Reviews of the other volumes were previously posted on LinguistList:
Vol 1&3 October 2001
Vol 4 September 2001
Vol 5 November 2001
Vol 6 September 2001
Kasher, Asa, ed. (1998) Pragmatics: Critical Concepts, Volume II: Speech act
theory and particular speech acts. Routledge, vi+490pp, hardback ISBN
0-415-16938-0, Routledge Critical Concepts series.
Reviewed by Bert Bultinck, Department of Linguistics, University of
Antwerp/National Fund for Scientific Research-Flanders.
I will not repeat the comments that the reviewers of the other volumes of
Kasher's six volume set have made. I will simply confirm that I too find it
hard to believe that this volume does not have a list of references that
indicates the source of each contribution. Michiel Leezenberg, who reviewed
the first volume (Linguistlist 12.2673, 10/25/01), noted that the first
volume does contain a list of references (which is, however, not complete).
Below I will summarize the texts that were selected for this volume on
speech acts. Occasionally, some critical remarks will be made in the
separate discussion of each text. A short general evaluation will be given
at the end of this review.
The volume on speech acts kicks off with part of J.L. Austin's
ground-breaking How to do things with words. This text can be said to have
founded scientific research into the phenomenon of "speech acts". It is
somewhat surprising that only 21 pages are devoted to Austin's essential
lectures. Of course, this sort of phenomenon is typical of anthologies and
reference works: due to limitations on length, difficult choices have to be
made. Nevertheless, I think that in this case it would have made sense to
reprint a much larger portion of the book. Because it contains many complete
articles alongside mere fragments of books and longer articles, Kasher's
"Pragmatics"-set falls in between a full-fledged anthology and an
"introduction" to critical concepts. Its subtitle suggests that the set is
meant as the latter, but the volumes lack the contextualizations that would
make it a genuine introduction to the most important principles and
concepts in pragmatics. Users of the present set might be tempted to draw
conclusions about, e.g., How to do things with words, on the basis of an
incomplete text. Contextualizations of each text might have prevented such
potential misreadings. They would have also contributed to much-needed
theory-formation in pragmatics in general.
The selection from How to do things with words starts with lecture VIII,
the first sentence of which begins as follows: "In embarking on a programme
of finding a list of explicit performative verbs, it seemed that we were
going to find it not always easy to distinguish performative utterances from
constative [...]". A start in medias res, certainly, but unfortunately so
abrupt that readers looking for more information on the critical concept
"speech acts" will not understand what Austin is talking about. Readers who
do, have probably read Austin's book before and will not use this reprinted
fragment at all - they will refer to the original publication. Moreover, the
editor sometimes seems to realize that not everything is clear: when Austin
talks about the "descriptive fallacy" he inserts a note in the text in
square brackets referring readers to the first lecture of How to do things
with words. But this first lecture is not to be found in this volume and
Kasher's rather abstract reformulation ("overlooking the possibility that
words embedded in apparently descriptive statements do not serve to indicate
features in the reality reported") will not be clear enough to readers who
have not read the entire book.
In the lectures selected, Austin explains his distinction between
locutionary, illocutionary and perlocutionary act. He devotes attention to
the fact that there must be "uptake" in order for an illocutionary act to be
successful (the audience must understand the intention behind the
illocutionary act and respond to it adequately). Crucially, he tries to
define performatives by doing paraphrasis-tests (e.g., by means of "In
saying p, I was V-ing", in which V stands for a performative). The selection
abruptly ends with the unanswered question "But what then is the relation
between performatives and these illocutionary acts?".
It is quite astounding to see that classic examples of performatives (such
as "I baptize this ship Elisabeth II") are only mentioned in passing, or not
at all. Also, while the selected fragment does address the question of how
to determine what performatives are (or explicit performatives, or speech
act verbs), the result of Austin's meditations is not included. Neither is
the crucial "I hereby S you to"-test. The reader has to wait until page 493
to see where Austin is heading.
The next selection is called "Mood and language game" and is written by Erik
Stenius. It is clearly meant as an illustration of the argument that
Austin's ideas were not as original as is sometimes believed: Wittgenstein's
contrast between the "sentence-radical" and the "modal element" is clearly a
precursor of Austin's "locutionary act" and "illocutionary act". Stenius'
central question is: how is the modal element to be characterized
semantically? In order to make this clear, Stenius uses Wittgenstein's
concept of a language-game. He suggests that the "mood" (Austin would say:
the "force") of an act of conversation can be compared to an extra mark; a
code that accompanies the other marks that simply show a state of affairs,
"the descriptive content of the sentence".
