In Fall 2001, at Brown University, for a philosophy course taught by James Dreier, I wrote a short paper about expressivism. In metaethics, “expressivism” is a theory that maintains that ethical pronouncements express attitudes, not facts. For many people, the expressivist theory is counterintuitive because ethical discourse does appear to make truth claims.
Allan Gibbard defended expressivism in Wise Choices, Apt Feelings (Harvard University Press, 1990, especially Chapter 5, “Normative Logic”), and Dreier reformulated, generalized, and expanded Gibbard’s theory to account for why people often speak of normative judgments as true or false (“Transforming Expressivism,” Nous, 1999).
In my own analysis at the time, I thought special attention should be paid to the moment at which norms are mapped onto propositions. The content of the specific norms and propositions would determine, in my view at the time, whether the Gibbard/Dreier theory could properly be called “expressivist.”
Transforming expressivism: Worlds in which a statement is true
If I say to you, “For goodness sake! Don’t do X. That’s terrible!,” what information am I giving you? Gibbard’s answer: The content of my statement is the set of factual-normative worlds in which my statement is true.
w1: the world in which John is a priest
n1: a complete set of norms according to which it is wrong for priests to drink
In the factual-normative world <w1, n1>, it is wrong for John to drink. There are other possible factual-normative worlds, too, in which it is (or would be) wrong for John to drink. The set of these worlds, according to Gibbard, is the content of the statement “It is wrong for John to drink.”
Sodium chloride is the chemical name for table salt. You know this—if you had forgotten, I just reminded you of it—but imagine someone who does not know it or has forgotten it. They might be alarmed at the idea of putting “sodium chloride” on the table although they accept the norm of using “salt.” Since the terms represent the same substance, these normative judgments are always contradictory. Expressivism, when using “possible world” semantics, fails to account for the contradictory judgments of “don’t eat it” and “eat it.” (The expressivist, for their part, probably doesn’t believe that their ethical expressions are propositions at all and thus does not believe that their expressions can contradict each other.)
To account for the apparent contradiction, Dreier wanted to reorganize Gibbard’s sets of possible worlds (which, again, are the contents of normative statements) into subsets based on a shared norm. The contents of normative statements are “incomplete propositions” or “propositional functions” that can be true or false relative to a set of norms. The statement It is right/wrong to do X is true or false depending on what else is true or false. The set of norms maps itself to some other more detailed proposition that can be more simply true or false.
The statement “Megan ought to fight” sounds normative. According to indexical theory, as Gibbard pointed out, it really means that a certain set of norms requires Megan to fight. It’s basically descriptive, although it becomes normative insofar as it implies that you endorse the norm you describe.
There are, of course, other sets of norms that do not require Megan to fight. So indexical theory can account for moral disagreement. Conversational context often assumes or creates a shared normative system, relative to which normative propositions can be true or false. When I wrote my paper, I suggested other possibilities for explaining disagreement. The normative statement could express additional information like “we have reason to adopt or obey that set of norms” or “we act as if that set of norms were true.” This is information about which people can disagree.
Indexical theory is a type of cognitivism. However, Dreier thought indexical theory was, on the whole, similar to a noncognitivist expressivist theory (i.e. one in which ethical expressions represent noncognitive attitudes rather than propositions).
Suppose we focus on the speaker’s attitudes rather than their beliefs about external facts. Using a previous example, when an expressivist says “it is wrong to put sodium chloride on beans and right to put salt on beans,” their own ignorance of chemical labels explains why they hold this self-contradictory belief, and this factual ignorance determines their attitude.
In my paper, I pointed out that it is reasonable not to ingest chemicals labeled with unrecognized names. If you don’t recognize the term “sodium chloride,” you inhabit a factual-normative world in which it can, for you, simultaneously be wrong to put one mysterious substance on food (when you don’t recognize the label “sodium chloride”) and right to put another substance on food (when you recognize the label “salt”). This has to do with knowledge, intent, and caution. We are speaking of someone’s feelings and attitudes, not objectively describing chemicals. There is no contradiction.
I suggested another example with a more obviously ethical application: deciding to donate to charity. An individual donor has different knowledge and comfort levels about how different charities operate. One might propose that it can be right or wrong to donate to a specific charity depending on the donor’s knowledge and feelings about that charity; this is not a contradiction of the form donate and don’t donate.
Am I using an expressivist-compatible propositional theory to explain normative logic (as Dreier suggested), or am I ditching propositional theory altogether and claiming that ethical statements shouldn’t be considered as propositions at all? I left that question open in my short class paper 20 years ago. Perhaps someone today has an idea of where to take this.