See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/261180528

Locution, illocution, perlocution. In: Pragmatics of Speech Actions. Ed. by M.

Sbisà and K. Turner. 25-75. Mouton de Gruyter. 2013. 9783110214383

Chapter · January 2013

CITATIONS READS

2 11,147

1 author:

Marina Sbisà
University of Trieste
64 PUBLICATIONS   816 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

assertion among the speech acts View project

Dynamics of illocution View project

All content following this page was uploaded by Marina Sbisà on 02 May 2020.

The user has requested enhancement of the downloaded file.

Locution, illocution, perlocution
Marina Sbisà

published in: M. Sbisà & K. Turner (eds), Pragmatics of Speech Actions, Handbook of
Pragmatics 2, Berlin: Mouton de Gruyter, 2013, pp. 25 75

1. Introduction

When we speak, we articulate sounds with our vocal organs, and we do so in such a way
that they can be taken to belong to some natural language, conform to its rules, and
express a certain meaning. We usually do something else as well. Our speech has more
or less precise goals, achieves or fails to achieve them, may express intentions or other
mental states, may produce consequences of various kinds (sometimes unintended),
and so on. It may be said that we “use” language to communicate, for strategic
purposes, to express emotional or other psychological states, to persuade, and even to
carry out such peculiar activities as joking or play acting. This heterogeneous set of
things we do when we speak or in the performance of which speech plays a major role
has been analysed by philosophers and linguists in the tradition of speech act theory,
and primarily by the British philosopher John L. Austin, who proposed the three fold
distinction between the locutionary, illocutionary and perlocutionary acts, which is the
topic of this chapter.
I will firstly present and analyse Austin’s distinctions. Then I will focus on the
reformulations that each of the three notions involved, that is, locution, illocution, and
perlocution, have undergone in the subsequent development of speech act theory. In
doing so, I will stick to the work of some major authors or trends of thought, aiming to
show how differences in the philosophical background assumptions create differences in
the descriptions of what we do when we speak and in the corresponding speech act
theoretic notions. In conclusion, I will propose some reflections on philosophical
problems concerning the various ways of conceiving of what we do when we speak that
speech act theory makes available and provide an evaluation of the role that these
notions and conceptions play (or should be able to play) in the analysis of discourse and
conversation.