John Searle's famous classification of conditions for speech acts is
reproduced in this section. Starting from a case study (the speech act
"promise"), Searle investigates what conditions must be fulfilled so that
the speech act can be called successful and non-defective (or, in Austin's
terms: "felicitous"). Searle divides these conditions into four classes:
propositional content conditions, preparatory conditions, sincerity
conditions and essential conditions. He gives examples for a number of
speech acts. Also his famous "expressibility principle" is explained and
Kent Bach and Robert M. Harnish's remarks on the Speech Act Schema are next:
the selected fragment concentrates on their discussion of the so-called
"mutual contextual beliefs". These beliefs are considered to be central in
bridging the gap between the meaning of the sentence uttered and what the
speaker is saying. Also the "linguistic presumption" and the "communicative
presumption" are explained and defended as elementary principles.
Donald Davidson's contribution on moods and performances is in fact an
attempt to rescue his truth-conditional, Tarskian account of meaning from
the attacks of speech act theory. Davidson starts from Dummett's discussion
of Frege's assertion sign or judgment-stroke: Davidson is bothered by the
close connection that Dummett perceives between the indicative mood and
assertion. The relationship between the grammatical form of a sentence and
the force of an utterance cannot be a conventional one: every joker,
storyteller and actor exploits the grammatical indicative mood but does not
necessarily assert anything. Davidson's own account, however, resembles
Frege's proposal rather closely: he postulates the existence of a
"mood-setter" alongside the classical truth-conditional locution. The order
Put on your hat then means: 'My next utterance is imperative. You will put
on your hat'. Davidson contends that both parts of this meaning have
truth-conditional content, "but the combined utterance is not the utterance
of a conjunction, and so does not have a truth value" (79).
A short fragment on the so-called Performative Hypothesis is next. Jerrold
Sadock discusses the origins of the hypothesis, which he eventually locates
in Katz and Postal's (1964) book. He also mentions the counterarguments
against the hypothesis, but tries to rescue it from at least one of the
attacks: the ambiguity of the illocutionary force of one utterance can be
reduced by reference to the vagueness of lexical items or by Gricean
The next chapter consists of part of Leech's Principles of Pragmatics (1983)
and concentrates on a set of postulates that Leech uses to distinguish the
semantic from the pragmatic, trying to steer a middle course in between the
"grammaticists" (cf. the Performative Hypothesis) and the "pragmaticists"
(cf. Searle's statement that a theory of language is a theory of action).
Leech also discusses the rule-governed nature of semantics and the
principle-controlled nature of pragmatics, the role of "convention" in both
domains, the relation between sense and force and the application of Grice's
maxims to Searle's account of (in)direct speech acts. The text closes with
an elaborate description of pragmatics as a special instance of
Anna Wierzbicka's text on the semantics of illocutionary forces tries to
answer the question whether these forces are indeterminate, as claimed by
many authors. Her thesis is that "the supposed indeterminacy of
illocutionary forces is largely an artefact of inadequate syntactic and
semantic analyses" (116). Even though specific types of utterances cannot be
rephrased in terms of single speech act verbs, Wierzbicka proposes that
their precise meaning can be determined by means of "bundles of components".
Instead of drawing up continua of illocutionary acts (e.g., from "most
prototypical imperative" to "most prototypical interrogative"), each
construction should get a "discrete" description. She emphasizes that not
everything can be calculated on the basis of general, rational principles of
communication. The 'suggestion' sense inherent in specific Why don't you
constructions has to be taught to the learner of English. She won't be able
to figure it out by herself. It is the task of the linguist to specify the
general meaning of these kinds of constructions. The rest of the reprinted
chapter offers such analyses (constructions as "conversational strategies",
tag-questions, constructions signifying personal abuse etc.). It ends with a
short cross-linguistic section.
Daniel Vanderveken's comments on the logical form of illocutionary acts are
next: in this section, he explains how illocutionary force can be divided
into six components and how this set can be the basis of a recursive
definition all sorts of speech acts. The six components are: an
illocutionary point (Searle's classification of speech acts), a mode of
achievement of an illocutionary point (e.g., when requesting something, the
speaker should leave the hearer the option of refusal), propositional
content (in fact, some illocutionary forces impose conditions on the set of
propositions that can function as propositional contents of a certain
illocutionary act), preparatory and sincerity conditions (cf. the well-known
Searlean conditions are explained and exemplified), and degree of strength
(supplication is stronger than request). Five primitive illocutionary forces
are defined on the basis of these components.
