affirm premise 1 without having a negative attitude toward the relevant action, but someone could not affirm P without having a negative attitude toward the relevant action. But then it seems that P is conveying different things in each case. When it's embedded in a conditional, the expressivist's P can convey one thing. When it's asserted, it can convey another. But that's not how validity works. The P's have to convey the same thing. So if you think there are valid moral arguments, expressivism faces a problem.

I suppose expressivists could argue that the statement in premise 1 (and other moral statements with conditional form) isn't really a conditional. But then what appears to be an obvious instance of a logically valid argument isn't one after all. More strongly, deductive logic could not be used in moral discourse. Expressivists who aim to preserve moral discourse -- including moral reasoning -- should find this troubling.
___________________________________________
Footnotes:

See chapter 2 of Sharfer-Landua, Moral Realism: A Defense, and Van Roojen, "Rationality and Expressivism."