John L Jerz Website II Copyright (c) 2015

Illocutionary Acts and Sentence Meaning (Alston, 2000)

William P. Alston

"the account should allow for most of the semantic judgments of fluent speakers of the language to be at least roughly correct... We have to allow for some errors in intuitive judgments, and we have to allow for those judgments to be relatively unrefined."

"Where an utterance is out of order it is because the utterance is not in conformity with what is required for an utterance of that sentence in that kind of context by some applicable rule; and when it is in order it is in conformity with those requirements."

JLJ - Humans are typically intuitive experts at making and interpreting speech acts, of all kinds.

If Artificial Intelligence proceeds through the equivalent of a Margaret Archer-esque "internal conversation", then detailed and specific understanding of the types of questions we ask ourselves and the answers we provide - perhaps in notational form as brief as an utterance - is one starting point for advancing AI.

Even more specifically, what if we made our machine an effective "expert" at generating "maybe moves" in standard game positions, then interpreting these responses as "utterances"? We tell the machine how much it should initially "care" about them as part of token Arie de Geus-style scenarios or even Francois Jullien-style "sketches" for Ludwig Wittgenstein-like "going on" - effectively a self-generating "diagnostic test" for estimating the "adaptive capacity for mobilizing coercion".

What emerges from this process is ideally more than just a good move to play in a game - we have also strategically developed a resilient "position" - a stance from which to confront an opponent in a complex and developing game, where the future is unknown and unknowable.

Our machine - perhaps now only "artificially" intelligent - might just be able to "play" a complex game of strategy at a high level, even in complicated and unclear positions with loose or tight coupling among the game pieces, armed with only this simple human-like tool of reflexivity.

But first, we would need to reformulate the actions of the machine, much as how humans expertly make and interpret speech acts... we would need knowledge of how to do this.

At a basic level, an utterance is "performed" verbally because it is deemed to be (at least somewhat) relevant to the "predicament" that "we" are in, that the creative act of determining how to "go on" can take it (the utterance) in stride, in turning-the-crank to determine what to do next. Perhaps even "silence" is an extreme form of utterance, one that deems "ponderance to generate a high-quality later utterance" or a "yielding-of-the-floor-to-someone-else" as the best way to "go on".

p.2 The taxonomy emerges from a critical discussion of J. L. Austin's path-breaking work, How to Do Things with Words (1962)...

1. Sentential act - Uttering a sentence or some sentence surrogate.
2. Illocutionary act - Uttering a sentence (or sentence surrogate) with a certain content, the sort of act paradigmatically reported by "indirect discourse", as in 'Jones asked where the nearest newsstand is'.
3. Perlocutionary acts - Producing an effect on some audience by an utterance.

...A typical act of speech involves all three.

p.3 the most general categories of illocutionary acts...

1. Assertives: merely asserting, acknowledging, concluding, remarking, insisting.
2. Directives: ordering, requesting, suggesting, imploring.
3. Commissives: promising, contracting, betting, etc.
4. Exercitives: adjourning, appointing, nominating, pardoning, etc.
5. Expressives: thinking, congratulating, expressing contempt, relief, enthusiasm, delight, etc.

p.11 My first task in this book is to develop an account of illocutionary acts (IA's).

p.11 all twentieth-century discussion of speech acts stems from J. L. Austin's seminal work, How to Do Things with Words

p.11-12 'utterance' is to be taken to range over the production of any linguistic token, whether by speech, writing, or other means.

p.28 In parallel fashion we will widen the notion of an utterance, 'Utterance' will extend over both any production of a sentence token or of other linguistic tokens as elliptical for a sentence, and any production of a sentence surrogate... Our widened notion is still restricted to devices that are fitted by rule or convention to have the same communicative function as some sentence might.

p.54 The notion of taking responsibility for the satisfaction of a condition... 'Take responsibility' must be understood in a special way... It is... like the way in which, when I become the head of a department or agency, I take responsibility for the efficient and orderly conduct of its affairs, including the work done by my subordinates. I am responsible for all that work, not in the sense that I have done it all myself, but in the sense that I am rightly held to blame if the work is not done properly. I am the one who must "respond" to complaints about that work.

p.58 The Crucial Role of Rules

I have been speaking in terms of U's instituting a certain normative state of affairs in uttering S. In the earlier version (D5.) what U does in R'ing certain conditions is to bring it about that certain states of affairs are grounds for blame or censure, whereas previously they were not... we need to consider more generally what is required for the existence of a normative state of affairs, like blameworthiness or incorrectness or the opposites.

