Contribution
A co-author’s restatement of the expressed-attitude theory of illocutionary acts (EA-TIA) of Bach and Harnish (1979), with a critical survey of the alternative that commitment, not expressed attitude, makes an utterance the act it is. It finds Searle’s and Alston’s normative accounts unexplained or false at the point that matters. It proposes that the two kinds of theory may pick out the same facts. The hard question, what undertaking a commitment amounts to, is left “to another occasion” (abstract).
Key insight
An intention-based theory and a commitment-based theory of illocution may not be rivals at all. If expressing the belief that P is the same as taking responsibility for P being the case, both theories say the same thing in different vocabularies (pp. 34–35, 38). A choice between them would then have to rest on something other than the illocutionary facts.
Assumptions
- The utterance of a sentence in a context is not sufficient for a speech act; theories differ by what they add (abstract).
- Belief-desire psychology as the frame for explaining communication: the Gricean programme hopes to put “communicative uses of language within the reach of traditional psychological explanations” (p. 16).
- “Internal” and “external” apply to force, not content. Harnish accepts “certain forms of externalism with respect to content” (p. 15, n. 9).
- Expressing in the non-achievement sense. One need not have the attitude one expresses: one expresses “the belief that P”, not necessarily “one’s belief that P” (pp. 17, 34).
Key results
- The Austinian paradigm’s shortfalls (pp. 12–14). No analysis of convention, no principled taxonomy, no link to successful communication, no explanation of why speakers say what they say or how hearers understand, and no tie to the structure of a language.
- EA-TIA (pp. 15–16). The Gricean answer goes “internal”. “Suppose a central class of illocutionary acts is communicative in nature, and suppose that the performance of illocutionary acts in this class can be defined in terms of the propositional attitude expressed by the speaker, and the understanding of such communicative acts can be defined as the recognition of such propositional attitudes by the hearer.” Recognition “is guided by a shared system of inference strategies, presumptions, and contextual information” (p. 15). Then “kinds of (communicative) illocutionary acts can be defined in terms of the kinds of attitude(s) expressed”, and “successful illocutionary communication can be characterized as the recognition of the expressed propositional attitude” (p. 15).
- Expressing (pp. 16–22). Bach and Harnish: “For S to express an attitude is for S to R-intend the hearer to take S’s utterance as reason to think S has that attitude” (1979, p. 15, quoted p. 19), where to R-intend is to intend that the intention be recognized. Kemmerling’s attempt at a stronger, “cp-analytical” notion fails, because lying, sarcasm and microphone tests are normal, and so “collapses into something ‘empirical’ like Bach and Harnish” (pp. 20–22).
- Objections to EA-TIA (pp. 22–23). Among them, that it is not a single category: Austin, Searle and Alston give homogeneous theories, whereas “Strawson (1964) suggested a spectrum of conventionality, and Bach and Harnish (1979) sharply distinguished the two categories”, the communicative and the institutional (pp. 22–23). Another objection is that illocution involves commitment, which expressed attitudes do not capture (p. 23).
- Normative theories (pp. 23–24). Alston: the utterance is made the act it is “not by any ‘natural’ facts about the speaker … but by a ‘normative’ fact about the speaker – the fact that he has changed his normative position in a certain way by laying himself open to the possibility of censure, correction, or the like in case the conditions in question are not satisfied” (Alston 2000, pp. 70–71, quoted p. 24).
- Searle on commitment (pp. 25–28). Searle (2001) says speech acts create desire-independent reasons by “imposing conditions of satisfaction on conditions of satisfaction”. Harnish finds the phrase “quasi-mysterious”, and “exactly how does imposing (cosocos) create commitment? … we are just told that it does” (p. 27). Appealing to constitutive rules regresses, since it leaves open how speaking by rules creates commitment (pp. 27–28).
- Alston on rule subjection (pp. 28–33). Alston analyses taking responsibility (D5–D8) as subjecting one’s utterance to a rule “in force within a community”, with rules existing when deviations “evoke incorrectness judgments” (pp. 28–30). Harnish’s counterexamples are conversational implicature, where no illocutionary rule licenses “I will be in London” only if one will miss the party, and parents’ responsibility for their children, where community judgements “would be just symptoms of having undertaken a responsibility” (pp. 31–32). He concludes: “A speaker does not take responsibility for a certain set of conditions because his sentential act is subject to an illocutionary rule requiring it. Rather, S performs a certain illocutionary act because he takes responsibility for those conditions being the case” (p. 32).
- Expressing and taking responsibility linked (pp. 33–38). Harnish proposes “(E : C) In uttering e, S is expressing the belief that P : In uttering e, S takes responsibility for its being the case that P” (p. 35). He meets the objection that commitment comes in degrees and expressing does not (pp. 37–38). Conclusion: normative notions “do, however, seem connected to notions of being liable to criticism and blame. However, we suggested, Gricean notions of expressing an intentional state are likewise liable. So normative theory might at this level be simply another way of describing the same illocutionary facts” (p. 38).
