Contribution
It turns the “Advertisement“‘s objection to Harman into a general convergence result: the private-use and public-use programmes in functional role semantics, pursued until they handle Twin Earth, the travelling speaker, borrowed reference and the division of linguistic labour, become the same two-factor theory. That makes room for narrow content inside any use theory, without arguing that narrow content is explanatory.
Key insight
Whatever extra machinery a long-arm role needs to get reference right is a theory of reference in disguise. Subtract it and what remains is the internal role, which is the same in the twins. So the choice between “one factor” and “two” is a choice of bookkeeping, and n. 31 says which bookkeeping suits which purpose.
Assumptions
- Thinking in a representational system, though not necessarily sentences in the brain: “structured states” suffice (p. 160).
- Functional role includes inferential role “plus more”, including causal relations to inputs and outputs; a reductionist version must state it non-intentionally (pp. 160–162).
- Putnam’s Twin Earth verdict that the two dialects’ ‘aluminum’ utterances differ in truth conditions is granted (p. 167).
- Public use “plus private use”: Block sees no non-behaviourist reason to include thinking out loud and exclude thinking quietly (p. 174).
Key results
- Five facts explained (pp. 162–166), as listed in the LIT. Functional role semantics supports no Davidsonian truth-maximising or Dennettian rationality-maximising charity (p. 166).
- Non-reductionist failure (pp. 168–169). Defining a term through other terms’ truth conditions cannot separate ‘aluminum’ from twin ‘aluminum’, since the defining terms share extensions; with synonyms it only moves the question to ‘munimula’.
- Two-factor option (pp. 170–171). The internal factor explains the shared behaviour (sticking a bus token in the fuse box on either planet); the second factor, for example a causal theory, explains the difference. “The functional role factor determines the nature of the referential- truth-conditional factor (without determining reference or truth conditions themselves).” Accepting it does not literally require narrow content (p. 171).
- Metaphysical picture for private use (pp. 172–173). Inner representations are “semantically active” (Haugeland); public tokens have life only in our deliberate use.
- Public use (pp. 174–178). Movie-camera public use is identical on both planets; semantic descriptions of use beg the question; long-arm roles work only with “normal context” (Harman) or “dominant causal source” (Evans, Devitt), with borrowed reference reaching back to Moses, and with experts for elm and beech. The internal part, “at least partly inferential”, is the same for both communities and “can be identified with the narrow meaning … if narrow meaning exists”; the external part “looks very like a traditional theory of reference” (p. 178).
- Appendix (pp. 179–181). Burge cases cannot be traced to different causal sources, since the counterfactual language change may never reach Fred. “Functional role in a language” must be cashed through the structure of deference; it cannot be an average, since a whole community can misuse a word (‘brisket’). Individual roles still matter: the fortnight case and ‘tharthritis’.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Functional role semantics explains five facts about concepts and thoughts | moderate; sketches | pp. 162–166, “I don’t pretend for a moment that these brief sketches settle the hash” |
| C2 | Internal inferential role cannot by itself yield Twin Earth truth conditions | strong, given Putnam’s verdict | extension argument, pp. 168–169 |
| C3 | A public-use theory adequate to truth conditions is equivalent to a two-factor theory | moderate | three reference phenomena, pp. 175–178 |
| C4 | Narrow content is left room, not established | stated by the author | p. 178, “I haven’t tried to answer these questions” |
| C5 | Burge cases require functional role in a language with deference, and individual roles also matter | moderate–weak | appendix, examples only |
Concepts
- Use theory of meaning — functional role, conceptual role or procedural semantics (p. 157).
- Inferential role vs functional role — the latter adds causal links to inputs and outputs (p. 160).
- Long-arm / short-arm roles — roles extending into the world versus stopping at the sense organs and effectors (p. 175).
- Functional role in a language — a community-level role shaped by the structure of deference, not an average of individuals (p. 180).
Connections
- “Advertisement for a Semantics for Psychology” (LIT-648): cited for the determination of the referential factor (p. 170) and for indicator and Gricean semantics (n. 14).
- Harman’s “Conceptual Role Semantics” (1982) is the long-arm theory under examination; Field, McGinn and Loar are the two-factor sources (n. 16).
- Burge’s individualism papers are the appendix’s target (n. 30).
- Drayson (LIT-112, NOTE-117) cites this paper as “Block 1987”.
Bearing on the record
- NOTE-117 (C3 and its limitations) can now point to this LIT. The paper supports Drayson’s claim that functional role semantics can let content depend on communicative use, but on Block’s own conclusion that dependence sits in the referential or community factor, while the role of the inner representation is narrow. If Toon’s worry is that inner representations are “locked away” from convention, Block’s answer is that their content is two-factored, and only one factor is locked away. That reading is mine.
- No instruction for machine-learning practice.
Limitations
- The convergence argument is by cases (traveller, names, experts); it is not shown that every reference-fixing device has a long-arm analogue that separates cleanly.
- Narrow content is defended against no objection; p. 171 lists the objections and declines to answer them.
- The appendix is a sketch, with no account of how deference fixes roles.
Open questions
- Is the “functional role in a language” of the appendix a third factor, or can it be absorbed into the referential factor, as the “Advertisement” (p. 630) left open?
- Campbell’s reply (Proceedings of the Aristotelian Society 88, 1987/8) is not in the record.