p.58 What is it for an action to be blameworthy or the opposite, or to be incorrect, out of order, or the reverse? Isn't it a matter of whether certain rules, regulations, norms, or principles are being conformed to or violated? ...Where an utterance is out of order it is because the utterance is not in conformity with what is required for an utterance of that sentence in that kind of context by some applicable rule; and when it is in order it is in conformity with those requirements... The claim that socially entrenched rules are required can be supported

p.63 We will, from now on, think of R'ing as essentially a matter of placing one's utterance under the jurisdiction of a certain rule, at the same time keeping in mind the other formulations as bringing out various implications of the rule subjection made explicit in D8. And since R'ing is central to illocutionary act performance, illocutionary acts are essentially what we may call rule-subjection acts, acts that essentially consist in subjecting a lower-level act to a certain rule. [JLJ - echoes of Foucault's discursive formation]

p.71 An utterance is most basically made into an illocutionary act of a certain type by virtue of a normative stance on the part of the speaker.

p.78 The D8. account of R'ing, as subjecting one's utterance to a rule governing the sentence uttered, is only a rough sketch and leaves many questions dangling. What sorts of rules are we envisaging? How are they to be formulated? In what sense do they "exist" in the community antecedently to a given speaker's bringing an utterance under one of them? What place do they have in the structure of a language?

p.83 A "conventional fact" is one that holds by virtue of certain conventions, rules, or regulations. It is because rules or conventions are in force in the relevant society, and apply in a certain way to the case at hand, that the fact in question obtains.

p.85 Austin's classification of illocutionary acts was a first sketch...

1. Verdictives. "Verdictives consist in delivering of a finding, official or unofficial, upon evidence or reasons as to value or fact, so far as these are distinguishable." (152)...
2. Exercitives. "An exercitive is the giving of a decision in favor of or against a certain course of action, or advocacy of it. It is a decision that something is to be so...." (154)...
3. Commissives. "The whole point of a commissive is to commit the speaker to a certain course of action." (156)...
4. Behabitives. "Behabitives include the notion of reaction to other people's behavior and fortunes and of attitudes and expressions of attitudes to someone else's past conduct or imminent conduct." (159)...
5. Expositives. "Expositives are used in acts of exposition involving the expounding of views, the conducting of arguments, and the clarifying of usages of references." (160)

p.86 To perform a commissive is to "commit" oneself to a certain course of action; that is, to take on an obligation to do so-and-so.

p.86 Exercitives... "It is a decision that something is to be so, as distinct from a judgment that it is so" (154). This captures the "executive" character of an exercitive, that it is the bringing into being of a certain state of affairs, rather than the "recording" of a preexisting state of affairs.

p.88 I next turn to what Austin calls "verdictives", "the delivering of a finding, official or unofficial, upon evidence or reasons as to value or fact..." ...a verdictive is "giving a finding as to something - fact, or value - which is for different reasons hard to be certain about" (150). Matters may be "hard to be certain about" for a variety of reasons - indeterminacy of concepts or lack of conclusive evidence, for example.

p.152 the account should allow for most of the semantic judgments of fluent speakers of the language to be at least roughly correct... We have to allow for some errors in intuitive judgments, and we have to allow for those judgments to be relatively unrefined.

p.154 The fact that an expression has a certain meaning is what enables it to play a distinctive role (to be used in a certain way) in communication.

p.191 what renders a sentence standardly usable to perform an illocutionary act of a certain type is that the sentence is governed by a certain rule.

p.192 A sentence's having a certain illocutionary act potential consists in its being subject to a certain illocutionary rule.

p.195 To perform a illocutionary act is to utter a sentence, subjecting one's utterance to a rule that implies certain conditions of correctness for this utterance.

p.201 1. An I-rule lays down necessary and sufficient conditions for the utterance of the sentence it governs... 2. Being governed by a certain I-rule endows a sentence with a certain meaning.

p.251 A regulative rule is one that lays down conditions under which actions of a certain sort are required, permitted, or forbidden. I-rules lay down necessary and sufficient conditions for the utterance of sentences.

p.252 Thus a rule to the effect that one must stop one's car when the light is red "regulates" behavior, namely, stopping one's car, that is not dependent on that rule for its existence. There could be such activities as driving and stopping automobiles even if there were no such rule.

p.252-253 I-rules... are clearly regulative, since they regulate sentence utterance... On the other hand, I-rules seem to qualify as constitutive rules since they "create or define new forms of behavior", namely, illocutionary acts... Thus we are faced with a dilemma. How can the same rule be both regulative and constitutive? ...the terms 'regulative' and 'constitutive' do not stand for different types of rules. They stand for different ways in which a rule is related to action concepts.

p.254 any regulative rule can also qualify as a constitutive rule... if we take advantage of the possibilities it presents for concept formation. But there are formulations called 'constitutive rules' by Searle that do not in any way qualify as regulative. These are "rules" that specify under what conditions an action, state of affairs, or whatever counts as having a certain conventional status... It is not clear why such principles should be called 'rules' at all. They simply make explicit the conditions under which some state of affairs "counts as" scoring... or possessing some other conventional status... They exist by virtue of social consent