2. John L. Austin on locution, illocution and perlocution

Locutionary act, illocutionary act and perlocutionary act are the names given by John L.
Austin to three aspects of what he called “the total speech act in the total speech
situation” ([1962] 1975: 52,148). Austin thinks that any feature of a speech act and of
the situation in which it occurs may be relevant to its meaning and to the assessment of
the speech act’s correctness, which, according to him, can never be reduced to the
logician’s assessment of truth and falsity (cf. e.g. Austin 1975: 52).
So, in a way, the speech act as a whole is a single, complex phenomenon or even, as he
writes, “the only actual phenomenon which, in the last resort, we are engaged in
elucidating” (Austin 1975: 148). But elucidating such a phenomenon involves attaining
some level of abstraction. As Austin says in announcing his distinction between locution,
illocution and perlocution, doing something is a vague expression, and there is a need to
reconsider “the senses in which to say something may be to do something, or in saying
something we do something (and also perhaps to consider the different case in which by
saying something we do something) (1975: 91–92). This reconsideration leads to
identifying a number of differing abstracted acts, which co exist in standard cases within
the same total speech act, but may fail independently of one another in non standard
cases, thus engendering “different kinds of nonsense” (1975: 147), and can therefore
(we may conclude) be appreciated and assessed independently of one another.
At this point the contemporary reader may observe that acts, as Austin conceives of
them, must be queer entities. The remark makes sense, especially from a perspective in
which acts are viewed as reducible to physical movements (let alone the relationship of
these with the neural events making them possible), and any accepted entity obeys
materialistic constraints. This was not Austin’s perspective. In some of his philosophical
writings, indeed, Austin reveals himself to be an ontological pluralist, a philosopher who
places no limits on the number of ontological kinds or realms to which things (and even
one and the same “thing”) may belong. In his theory of perception (Austin 1962), he
explicitly refuses the dichotomy between “material things” and inner, psychological
entities. While both materialists and idealists are monists, Descartes was a dualist, and
Frege admitted of three distinct ontological realms (Frege [1918] 1956), Austin intends
to reject the philosophical practice of counting worlds as if their number, allegedly small,
were of some import to philosophy (Austin [1949] 1970). In the context of his
philosophy, therefore, applying materialistic constraints to the notion of an act (or for
that matter, an action) would lead to misleading interpretations. Leaving metaphysical
questions aside, let us only recall that, for Austin, no ontological claims were admissible,
apart from those stemming from the observation of the way in which we ordinarily talk
(cf. Austin 1962). And in a way, through the examination of the senses in which doing
may be ordinarily used in connection with saying and the consideration of the different
kinds of flaws and dimensions of assessment to which the speech act is liable, Austin
appears to believe that also the “abstracted” acts such as the locutionary, illocutionary,
and perlocutionary act are legitimate objects of our attention.
2.1. Locution
The locutionary act (Austin 1975: 92) can be identified as the act of saying something1,
but since saying something may have different senses, its analysis has to proceed
further, leading to distinguishing the phonetic act, the phatic act and the rhetic act. The
phonetic act is the uttering of sounds and is performed whenever we speak (not in the
same way, though, when we use language in writing), but is not itself speech. One
might utter vocal sounds that are not speech, as children do when they are about ten
months old. The product of phonetic acts, I would like to add, is continuous as opposed
to discrete: an analogical recording can be made of it, and it is studied by phoneticians
in all its shades of pitch, volume, sound quality, etc. The phatic act is again the uttering
of sounds, but “conforming to and as conforming to” a language (1975: 92). Its product,
that is, are not continuous sounds, but discrete tokens of phonemes, morphemes, and
other linguistic structures. Performing phatic acts is already speaking a language, but in a
broad sense, including for instance practising a language one does not fully understand.
Writing, of course, must comprise a level of linguistic activity equivalent to the phatic
act, to be carried out by means other than vocal utterance. The rhetic act is the uttering
of words (or production of written words, etc.) endowed with meaning, which may be
“sense”, “reference” or both (1975: 93).
Austin leaves outside the scope of his analysis the problem of how meanings (senses
and/or references) attach to words. He also neglects to define sense, reference, and
meaning, as if the uses of these words which the analytic philosophy of language
inherited from Frege were self explanatory. But the Fregean view that every linguistic
expression has a sense, which is an entity belonging to the realm of thought and
connecting the linguistic expression to what it refers to, is not shared nowadays by all
philosophers of language and was not so shared in Austin’s times. Austin himself does
not accept it completely, since he tends to assign “sense” to predicates and “reference”
to singular terms (cf. 1975: 97, [1953] 1979c). This lack of care in defining the rhetic act
is undoubtedly a flaw of Austin’s analysis.
It is perhaps useful to recall that Austin proposed to distinguish phatic acts from rhetic
acts as the acts that are reported in inverted commas (oratio recta) as opposed to those
that are reported in “indirect speech” introduced by that or to (oratio obliqua). Here are
some examples:
(1) It is getting late
(1a) Phatic report: “I said to him ‘It is getting late’”
(1b) Rhetic report: “I said to him that it was getting late”
(2) Go away!
(2a) Phatic report: “I said to him: ‘Go away!’”
(2b) Rhetic report: “I told him to go away”
One can easily see from these examples, as well as from Austin’s own (1975: 95), that
reporting a speech act qua phatic act involves preserving the precise words spoken, but
not necessarily understanding nor clarifying what it means, while reporting a speech act
qua rhetic act requires the use of words that may need to differ from those actually
spoken in order to express or even clarify what the utterance means. Insofar as the
phatic act and the rhetic act are identified by means of these different speech reporting
operations, it is difficult to represent them as parts of the locutionary act that can be
merely added one to the other to make up the whole, but there is no indication that this
was ever Austin’s intention. Austin never defines the complete locutionary act (apart
from the initial consideration that it is the act of saying something); he exemplifies his
conception of the locutionary act by means of an example which is again a report:
(3) Shoot her!
(3a) Locutionary report: “He said to me ‘Shoot her’ meaning by ‘shoot’ shoot and
referring by ‘her’ to her”.
But it is clear that not all that would belong to a rhetic report is inherited by the alleged
locutionary report and, moreover, such a locutionary report sounds quite unnatural
from the ordinary language point of view (which should have made Austin himself
suspicious of it).
A further puzzle about the rhetic act is noticed by Austin himself, but only in passing.
Rhetic reports must take into account sentence type (declarative, imperative,
interrogative, or other types if the natural language in question has them). In our
example (2), the rhetic report of “Go away!” uses the reporting formula I told him to as
opposed to I said him that. So, as Austin observes (1975: 95, 97), the rhetic report of a
question will have to use I asked him whether or I asked him what (who, which, how…).
A first, broad orientation as to the kind of illocutionary force of the speech act cannot be
separated from the rhetic act, nor therefore from the locutionary.
One might argue from this to the conclusion that Austin’s distinctions are incorrectly
drawn and that the rhetic level of the locutionary act is already illocutionary. It is in
conformity with this line of reasoning that most speech act philosophers replaced the
Austinian notions of the rhetic and the locutionary act with that of expressing a
proposition. Since propositions are truth bearers and expressing a proposition is
equivalent to assigning truth conditions to an utterance, the whole speech act appears
to comprise a lower, truth conditional level and an upper, illocutionary one. Linguistic
markers of the illocutionary act such as sentence type and mood, albeit part of what is
uttered, cannot belong to the lower, truth conditional level because they are not part of
the expression of truth conditional content. Thus, isolating truth conditions involves
taking sentence type and mood as pertaining entirely to illocution. An opposite line of
reasoning would take speech reports in terms of saying that, telling to, asking whether
as rhetic and therefore locutionary reports (see Hornsby 1988). According to this view,
there is at least one sense of alleged illocutionary verbs of the most general kind, such as
saying, telling, and asking, in which these verbs do not design broad illocutionary act
types, but kinds of rhetic act.2 Indeed, Austin himself admitted the ambiguity of saying
as sometimes merely locutionary (phatic or rhetic) and sometimes equivalent to stating
(cf. 1975: 167, 168). Perhaps, in order to make sense of Austin’s rhetic act, we should
distinguish two senses also for verbs such as telling and asking, or, at least, two poles
between which their uses oscillate.

2.2. Illocution
The illocutionary act is introduced by Austin as the kind of act that we generally eo ipso
perform in performing a locutionary act (1975: 98). Illocutionary acts are also taken by
him to be ways in which language is “used”, kinds of “use of language”, at least in one of
the senses of this expression. Albeit being part of the movement known as ordinary
language philosophy, for which the meaning of a linguistic expression was to be
analysed by considering its use, Austin was critical of any easy identification of
locutionary meaning with use. But he also refrained from identifying all “uses of
language” with illocutionary acts. There is according to him no uniform notion of use of
language, rather, there are different senses in which we can speak of uses of language:
one connected with locutionary meaning, another with illocution, a third with the
achievement of extra linguistic goals, and a fourth linked to so called “non seriousness”
or aetiolation. For this very reason, Austin chooses to replace the umbrella expression
use of language with a more detailed terminology.