Michael Dummett's notes on sense, force and tone are heavily inspired by
Dummett's reading of Frege. He takes up the quarrel with Davidson, and
addresses the difficult status of "convention" in an account of
illocutionary force. Dummett tries to demonstrate the importance of
distinguishing the linguistic act from the intention that is behind the act.
He claims that the matter is very complex, but that at the very least sense
has to be distinguished from force. At the end, Frege's concept of "tone" is
explained, by means of the difference between and and but.
Geach's text on assertion kicks off Part Four, which focuses on specific
speech acts. A rather long introduction tries to disentangle the concepts
"sentence", "proposition" and "statement". A proposition is a form of words
in which something is propounded, put forward for consideration. A statement
is an asserted proposition. The confusion between the terms "proposition"
and "statement" is a source for many misunderstandings, e.g., Ryle's
critique on the "code style" of the modus ponens, Oxford-based amendments to
the logic of truth-functional connectives (or, and) and negation. Geach
defends Frege's decision to adopt an "assertion sign", because that sign
enables philosophers to distinguish propositions from assertions. The fact
that a proposition may be asserted or unasserted does not change anything
about the proposition: it always remains the same. This insight is called
"the Frege point" by Geach and is presented as the solution to a large
number of misunderstandings. The final part of the paper criticizes various
attempts to reduce the "assertion sign" to a concept (e.g. "existence" or
"it is true that") and discusses Austin's tendency (in a not overly generous
reading of Austin's texts) to avoid speaking of propositions altogether. If
people refuse to recognize "propositions" as relevant concepts, then all
syllogistic reasoning breaks down.
Michael Dummet's paper on assertion first establishes that a theory of sense
and reference alone cannot explain communication as a human action. Human
intentions must be integrated into the theory in order to show what the
point is of communication. Equally crucial are a set of conventions:
"Assertions take place against the background of a custom of uttering them
with the intention of saying something true" (228). Dummett looks for the
codification of speech acts in English, and suggests, e.g., that we may draw
up an 'obedience-table' for disjunctive commands, just like we can draw up
truth-tables for 'or'.
Robert Stalnaker introduces a number of concepts that are meant to capture
the various interpretations a proposition can have in various contexts. He
proposes a two-dimensional matrix as truth-table for any expression in a
given set of contexts: the vertical axis represents possible worlds in their
role as context. These determine what is said. The horizontal axis
represents possible worlds in their role as the arguments of the functions
which are the propositions expressed. The most important aspect of this
text, however, lies in Stalnaker's evaluation of (the effects of) assertion
in terms of a reduction of the so-called "context set". This perspective not
only allows him to define communication in terms of information (and
information in terms of possible worlds), but it also allows him to
formulate three very general principles guiding linguistic interaction,
focusing on the roles of content and context. These principles reformulate
Gricean insights, interpret the phenomenon of truth value gaps, and rephrase
the adage that what a statement says must be independent of any facts that
might be relevant to determining its truth.
Dietmar Zaefferer's text opens the second section of part four ("other
speech acts: general"). It tries to account for illocutionary force
indicators, in this case in German. The article mainly consists of a formal
reconstruction (in the Montague framework) of data on two uses of
interrogatives, namely the erotetic and the assertive use. There are
correspondences between these interrogatives and specific assertive
constructions (in specific readings). It is these correpsondences that
Zaefferer wants to point out. The article closes with a few more general
remarks on the semantics/pragmatics divide.
Deirdre Wilson and Dan Sperber tackle the relationship between mood and
non-declarative sentences. They concentrate on the distinction between mood
and force in the traditional analyses of imperatives and interrogatives.
They conclude that the traditional descriptions are not empirically
adequate. Their proposal consists of a two-layered schema: the syntactic
pattern (e.g., "interrogative") must be assigned some conventional intrinsic
semantic content, which can then be used as a foundation for an explanatory
account of force. This conventional content cannot be truth-conditional - it
is semantically indeterminate and has to be enriched pragmatically. It
encodes a rather abstract property of the intended interpretation: a
suggestion as to the direction in which the relevance of the utterance is to
Nuel Belnap discusses some examples of what he calls the "declarative
fallacy": e.g., Frege's famous "context principle" only takes into account
declarative sentences and is therefore, necessarily, incomplete. But Belnap
goes further and argues that any mentioning of "propositional content" as
the basis of each and every speech act is fallacious: it is a mistake to
suppose that the content of all speech acts can be identified with the
content of declarative speech acts. Also, truth-conditions, verification
conditions, and inferences are not enough for a compositional theory of
meaning. Belnap provides some suggestions as to how we could interpret
interrogatives and imperatives, in order to make the compositional theory of
meaning less incomplete.
