Ned Block (1986), Midwest Studies in Philosophy 10 (Studies in the Philosophy of Mind), pp. 615–678 — DOI-10.1111/j.1475-4975.1987.tb00558.x.
The brief’s citation (Midwest Studies 10, 1986) is correct. The running head of the scan reads “MIDWEST STUDIES IN PHILOSOPHY, X (1986)”; the “1987” in the DOI is Wiley’s.
Key takeaways
- Two factors (pp. 627–628). Meaning has “a conceptual role component that is entirely ‘in the head’ (this is narrow meaning) and an external component that has to do with the relations between the representations in the head … and the referents and/or truth conditions of these representations in the world”. Conceptual role is “total causal role, abstractly described”, abstracted to the relations that mediate inference, decision and the like, as herring have a culinary role inside their total causal role.
- Narrow meaning is motivated by explanation (pp. 619–625). Two people thinking “I am in danger” leap aside for the same reason; Teen and her Twin Earth duplicate share a thought about “Michael Jackson”; my twin and I both jump into the pool when our pants are on fire. What they share explains behaviour and fixes the function from context to referent. An analysis of variance over a bus of Earth and Twin Earth tourists would otherwise have to credit variance to a factor with no physical effect on them (p. 625).
- Conceptual role fixes the kind of reference, not the referent (pp. 643–644). Which theory of reference is true (causal, not descriptive) is a fact about how names function in our thinking, so conceptual role “determines the nature of the referential factor”; the same role maps to H₂O here and XYZ on Twin Earth.
- Harman’s long-arm roles collapse into two factors (pp. 635–637). To handle the visitor to Twin Earth, the elm/beech case and borrowed reference, long-arm roles must import “normal context”, experts’ roles and causal chains into the past. “Chop off” the part outside the skin and what is left is the short-arm role; what is chopped off is a theory of reference.
- The competitors (pp. 649–660). Non-reductionist semantics (possible worlds, situations, Davidson, Katz) do not aim to say what makes an expression meaningful. Searle’s “equivalent causal powers” is not naturalistic. Dretske–Stampe indicator semantics falls to Fodor’s disjunction problem, and Fodor’s own theory fails on both its “wheels”: epistemic idealisation cannot be specified naturalistically for theoretical sentences, and evolutionary teleology does not guarantee mechanisms aimed at truth. Without compositional structure, it cannot tell “the electrical conductivity is rising” from “the thermal conductivity is rising” (the Wiedemann–Franz law, p. 660).
- Functionalism for intentionality, not experience (p. 635). Block explains why he advocates a functionalism after “Troubles with Functionalism”: he will settle for a theory of human meaning, and his liberalism arguments “were strongest against functionalist theories of experiential mental states”. Pain is “a composite state consisting of a nonfunctional qualitative state together with a functional state”.
Standing in the record
Filed on 2026-10-03 at the owner’s request, with Block’s other papers on
functionalism, meaning, AI and perception. A theory of meaning and mental
content as philosophers pose it is philosophy-of-language; no anthology
topic holds it, and it carries no instruction for machine-learning
practice.
It is the record’s first full statement of functional (conceptual) role semantics, which the record had cited only at second hand.
- NOTE-117 (Drayson, LIT-112) names “Block 1987” as a counter-model to Toon’s mind-first metasemantics: functional role semantics on which mental representations take content partly from social and communicative use. Drayson’s reference list gives that as “Functional Role and Truth Conditions” (1987), held here as LIT-625, not this paper. This paper is where the theory Drayson leans on is stated in full, and its p. 630 leaves the social dimension of meaning to the referential factor or to “a distinct social factor”. NOTE-117’s limitation, that Block 1987 is “named, not developed”, can now be checked against both.
- NOTE-117’s RTM core. NOTE-117 takes RTM’s core to be that mental representations are individuated apart from attitudes, with structure inessential. Block’s representationalism is minimal in a similar way: he is committed only to representational states that “constitute a combinatorial system”, contingently and empirically (pp. 641–642). If CRS met serious empirical problems because of that commitment, he “would pursue a nonrepresentationalist version”, and he names Loar’s and Schiffer’s functionalist Gricean programme as one such avenue (p. 642).
- Searle (LIT-400, NOTE-349). CRS gives the systems reply a theory: the Chinese symbols would have autonomous meaning if they had the right conceptual roles. Block separates simulating a Chinese speaker’s processing from emulating it, and says only emulation is understanding (pp. 665–666).
- Teleosemantics (LIT-149, NOTE-165). Martínez’s teleofunctional account of content is in the family whose teleological “wheel” Block argues cannot be naturalised by appeal to evolution, because selection does not guarantee mechanisms aimed at truth (meiotic drive, pp. 659–660). That pairing is mine.
- Interpretationism (LIT-206, NOTE-138). NOTE-138’s C7 contrasts reward-based representationalism with interpretationism; Block’s p. 645 rejects truth-maximising charity and derives coherence-based charity from conceptual roles, which is a third position in that dispute.
- Neighbours held only as seeds: Prinz’s two wide contents (LIT-134) and Brandom’s inferentialism (LIT-106).
No relation is declared. The paper argues against Searle, Fodor, Dretske and Harman, none of whose primary statements of these views is in the record except Searle’s, and against Searle it diagnoses an ambiguity rather than refuting the argument.