2.2.1. Illocution and aetiolation

Austin’s insistence that the uses of language for joking or quoting or play acting are not
illocutionary acts (and therefore, are not uses of language in the same sense in which an
illocutionary act may be said to be a way of using language, see 1975:104) impressed
the French post phenomenological philosopher Jacques Derrida, who took inspiration
from it to mount a case against the viability of the very notion of a speech act (qua
illocutionary act in particular). According to Derrida (1972), the availability of linguistic
expressions (even performative ones, connected, in Austin’s view, to the performance of
illocutionary acts: cf. Doerge, this volume, and this article, Subsection 2.2.2) for use
without serious intention, amounting to what he also calls “iterability”, is a revealing
feature of language and Austin should not sweep it under the carpet by saying that non
serious uses of languages are not his main concern and giving them negative
characterizations such as “nonserious” or marginal and non standard ones such as those
implied by calling them “parasitic” or “aetiolated”.3 In Derrida’s perspective, there is
nothing serious and central to stick to. To evaluate whether Derrida’s objection hits the
mark, readers of Austin should consider that Austin’s real aim in setting aside non
serious uses was not to defend genuine speech actions as originating from the serious
intentions of the speaker (as Derrida seems to believe), but to criticize and surpass the
Wittgensteinian rhetoric of the “infinite uses of language”, which was widespread and
influential in the ordinary language philosophy environment. In his Philosophical
Investigations, Wittgenstein exemplifies uses of language (in his terminology, “language
games”) by means of a deliberately heterogeneous list, ranging from commanding or
describing or thanking to story telling, play acting and translating (1953: § 23). What
Austin wants to do is to show that, if attention to uses of language is to help us
understand language and speech, we should not stay content with such heterogeneous
lists, but should distinguish the precise sense in which each alleged use of language is so
called. Indeed, Wittgenstein’s list mentions or involves locutionary acts (or sub acts
belonging to the locutionary level), illocutionary acts, perlocutionary acts, various
coordinated conversational or social activities, and non serious speech activities.
Locution, illocution and perlocution, as well as aetiolation (“non serious” speech acts),
are for Austin not mutually exclusive kinds of use of language, but categories of analysis
that apply to any speech act, often one independently of the other. So, for example,
kinds of aetiolation comprise joking, quoting, play acting and writing fiction or poetry.
Kinds of illocutionary acts comprise assertion, question, promise, command and
apology. It may make sense to ask of an utterance functioning as a certain type of
illocutionary act, say, a promise, whether it is really meant or subject to one kind or
another of aetiolation. It may make sense to ask of an aetiolated utterance (e.g. one
spoken in play acting) whether it functions as one type or other of illocutionary act (even
if the illocutionary act is to be attributed to the character in the play in the name of
whom the utterance is spoken, and is addressed to another character, as opposed to a
real person).

2.2.2. Illocutionary acts, performatives, and force

Coming back to the illocutionary act, the act that we generally eo ipso perform in
performing a locutionary act (where illocutionary is derived from in+locutionary), Austin
emphasizes its being embodied (as it were) in the locutionary act and explains how a
number of linguistic devices may be used as indicators of its performance, from mood
and sentence type to specialized lexical items, from modal verbs to intonation or
punctuation (1975: 73–77). The speaker’s gestures and the circumstances of the
utterance may also be of help. The point of these devices is to indicate how the
utterance is to be taken, or to make clear its force. Austin borrows the term force from
Frege, who distinguished the thought expressed by a sentence from the “assertoric
force” with which the sentence is used by a speaker when she pronounces her judgment
that the thought expressed is true (Frege 1956). How ever, while Frege limits his
consideration of force to matters concerning the assessment of an expressed thought as
true or false (he contrasts assertions with hypotheses, in which this assessment is
suspended, and with questions, in which it is deferred to the interlocutor), Austin frees
force from the relationship with truth and falsity and generalizes the use of the term
force to all uses or functions of language belonging to the illocutionary level. The
illocutionary force of an utterance is thus, in Austin’s perspective, equivalent to the
illocutionary act that an utterance of that type by default performs so long as the
performance is not undermined by fatal flaws. It is in a broad sense a variety of meaning
(he notices that we can use the word meaning with reference to illocutionary force: “He
meant it as an order”, 1975: 100), but should be contrasted with meaning “in the sense
in which meaning is equivalent to sense and reference”, that is, in the sense in which
meaning pertains to the locutionary act alone.
Performative formulas such as I promise you that … or Passengers are warned to … (cf.
Doerge, this volume) are the most explicit devices for performing illocutionary acts,
increasing at the same time the specificity of the performed illocution (Austin 1975: 69–
73), while illocutionary acts performed by means of the use of a certain sentence type or
of a certain modal verb, and the like, are often negotiable (to borrow a concept from
conversation analysis; but cf. Austin 1975: 66, 72). Indeed, an imperative accompanied
by please may be a command, but also an entreaty or suggestion, and the way it will
actually work in the conversation in the course of which it is uttered depends also on the
way it will be responded to (let alone the way in which that response will be received by
the speaker). But the utterance of a performative formula such as I order you to … either
is successful in giving an order, or fails to do so because of some inappropriateness
or infelicity,
e.g. if the speaker has no authority over the addressee as regards the content of the
purported order (Austin 1975: 14–18, cf. Doerge, this volume). It should be noted that if
illocutionary acts can be explicitly performed by means of performative utterances, the
converse must also hold, namely, it must be the case that the acts performed by means
of performative utterances are illocutionary acts.4