The third part of part four concerns questions and starts with a text by
Ranier Lang on questions as epistemic requests. The bulk of the paper
consists of an overview of various attempts to define questions. The
semantic bases of questions are crucially linked to the desire of the one
posing the question to get to know something. Also, it is especially on the
discourse level (e.g. challenges to questions and answers) that we can hope
to find suggestions as to the general nature of questions.
Ruth Manor's article on the logic of questions and assertions takes a
pragmatic stance towards the problems at hand. She starts with a discussion
of the proper place of logic in the general analysis, drawing heavily on
Collingwood's theory of logic. In this framework, an assertion is always an
answer to a question; there is a logical priority of questions to assertions
in the sense that any assertion consists of answering a question - whether
it was asked or not. She applies Collingwood's insights to the theory of
speech acts and argues for a representation of both asking a question and
asserting a proposition in terms of question-answer pairs. She also presents
a way to formalize dialogue in these terms.
The fifth section of part four deals with commands and deontic speech. David
Holdcroft's contribution follows up on his analysis of the indicative as a
semantic restriction on its range of uses. This text concentrates on the
question whether it is possible to formulate similar restrictions as part of
the meaning of the imperative mood. He argues that imperatives can neither
be analyzed performatively, nor as expressions of intentions, wishes etc.
but must relate to states of affairs. He proposes that imperatives can be
associated with "conformity conditions", just like indicatives can be
associated with truth conditions.
Hans Kamp's text focuses on expressions of permission and obligation. It
identifies the relationships between the following permissive sentences:
(1) "you may take an apple", (2) "you may take a pear" and (3) "you may take
an apple or take a pear". The latter is stronger "in as much as the set of
worlds which a performative utterance of (3) adds to the options of the
addressee includes, but is not necessarily included in, the set added
through a performative use of (1)" (399). In addition, the relationship
between the utterances mentioned and their directive counterparts is
analyzed, and the truth-conditions for the permission sentences are
established. Finally, also the consequences for the semantics-pragmatics
distinction are reviewed.
James Forrester's contribution first asks why we should use deontic speech
at all. He answers that there may be many motivations, but one is considered
to be central: "to cause people to act or to refrain from acting in certain
ways". This is what he calls the "directive use" of deontic statements.
Forrester discusses differences between his analysis and R.M. Hare's theory
and formulates eighteen maxims of "deontic pragmatics". Deontic speech is
divided into the "legislative mode of deontic speech" (for which Forrester
formulates general rules for obligation and permission), and the "judicial
mode of deontic speech" (use of deontic language in particular cases).
The section on promises begins with a text on the relationship between
promises and assertions by Katharine Bath. Her account of promises also
entails a critique of Lewis' assertion that every utterance is governed by a
convention of truthfulness. The convention that gives the meaning of
promises is two-fold: the promises covered by law and those that are not.
She draws up a list of twelve conditions. She also discusses John Rawls'
analysis of promises. Her discussion of breaches of promises and their
consequences for the meaning of promises is explicitly framed in legalistic
terms, and also her account of assertions is inspired by jurisdictional
Margaret Gilbert tries to answer the question whether an agreement is an
exchange of promises or not. She refers to the law as one of the sources of
inspiration for her version of a specific part of speech act theory. She
defines three criteria of adequacy for a model of agreements, looks at
different kinds of promise (e.g., unconditional promise, internally
conditional promise, and externally conditional promise) and analyzes a few
case studies. She concludes that the agreements she has looked at cannot be
seen as exchanges of promises.
The sixth and last section of part four deals with performatives and returns
to Austin's classic text, How to do things with words. Included are the
passages in which Austin deconstructs the distinction between constatives
and performatives he had defended in his first lectures. The distinction
breaks down because to make a statement is also to "do" something. But there
remains a difference: with constatives the locutionary is more important
than with performatives and the correspondence with the facts is essential.
