The Frege–Geach problem is a major challenge for expressivism and other non-cognitivist theories in metaethics
. It questions how non-cognitivist accounts can explain the logical behavior of moral statements, especially when they are embedded in more complex sentences like conditional (“if-then”) statements.
The problem was developed in the 1960s by philosopher Peter Geach, based on the work of Gottlob Frege, and independently by John Searle.
The core of the problem
A non-cognitivist theory, such as expressivism, holds that moral statements do not express facts or propositions that can be true or false. Instead, they express non-cognitive mental states, such as emotions, attitudes, or prescriptions. For example, “Stealing is wrong” is not a statement of fact but is roughly equivalent to an emotional exclamation like “Boo, stealing!”.
The Frege–Geach problem exposes a critical weakness in this view by pointing to how moral language is used in logical arguments.
Consider this standard logical argument using modus ponens:
- Premise 1: If stealing is wrong, then getting your little brother to steal is wrong.
- Premise 2: Stealing is wrong.
- Conclusion: Therefore, getting your little brother to steal is wrong.
The non-cognitivist dilemma
The argument is logically valid, and its premises and conclusion seem consistent with how we use moral language. The non-cognitivist, however, struggles to explain this validity without either changing their theory or creating a fallacy:
- The assertion problem: In Premise 2, “Stealing is wrong” is asserted and, according to expressivism, expresses the speaker’s disapproval of stealing. But in Premise 1, which is a conditional statement, “stealing is wrong” is not asserted; the speaker is not necessarily expressing disapproval. An expressivist must explain how a moral sentence can maintain a consistent meaning and logical function in both an asserted and an unasserted context.
- The equivocation problem: A non-cognitivist might try to claim that “stealing is wrong” means different things in Premise 1 and Premise 2. But if the meaning of the term “wrong” changes, the argument becomes a fallacy of equivocation and is no longer valid. This directly contradicts the apparent logical validity of moral arguments.
- The compositional problem: An adequate theory of language must explain how the meaning of complex sentences is a function of the meaning of their parts. The Frege–Geach problem shows that expressivism lacks a consistent semantic account of moral terms that works for both simple assertive sentences and complex embedded sentences.
Responses to the problem
Philosophers have offered several solutions to the Frege–Geach problem, primarily by attempting to build a more robust non-cognitivist theory that can account for logical inference:
- Quasi-realism: Simon Blackburn argues that expressivists can earn the right to use truth-apt language, like conditionals, by developing a framework of “higher-order attitudes” that mimics realistic talk. The logical relations are then explained in terms of consistency and inconsistency between attitudes rather than truth values.
- Norm-expressivism: Allan Gibbard proposes a theory of “factual-normative worlds” where moral language expresses the acceptance of systems of norms. Logical validity, on this account, is defined by what is consistent within a system of accepted norms.
- The “being for” framework: Mark Schroeder suggests that moral terms express “being for” a certain course of action. This framework attempts to provide a compositional semantics that can account for moral terms in both asserted and unasserted contexts.
Non-Cognitivism in Ethics | Internet Encyclopedia of Philosophy