2.2.3. Illocutionary acts and their effects

The acts we perform in saying something, or illocutionary acts, have a number of effects
that Austin deals with very briefly (thereby leaving ample room for subsequent
interpretations and debates). These effects are:
(a) a reception of the speech act such that the speaker can be said to have “secured
uptake”, that is, made the meaning and force of the utterance available to the
audience;
(b) a conventional effect brought about by the illocutionary act insofar as uptake has
been secured;
(c) an effect consisting of inviting a certain response (when the inviting of that
response is conventionally attached to the performance of that kind of illocutionary act)
(Austin 1975: 116–118).
As to (a), Austin claims that the securing of uptake is required in order for the
illocutionary act to have been actually performed. But he is not thoroughly clear about
whether what is required includes actual uptake or just the speaker’s reasonable effort
to produce it. He speaks of “securing uptake”, which can be understood as the speaker’s
effort to make herself understood, but, as an example, he proposes the case of a
warning, which is not really carried out if, for example, the audience does not hear the
speaker’s utterance (if I shout my warning aloud, but you are already too far away to
hear me, I cannot say I did warn you, I merely tried to warn you). One position that may
plausibly be attributed to Austin is that uptake is secured when the speaker manages to
make it possible for the audience to understand. This indeed is already an achievement,
as implied by Austin in the case of warning, and therefore an effect brought about by
the speaker in issuing the utterance. But it does not follow from this that the audience
actually pays attention nor that any actual interpretation, even when it is indeed
misinterpretation, should count as uptake and contribute to validating the
corresponding illocutionary act as well as its attribution to the speaker.5
As to (b), what Austin literally says is only that “the illocutionary act ‘takes effect’ in
certain ways, as distinguished from producing consequences in the sense of bringing
about states of affairs in the ‘normal’ way, i.e. changes in the natural course of events”
(1975: 117). So, the effect that the illocutionary act “takes” (provided that uptake is
secured) is not a change in the natural course of events, which we may assume to be
brought about by means of causal connections. One may be tempted to say, anticipating
Grice’s terminology (but not his concept of non natural meaning, for which see
Kemmerling, this volume), that Austin introduces this second effect of the illocutionary
act as in some sense “non natural”. To explain this, we may resort to the contrast
between nature and culture, between facts (as objects of the natural sciences) and
values or norms, or between what there is (the ontological dimension) and what there
should be (the deontic dimension). On the one side of the divide we find causal
connections of events, on the other side we find human elaborations and choices, tacit
agreements, conventions, and rule following behaviour. We may also think of the later
distinction between brute and institutional facts (Searle 1969: 50–53, 1995). Indeed,
Austin never addresses the analysis of institutions, but brings various kinds of
institutional and ritual activities together with other more informal social ones under the
notion of “conventional procedure” he employs in his explanation of the conditions in
which an utterance may be performative (1975: 26–28, cf. Doerge, this volume). It
should be recalled in this connection that such so called conventional procedures, while
enabling performative utterances to be operative, also provide the framework in which
illocutionary acts can be performed, a framework which essentially envisages
“conventional effects”. Thus, the word “conventional” can be used, in an Austinian
framework, to characterize the effect which illocutionary acts are said to “take”. In Sbisà
(1984, 2001, 2006), I have claimed that such an effect can be represented as the
creation, cancellation or change of deontic states of affairs concerning the participants
in the ongoing interaction (that is, states consisting of the conjunction or disjunction of
an agent with a modal predicate belonging to the deontic kind). In Sbisà (2007), I
have stressed that changes at the deontic level have a feature central to “conventional”
acts in Austin’s sense, which is defeasibility. The creation of an obligation (e.g. as the
outcome of an order) is annulled if the order is discovered to be infelicitous and thus a
misfire: you did not do what you were told to do, perhaps, but this was no disobedience,
because, since the order was a misfire, you were under no obligation to do what the
speaker’s utterance purported to impose on you. We will come back to conventionality
and its role in illocutionary acts below (Subsections 4.1.1, 4.3).
As to (c), the inviting of a response is proposed by Austin as an optional feature of
illocutionary acts. Some illocutionary acts invite by convention a response of a certain
kind: e.g. a request, qua request, invites compliance. But other illocution ary acts, while
apt to provoke certain responses on certain occasions and in certain contexts, do not
invite one specific kind of response in this same way. The classic example is assertion.
But one might also contrast requests (which would not be requests if they did not invite
compliance) with advice. A piece of advice only needs to be recognized as such in order
to be such (and model the relationship between speaker and addressee accordingly). It
need not invite a specific response, but can leave it up to the addressee in what way
and to what extent to follow it.

2.2.4. The problem of the classification of illocutionary acts.

In the last chapter of his How to Do Things with Words, Austin presented a tentative
classification of illocutionary acts, which has since been criticized by almost everyone in
speech act theory. But what is he really doing? Let us take a closer look at his line of
reasoning.
If we want to distinguish classes of illocutionary acts, we first need a list of illocutionary
acts which will then be distributed among classes. In this context, a list of “illocutionary
acts” can only mean a list of illocutionary act types: not so much a list of individual
occurrences of illocutionary acts (individual performances of acts done “in” saying
something), as a list of the different kinds of things that we may do “in” saying
something. Austin is in no doubt that the best approximation to such a list can be
obtained by identifying the performative verbs available in a natural language (such as
English, in his case) as those verbs that, at the first person present indicative active, can
be performative (while their use at the second and third person present indicative
active, or in the past tense, yields mere reports). Thus we find out the names of the
illocutionary act types that are available to the speakers of that language. Austin’s
assumption draws upon his own linguistic phenomenological approach to philosophy,
particularly upon the idea that natural languages encode useful distinctions and,
perhaps, all the distinctions that ordinary social interactions in a given culture require.
So he uses the lexicon of illocutionary verbs to delimit the field of illocution and identify
the illocutionary act types to be subjected to classification.
But how can one classify act types? Austin’s way is to draw upon resemblances among
the action schemes or scripts that specify the meaning of the verbs on his list. The
salient features he considers include the properties that the speaker is to possess in
order for her to be the agent of an act of a given type (and name) and the kind of
conventional effects that acts of that type (and name) are designed to achieve. So,
Austin sketches out five “classes”, which are indeed fuzzy sets each built around a small
number of illocutionary act types by which, as he says, the class is “typified” (1975:
151). He summarizes his distinctions by characterizing verdictives as acts of exercising
judgment, exercitives as acts of exercising power or influence, commissives as acts of
assuming an obligation, behabitives as acts of adopting an attitude, and expositives as
acts of clarifying reasons, arguments, and communications (1975: 163; for the
definitions and their discussion see also 1975: 151–162, and Kissine, this volume, Table
1).
Quite openly, and nearly programmatically, Austin’s “classification” is not intended to
divide illocutionary act types exhaustively into a series of non overlapping classes. He
accompanies the characterization of each class with comparisons with the other four,
thereby pointing out overlaps and ambiguities and highlighting hybrid cases, that is,
cases of illocutionary act types that possess aspects belonging to prototypical members
of one class together with aspects belonging to prototypical members of another
(permitting, for instance, is the exercise of a power, but also commits the speaker to a
certain line of conduct, and commending is the taking up of an attitude as well as an
exercise of authority: Austin 1975: 156–157). He even considers the possibility that
belonging to one or the other class might be a matter of degree (1975: 164).
So, Austin’s classification of illocutionary acts helps us see what he is concerned with
when he speaks of illocution: complex procedures envisaging circumstances of
execution, competence, position or capacity of the participants, conventional effects,
and linguistic forms suitable to making it clear what procedure is invoked. These
procedures are constitutive of illocutionary act types, and their variety in a given society
and culture is reflected in the availability, in the language of that society and culture, of
names for illocutionary act types, that is, verbs allowing for performative use.
Illocutionary act types may be prototypical or marginal with respect to the class to which
they belong. Nothing prevents occurrences of illocutionary acts from being hybrid or
marginal realizations of illocutionary act types.