J.O. Urmson's comments on Austin's theory of performatives start from the
assumption that Austin's theory of performatives in How to do things with
words is less satisfactory than his earlier attempts. Urmson elaborates on
this earlier version, demonstrates its merits and concludes that its basic
insight is that performatives rely on non-linguistic conventions for their
A short fragment by Fran�ois Recanati proposes to replace the definition of
performative utterances as serving to perform illocutionary acts by one in
which they aim to bring about (and not simply describe) a state of affairs.
It also discusses the relationship between performatives in general and
Searle's text on how performatives work first of all defines performatives:
these are illocutionary acts that can be performed by uttering a sentence
containing an expression that names the type of speech act. It then
addresses the question how it is possible that there is a class of sentences
so that we can perform the action named by the verb "just by saying
literally we are performing it" (522), but that is only one among a series
of other problems and puzzles. A large part of the text consists of a
discussion of earlier proposals to answer these questions. An attempt is
made to derive the declarational character of performatives from their
assertive character, but that does not work. The crucial step is to realize
that the self-guaranteeing character of performatives derives from the fact
that these utterances are self-referential to a verb which contains the
notion of an intention as part of its meaning. These acts can be performed
by manifesting the intention to perform them.
Kent Bach and Robert M. Harnish reply to Searle's account of performatives
argues that Searle confuses performativity with communicative success. They
defend their model against Searle's attacks and discuss Searle's objection
to statement analyses. They also argue that ordinary performatives are not
declarations and claim that, in fact, performatives are not very special at
all: "they are but one example of standardized forms of words used to
perform speech acts indirectly" (554).
In the general, theoretical part on speech acts (called Part Three), Austin
and Searle receive relatively little attention, while Wierzbicka's text is
allotted too much space: her text is very valuable, but her point is quite
simple ("do not squeeze particular constructions into general frameworks,
but examine them in their own right") and the many examples that illustrate
her point add nothing to her theoretical stance (again, this entails nothing
with respect to the value of the concrete analyses in themselves; it is just
that you would not expect to find these analyses in an introduction to
"critical concepts"). Together with the selection of Stenius' text (which is
also, in itself, a very valuable contribution), this is one of the puzzling
decisions of the editor. More of Austin's classic text should have been
reprinted; and the two fragments that are reprinted should have been grouped
together (now they are more than 450 pages apart).
Part four groups together analyses of specific speech acts. It is somewhat
surprising to see more than 340 pages devoted to rather concrete, often
long-winded analyses of specific speech acts, especially when the other
volumes were found to lack a few crucial texts (cf. the other reviews on
Linguistlist). Also, the unity of this fourth part is only apparent: only
three of the six subsections really correspond to what you would expect to
find under the rubric "particular speech acts", namely the sections on
questions, commands and deontic speech and promise. The three others are
either much more general (the section aptly called "other speech acts:
general" does not deal with general surveys or classifications of other
speech acts, but concentrates on the question of mood and force), much more
philosophical in style and content (the part on assertions) or even
explicitly critical, and with good reason, of the separate status that a
certain type of speech act has (the section on performatives). The class of
performatives is not on the same level of generality as the class of
commands, promises or even questions.
The set does not offer a discussion of the criteria that were used to select
texts. Equally unfortunate is the lack of contextualizations. Maybe a new
edition could provide short introductions to each text, locating the
fragment in the history of pragmatics, demonstrating its relevance, and,
crucially, indicating its place in the work of the author. When the fragment
is taken from a book, it is absolutely essential that there is some
indication of the main theses of the whole book and how these relate to the
specific claims made in the fragment.
On the whole, however, this volume offers a rich overview of the many (kinds
of) discussions that have been triggered by Austin's trail-blazing work. The
specific selection of texts can of course always be criticized, and there
will be infinitely many gaps for anyone who expects completeness.
Completeness, however, should not be expected from this introduction to
"critical concepts" in pragmatics. Some of Kasher's choices are clearly
political, in the sense that particular paradigms, approaches, and authors
receive much more space than one would expect. While this may be seen as a
shortcoming, it must also be admitted that this is unavoidable. And a more
generous evaluation would consider it as an opportunity to discover new
texts, approaches and hidden agendas in the domain of speech act theory.
Katz, J.J. and P.M. Postal. (1964). An integrated theory of linguistic
descriptions. Cambridge: MIT Press.
Leech, G. (1983). Principles of pragmatics. London: Longman.
ABOUT THE REVIEWER
Bert Bultinck is a research assistant (Fund for Scientific Research -
Flanders) at the University of Antwerp and works on Grice's legacy, the
semantics/pragmatics interface and the meaning analysis of numerals.