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | On EA-TIA, communicative illocutionary acts are individuated by the attitude the speaker expresses, and understood by its recognition | definitional, reported by a co-author | pp. 15, 19 |
| C2 | Bach and Harnish separate communicative from institutional acts | reported | pp. 22–23 |
| C3 | Searle does not explain how a speech act creates commitment | moderate | analysis of Searle 2001 (pp. 25–28) |
| C4 | Alston’s analysis of taking responsibility as rule subjection is false | moderate | two counterexample types (pp. 31–32); compositionality worry (pp. 30–31) |
| C5 | Expressing the belief that P and taking responsibility for P may be equivalent | weak; a proposal | the biconditional (E : C), one objection answered (pp. 35–38) |
| C6 | Normative theories have no advantage over Gricean ones at this level | weak; conditional on C5 | p. 38, flagged as needing more work |
Concepts
- EA-TIA: the expressed-attitude theory of illocutionary acts, Bach and Harnish’s Gricean theory (p. 15).
- N-TIA: any normative theory of illocutionary acts, using commitment, responsibility, blameworthiness (p. 23).
- Expressing an attitude (Bach and Harnish): R-intending the hearer to take the utterance as reason to think the speaker has it. It is non-achievement: the speaker need not have the attitude (pp. 17, 19).
- R-intend: intend reflexively, that is, intend that the intention be recognized (p. 18).
- R’ing that P (Alston): taking responsibility for its being the case that P; in D5, knowingly taking on liability to blame if not-P (p. 28).
- Internal and external, of force: whether what makes an utterance an illocution lies in the speaker’s states or outside them (p. 15, n. 9).
Connections
It summarizes, by one of its authors, Bach and Harnish’s Linguistic Communication and Speech Acts (LIT-1040, unread). Its foils are Alston’s Illocutionary Acts and Sentence Meaning (2000) and Searle’s Rationality in Action (2001), neither held. On the record’s side, Strawson 1964 (LIT-1043, read in NOTE-736) is the source of the “spectrum of conventionality” Harnish credits. Searle’s status functions (LIT-482, read) are the institutional half CLAIM-221 grants. Berdini (LIT-1877, read alongside) cites this article as facing normativity with a solution other than hers.
Bearing on the record
- CLAIM-221: partly. It confirms the rival’s shape and adds a twist, but says nothing on teasing. The claim describes the intention-based line as individuating communicative acts by speaker-internal states, and conceding institutional acts to convention. A co-author’s text bears both out. EA-TIA goes “internal” (p. 15). Acts are defined by the attitude expressed (p. 15). Bach and Harnish “sharply distinguished” communicative from institutional acts (pp. 22–23). The 2026-10-13 correction, that uptake is a condition and not only evidence, holds here too: communication succeeds by recognition (p. 15). But what individuates the act is the attitude the speaker R-intends to be recognized, and recognition only measures success. The new point is Harnish’s argument that the normative alternative may be equivalent (p. 38). Alston’s normative theory makes the act a matter of the speaker’s liability “to censure, correction, or the like”, a relation to others (p. 24). If Harnish is right, that relation and the expressed attitude pick out the same acts, and the dispute CLAIM-221 describes may be partly verbal. The article does not mention teasing, reproach or standing. What CLAIM-221 waits on for teasing, Bach and Harnish’s own treatment, is not met.
- QUESTION-073: partly. The question asks whether teasing and reproach are individuated by communicative intention or partly by a jointly held frame and standing. On the evidence here, EA-TIA answers “intention” for every communicative act. It handles context as “contextual information” and mutual presumptions that guide the hearer’s inference (p. 15), not as part of the act. That is the intention-based answer stated by its author, with no case of a fixed intention and a changing act. Its suggestion that liability to censure (a relation) may coincide with the expressed attitude bears on the 2026-10-13 note’s open point: whether standing that enters through the intention counts as a relation. The article does not settle it.
- No instruction for machine-learning practice.
Limitations
- Not the book. EA-TIA is summarized in two pages (pp. 15–16). The book’s inference schema, its presumptions and its taxonomy are not presented.
- Assertion-centred. The positive argument (E : C) is made for belief and assertion only, though EA-TIA covers all communicative acts (p. 33).
- Left open by the author. “The really hard work is determining what undertaking a commitment or responsibility amounts to. That we must leave to another occasion” (abstract).
Open questions
- Does (E : C) extend from assertion to directives and to acts like teasing, whose expressed attitude is not a belief?
- If liability to censure and the expressed attitude coincide, are they equivalent for an audience that does not share the speaker’s standing? That is the case QUESTION-073 asks for.