2.3. Perlocution
The perlocutionary act (where perlocutionary is derived from the Latin per ‘through’, ‘by
means of’+locutionary) is introduced by Austin as the kind of act that we may perform
by saying something, that is, by performing a locutionary act and therein an illocutionary
act (1975: 101, 108). Examples of perlocution are convincing someone that things are
so, persuading someone to do something, alertingsomeone about some impending
danger, reassuring someone about not being left alone. Other examples may be
surprising someone, or misleading someone, by one’s speech act.
Perlocution occurs only when some consequential effect is produced in some receiver of
the speech act because of some feature of the speech act itself, so that its speaker can
be taken to be responsible for that consequential effect.6 Suppose a speaker reproaches
her addressee harshly and then he kills himself. Did she make him commit suicide?
Perhaps so, if her reproach played a major role in triggering his suicidal behaviour.
Perhaps not, if a major role was played by some other reason or by an inadequately
treated pathological state of depression. Contrary to the effects of the illocutionary act,
illustrated above in Subsection 2.2.3, the consequential effect whose production is
constitutive of the perlocutionary act must not be a conventional effect, that is, it must
not affect a purely deontic state, such as a state of right, obligation, entitlement (as
Austin says in his 1975: 102–103, the speaker’s commitment which is an effect of
promising is a matter of illocution and not of perlocution). In general, one can say that
all Austinian examples of perlocution involve either psychological attitudes or actual
behaviour of some of the participants on the scene of the speech act. By the way, these
effects are typically not defeasible. The fact that something non defeasible, that is, not
dependent on conventions must occur goes hand in hand with the fact that there can be
no explicit performative formula for perlocutionary acts. In saying “I persuade you to
buy my old car”, whatever the situational context, I am not eo ipso persuading you to
buy my old car: my funny utterance might serve to announce what I am trying to do and
what I believe will happen, but whether I do persuade you depends on whether you are
actually persuaded, which requires your adoption of a certain psychological attitude.
Compare this example with “I propose you buy my old car”. Unless there are infelicities
that make the illocutionary act misfire (suppose the car is not mine), this utterance is
itself a proposal and commits the speaker to selling her car to the addressee if he
accepts to buy it, inviting the addressee to provide a response of acceptance or refusal.
Suppose that by saying “I propose you buy my old car” I persuade you to do so and thus
perform the perlocutionary act of persuading you to buy my old car. Some participant or
bystander might report our exchange this way (“She persuaded him to buy her old car”),
but it will be a report, not a first person, explicit performative utterance.
While an actual, non defeasible consequential effect of the speech act is necessary for
perlocution, the speaker’s intention to produce that effect is not indispensable (at least
according to Austin 1975: 106). A speaker may attempt to achieve a certain
perlocutionary effect without succeeding (and in this case she has not performed the
corresponding perlocutionary act), or may not intend to produce a certain effect, which
nevertheless occurs, so that, insofar as her speech act has played a role in triggering the
effect, she has performed the corresponding perlocutionary act.
A distinction that might be of help while exploring the heterogeneous field of
perlocution is the one between the achievement of perlocutionary objects and the
production of perlocutionary sequels (Austin 1975: 118). This distinction aims at isolating
those cases of perlocution in which the perlocutionary effect is closely linked to the
illocutionary act performed, perhaps by definition (the illocutionary act would not
belong to the type it does if it were not designed to aim at that perlocutionary effect). In
these cases, it is highly unlikely that the production of the perlocutionary effect be
unintended. If I say “Do not do it”, I (provided I have the authority to do so, etc.) am
forbidding you to do something, and my illocutionary act is closely linked to the aim of
deterring you from behaving that way. That is, my act of forbidding has deterring as its
perlocutionary object. In certain cases, there is coincidence between the response
invited by convention by the illocutionary act (see above, 2.2.3 (c)) and the
perlocutionary object that the speech act is aimed to achieve. This is certainly the case
with commands and requests. In these cases, one might want to say, against Austin, that
there is a conventional element in perlocution after all. But Austin attributes this
conventional feature to the illocutionary act, saying, as we have seen, that it is an effect
of certain illocutionary acts that they invite by convention a certain response. So, while
the inviting of the response is an effect of the illocutionary act, the response itself is the
perlocutionary object of the speech act, and the actual production of the response is the
achievement of that perlocutionary object, constitutive of the speaker’s perlocutionary
act. In this framework, it becomes clear that the illocutionary act is the means by which
the perlocutionary act is accomplished.
The achievement of perlocutionary objects is contrasted with the production of
perlocutionary sequels. This label covers a very heterogeneous field: all those cases in
which some aspect of a speech act, locutionary or illocutionary, produces an effect of
the non conventional kind, whether intended or unintended, which lacks any regular
association with the illocutionary force of the speech act that triggers it. Here are some
tentative examples of perlocution belonging to this kind:

(a) I know you like to do what I forbid you from doing. So I say, “Do not do that”. You
immediately do it. Getting you to do that is not the perlocutionary object of my act of
forbidding (rather, it would be the perlocutionary object of a command, request,
suggestion or recommendation, but I did not perform any one of these). But it is the
effect my act of forbidding has on you in these circumstances, and is actually an
intended effect, given what I know about your psychological inclinations. So, it is an
intended perlocutionary sequel of my act of forbidding.
(b) You are extremely afraid of dogs. On entering the garden of your new friend’s house,
you see a notice reading: “Dangerous dog”. It is a warning, having the perlocutionary
object of alerting its readers about the presence of our friend’s dog and keeping them
from approaching the animal. But you are not merely alerted, you feel alarmed and
start panicking. The warning produced in you an unintended perlocutionary sequel.
(c) You know that Ron has quarreled with his friend Jon, and does not want even to hear
his name. You are drinking a coffee with Ron and talking to him of whatever comes to
mind. At a certain point Ron tells you that Jon has a new job. Of course, you are
surprised. This effect is not triggered by the illocutionary force of Ron’s utterance, so it is
beyond question that it is not the achievement of a perlocutionary object. What triggers
the effect is, rather, the content of the assertion and, more precisely, the fact that it
explicitly refers to Jon. So, your surprise is a perlocutionary sequel of Ron’s speech act,
triggered specifically by Ron’s making reference to him by mentioning his name.
(d) Tom is a very kind, shy person who never imposes on others. One day he takes part in
a treasure hunt with some friends. Faced with the various puzzles they have to solve and
tasks they have to perform, he proves to be extremely competent and starts playing a
leader like role in the group. When he first issues a command, his friends are surprised.
Commands bear no regular connection with surprise, so this effect cannot be the
achievement of a perlocutionary object. Maybe Tom’s command achieves also its
perlocutionary object of making his partners do something: but nothing stops a speech
act from having more than one perlocutionary effect. Though triggered by the
illocutionary force of the speech act, the surprise effect is a perlocutionary sequel.
The notion of perlocutionary sequel may help understand why the border between
perlocutionary acts and acts of getting people to do things merely by non verbal means
is fuzzy. Surprising someone or frightening him are effects that can be produced by
speech acts as their perlocutionary sequels, but of course, analogous psychological
reactions can be aroused by non verbal events and, in this case, there is no reason to call
such a production “perlocution”. Stopping someone from doing something may be the
achievement of the perlocutionary object of a prohibition and also (on occasion) the
perlocutionary sequel of any other speech act, but the agent may also be prevented
from doing that thing if hit on his head or tied up with rope and, in these cases, it is
obviously no perlocutionary effect. In contrast with this, consider obedience. It is
possible to obey and to get someone to obey only if some order (or similar illocutionary
act) is addressed to an agent by a certain speaker. Obedience, indeed, is the
perlocutionary object of orders. Additional, possibly non verbal means may be used, but
if the addressee’s performance is to be called “obedience”, there must be some order
with which he has to comply.

3. Meaning and the locutionary act

As I pointed out above (Subsection 2.1), Austin did not say enough about the locutionary
act or, as we may put it, about saying considered as itself a doing. What he said
appeared to most commentators to be controversial and even mistaken, particularly
because of gaps and overlaps between the phatic act and the rhetic act, the rhetic act
and the complete locutionary act, and even the locutionary act and the illocutionary act.
Only a few speech act theorists kept using his term locutionary act and not without
adjustments. Other philosophers adopted new terminology, introducing for example the
propositional act or the propositional content of the speech act, thus marking a
conceptual change with respect to the articulation of the speech act initially put forward
by Austin, which did not include propositions. Still others have dealt with meaning in the
context of various views considering speech as an activity. This has mostly been done by
way of a contribution to the conceptual frameworks of semantics or pragmatics and
without reference to the distinction of locution from illocution and perlocution. Insofar
as these proposals or theories approximate the general project of grounding meaning in
some kind of human action, we shall briefly mention some of them.

3.1. Propositional acts and contents

As said above, the proposition, an element typical of analyses of language in the
tradition of analytic philosophy is missing from Austin’s account of the act of saying. In
presenting the rhetic aspect of the locutionary act, Austin deals with sense and
reference as potentially disjoint from one another, thus implying that the unity of the
proposition (requiring both reference and predication) is not realized at the rhetic level
and therefore not at the locutionary level either. But many speech act philosophers held
that there are reasons for reserving a role for propositions in the analysis of the speech
act and that their expression should be collocated at a stage prior to that at which
illocution comes in. Indeed, the main philosophical intuition about propositions is that
they are constant across different, but synonymous (and inter translatable) ways of
expressing them. So “Snow is white” expresses the same proposition as “La neve è
bianca”, and “Sono stanca” (said by the author of this paper) expresses the same
proposition as “Marina is tired”. This intuition can be extended to the cases in which
what varies is not language or choice of words, but the illocutionary act the utterance is
designed to perform. An assertion, a command, a question, a wish, such as:
(4)
(a) Sam smokes.
(b) Sam, smoke!
(c) Does Sam smoke?
(d) May Sam smoke!

may in this perspective be analysed as applications of different illocutionary operators to

one and the same proposition or “propositional content”(cf. Searle 1969: 22–23).
John R. Searle (among the first – with William Alston and Peter F. Strawson – to pick up
Austinian themes about speech and action in the 1960s) modified the analysis of the
locutionary level of the speech act so as to include the act of expressing a proposition.
He replaced Austin’s phonetic and phatic acts with the utterance act, amounting to the
uttering of words (morphemes, sentences) and the rhetic act with the propositional act,
comprising the acts of referring and predicating (Searle 1969: 24). According to him,
utterances (4) (a d) can be analysed as performing the propositional act of expressing
one and the same proposition (a proposition that refers to Sam and predicates “smoke”
of him) and applying to it the illocutionary force of assertion, command, question and
wish respectively.
So in Searle’s theory, utterance acts are the vehicle of propositional and illocutionary
acts, insofar as their performance follows the rules for the performance of such acts. In
the same way in which making an x on a ballot paper counts as voting, an utterance act
can count as a propositional act and as an illocutionary act too: while utterance acts
consist of uttering strings of words, illocutionary and propositional acts consist of
uttering words in sentences in certain contexts, under certain conditions and with
certain intentions (Searle 1969: 24–25).
Searle also stresses that propositional acts cannot occur alone: one cannot perform a
propositional act without performing some illocutionary act at the same time (1969: 25).
This establishes some kind of a primacy of the illocutionary act over the propositional
act: the utterance of a complete sentence can never only be the expression of a
proposition and thus the mere saying of something. Every saying is also a doing, that is,
the performance of an illocutionary act, and indeed the complete sentence (duly
analysed) comprises, besides an indicator of the proposition expressed, also an indicator
of the type of illocutionary act it is designed to perform. There is no saying completely
free from being also, or giving rise to, a doing. This idea of Searle’s appears to support
Austin’s rejection of the dichotomy between performatives and statements (or
“constatives”; see Doerge, this volume) and even to give it full accomplishment, but it
does so at a price. Austin was wary of using the notion of expression, which (so it
appears) he felt to be unclear and imbued with mentalism (in the introspectionist
sense), but here we have an indispensable component of the speech act defined in its
terms. Moreover, and perhaps more importantly, while according to Austin that which is
said to be true or false is the accomplished speech act (provided it is an assertion or
statement, or some similar verdictive expositive speech act, see Austin [1950] 1979b,
1975: 142–145), according to Searle assessing an assertion according to truth and falsity
amounts to assessing its propositional content. So the introduction of the notion of
proposition into the analysis of the speech act provides the theorist with a truth bearer,
an entity to which truth or falsity pertain, or which (adopting the jargon of
semantics) can be said to possess a truth value. In his subsequent work (particularly
Searle 1979a, Searle and Vanderveken 1985), Searle sticks to the distinction between
illocutionary force and propositional content, which has in fact been generally accepted
in the mainstream tradition of speech act theory.
The notion of proposition, or of the propositional content of a force, is thus in important
respects equivalent to the truth conditional idea of “what is said”, according to which
what one “says” is determined by the truth conditions of the uttered sentence. But what
is said cannot depend only upon what sentence has been uttered. In the case of
indexical sentences, such as Sono stanca ‘I am tired’, context gives a decisive
contribution to the determination of the truth conditional content of the utterance, as
has been recognized already by Frege (1956). As John Searle recognized soon (1979a:
117–136), this contribution of context goes far beyond the interpretation of indexical
elements. In the following decades, other philosophers and linguists (see e.g. Travis
1997, 2000; Carston 2002; Recanati 2004, 2010) gave their support to contextualist
claims and analysed the ways in which the propositional content of an utterance
depends on context. According to most authors, what influences meaning is the
cognitive context pertaining to the speaker or supposed by the speaker to be shared
with the audience, while for others it is, instead, the situational context, that is, the
circumstances of the utterance (cf. Gauker 1998 for this contrast). Contextualism,
though, does not hark back to any hypothetical primacy of the illocutionary act over the
propositional content. The possible role of the kind of ongoing conversational activity
and of the conversational goals of speaker and audience, which might be of great
interest to a speech act theoretic perspective on the contextual determination of what
an utterance says, is hardly touched upon in contextualist literature.7
Jennifer Hornsby (1988) criticized Searle’s distinction between propositional act and
illocutionary act, remarking that it reduces the description of the locution to its
“indicative core”, that is its truth conditions. But utterances also involve mood, and as
Austin already noticed, rhetic (or oratio obliqua) reports of non indicative utterances
cannot be introduced by the formula X said that, but have to use other introductory
formulas. So, it is reasonable to take mood as belonging to the rhetic act, introducing
specifically rhetic notions such as “indicatively saying”, “interrogatively saying”, or
“imperatively saying”. Also Kent Bach (1994) suggests distinguishing the locutionary
act from the expression of a propositional content. In figurative (or implicature based,
see Bianchi, this volume) uses of language such as metaphorical assertion, the
locutionary act performed does not yet involve the expression of the proposition
which the speaker intends to convey and therefore does not yield the real truth
conditions of the utterance. Still, it may be deemed to “say” what it literally means
insofar as we stick to the consideration of the words uttered, their syntax and
morphology, and their standard lexical meaning in the language (or Kaplan’s
“character”, see Kaplan 1989). Thus, the particular case of metaphorical assertions
may show that a neutral, abstract notion of locutionary act and a truth conditional,
intentional notion of expressed propositional content can co exist and both be of use in
the analysis of speech activity.

3.2. The locutionary act: inferentialist perspectives

Kent Bach and Robert M. Harnish, in their volume Linguistic Communication and
Speech Acts (1979), propose a view of linguistic communication as inferencebased: the
speaker provides, by saying what he says, “a basis for the hearer to infer what he
speaker intends to be thereby doing” (1979: 4–5). Such a view belongs to a theoretical
framework quite distinct from Austin’s, that is, the intention based conception of
meaning proposed by Paul Grice (see his 1989a; Kemmerling, this volume), to which
we will return below (Section 4.2). They apply an inferential analysis to the process of
producing and understanding a speech act, starting from the speaker’s utterance of
a linguistic expression. Thus, their inferential analysis or “Speech Act Schema”
applies already to the Austinian level of the locutionary act. Bach and Harnish
distinguish between the utterance act (the act of uttering a linguistic expression
addressing it at an audience in a context) and the locutionary act (the act of saying, to
an audience in a context, that so and so)(1979: 3), keeping the Austinian label for the
latter. Between these two acts there is an intermediate level, which Bach and Harnish
describe as the speaker’s meaning something by the linguistic expression she utters.
The first step in the inferential path thus outlined leads the hearer from merely
hearing the utterance to understanding that the speaker is uttering a certain
linguistic expression and corresponds to the hearer’s recognition of the utterance
act. The second step leads from the fact that a certain expression is uttered to the
recognition of what the speaker means by it, which, in normal cases, is one of the
possible meanings that the expression uttered has in the language to which it belongs
(Bach and Harnish call this selected meaning “operative meaning” of the linguistic
expression; see 1979: 20–23).
A third step leads from the selected intended meaning to the identification of what the
speaker is saying to her audience (1979: 8–10, 24). These inferential steps, particularly
the second and the third, are made possible not only by the speaker’s verbal behaviour,
but also by the hearer’s mastery of the language, the presumption that hearer and
speaker share one and the same language (1979: 7, 20–21), and their mutual contextual
beliefs (i.e. the salient contextual beliefs that both speaker and hearer have, believing
each that the other has them, and that the other believes that she has them) (1979: 5,
20–21, 24). This third level corresponds, in Bach and Harnish’s Speech Act Schema, to
Austin’s locutionary act. It is at this level that it is determined, for example, what in the
world the speaker is referring to by the referring expressions she uses. The performance
and reception of the locutionary act are described by Bach and Harnish in greater detail
as involving both the expression of a proposition and the use of a sentence type
suitable to the further performance of an illocutionary act (see below, Subsection
4.2.2.2). Bach and Harnish, indeed, consider some main illocutionary force indicators
such as sentence type and mood as contributing to what is said in the locutionary act,
since indirect speech reports take them into account. When the sentence uttered is
declarative, what is said is specified by standard truth conditions, while for other
sentence types, Bach and Harnish propose a generalizing strategy that enables the
construction of indirect speech reports comprising declarative that clauses (1979: 9,
25).8
Inferentialist theories of communication have received greater impulse since the
publication of Relevance: Communication and Cognition by Dan Sperber and Deirdre
Wilson (Sperber and Wilson [1986] 1995). For these authors, the proposition expressed
by an utterance is inferentially retrieved by the hearer from a more abstract
propositional form, on the basis of contextual assumptions, in conformity with the
Relevance Principle (which prompts participants in a conversation to plan and
understand one’s own and the other’s utterances as optimally relevant) (see Bianchi,
this volume). The proposition retrieved in this way, being the explicit content of the
utterance, is called by Sperber and Wilson “explicature” (1995: 182). The nature and
functioning of explicatures has been discussed at length in Relevance Theory and in the
neo Gricean context (see e.g. Carston 2002, Levinson 2000) (see Bianchi, this volume,
Subsections 10–12). Most contextualist philosophers maintain that the propositions
actually expressed by utterances are also retrieved by the audience inferentially. It is to
be noted, though, that in both the relevance theoretic and the contextualist literature,
the expression of a proposition is generally not thematized as an “act” performed by
the speaker.

3.3. Speech act theory and use theories of meaning

Some philosophers of language (following Wittgenstein 1953 and ordinary language
philosophy) have preferred use theories of meaning, that is, theories according to which
the meaning of linguistic expressions depends on the way in which they are used, to
theories inspired by the truth conditional paradigm (such as Frege’s, Davidson’s, or
model theoretic semantics). Among these philosophers there are Michael Dummett,
William Alston, and Robert Brandom. Use theories offer intuitive foundations to
meaning and are in general connected with a view of speech as a social, rule governed
activity. One would expect such theories to shed some light on the locutionary act (or:
on meaning, as produced by locutionary acts). But do they?
One well known kind of use theory of meaning goes back to the work of Michael
Dummett (see e.g. his 1975). Dummett argues for a conception of meaning as use which
connects use with proof (since using a declarative sentence correctly – thereby making
an assertion – requires knowing how to prove its truth) and with inference (so that
purporting to analyse the use of a linguistic expression amounts to considering the rules
of inference according to which it is employed.9 Brandom (1994, 2000) presents a
particularly complex kind of inferentialism, explicating meaning in terms of the
inferences to which the speaker is committed by using a given linguistic expression in an
assertion. A use theory of meaning more closely connected to speech act theory is that
of William Alston (1964, 2000). Alston (1964: 32–39) aims at an explicit analysis of
semantic concepts, inspired by the Wittgensteinian principle that meaning is use. He
connects sentence meaning with the communicative function that the utterance of the
sentence is designed to perform, and considers sameness of communicative function as
sameness of meaning. He borrows Austin’s term illocutionary act to indicate the
relevant kind of communicative function. Two sentences, he claims, have the same
meaning if they are commonly used to perform the same illocutionary act and thus have
the same illocutionary act potential (1964: 36). To illustrate sameness of illocutionary act
potential, Alston proposes pairs of sentences such as Das ist gut / That’s good, That’s my
paternal grandmother / That’s my father’s mother, and also Can you reach the salt? /
Please pass the salt. He further developed his view that sentence meaning is
illocutionary act potential, arguing against objections and alternative proposals, in his
(2000). A general problem with his view appears to be the difficulty to define the notion
of illocutionary act potential in more than merely intuitive terms. Moreover, once it is
claimed that for a sentence to have a certain illocutionary act potential is for that
sentence to be governed or to be subject to a certain rule (Alston 2000: 190), we are left
with the problem of how the rules actually governing sentences (and a given sentence
in particular) should be identified.10
Use theories of meaning encounter problems when they attempt to clarify the meaning
of particular linguistic expressions (going beyond basic examples): describing use,
however defined, is far from easy and the specification of rules raises problems about
their nature and the activity of rule following. Use theories may also encounter
difficulties in accounting for reference (especially when the theory has an internalist
character, that is, is mainly concerned with relationships of words and concepts to other
words and concepts in the mind) and in analysing the relationship between complex
linguistic expressions and their components (since use, qua communicative function, is
more easily identified for linguistic expressions of the sentential level). When (as in the
case of Alston) they present themselves as contributions to speech act theory, these
views reveal a structural peculiarity. Speech act theory typically admits of (at least) two
different levels or kinds of use of language, co existing within the total speech act.
Austin distinguishes the locutionary act from the illocutionary, both of which may be
said to be uses of language, and Searle distinguishes the expression of propositional
content from the indication of illocutionary force. Proposers of use theories of meaning,
since Wittgenstein’s Philosophical Investigations (implicit target of criticism in Austin
1975: 104, see above, 2.2.1), need not be interested in distinguishing such two levels
and, insofar as the uses of language they are concerned with belong de facto to the
illocutionary level, display a tendency to explain meaning by tracing it back to
illocution, thus opening the way to the conflation of meaning and force. Thus, in Alston
(2000) the locutionary or propositional aspect of the speech act is absorbed within finely
cut illocutionary act types and, in Brandom (1994, 2000), what content is asserted
appears to depend on the commitments that its assertion produces (rather than the
other way around).11
Generally speaking, these use theories of meaning, in spite of their thematic closeness
to speech act theory and sometimes their terminological loans from it, fail to give
pertinent contributions to the characterization of the notion of sayingas itself a doing, to
which (in the most classical speech act theoretic framework) the notion of meaning
should pertain.
It should be said, however, that there is another theory of meaning, closely connected
with speech act theory, which might be classed with use theories: that of Paul Grice (see
Kemmerling, this volume). The kind of use of words on which Grice (at least in his first
study on this topic: [1957] 1989b) bases his notion of meaning is not the locutionary or
the illocutionary, but the perlocutionary (see above, 2.3 and below, 5): the speaker’s
intent to achieve a certain psychological effect on her audience lies at the core of Grice’s
analysis of the complex intention, aiming to be recognized by the audience and to
achieve its effect thanks to this, which he takes to be constitutive of meaning. But there
is an important aspect of Grice’s philosophy of language which is opposed to the views
of use theorists: he accepts to define “what is said” (meaning qua propositional content)
in terms of truth conditions (Grice [1975] 1989e). A specific contribution of Grice’s that
might be brought to bear upon the understanding of the locutionary level of the speech
act is his distinction between the dimensions of dictiveness and formality: the former
has to do with whether the relevant meaning is part of what an expression says as
distinct from implies, the latter with whether the relevant meaning is conventional or to
be retrieved inferentially (Grice 1989a: 360–362). We will come back to Grice’s analysis
of meaning in terms of intentions below (Section 4.2), since it has been very influential
on re definitions of the illocutionary act.

4. Vagaries of the illocutionary act

Austin meant the illocutionary act to be the core of his theory. He complained that
philosophers and other scholars concerned with language were used to paying attention
to the locutionary level (saying something, possibly true or false) and to the
perlocutionary level (doing something by saying something), largely missing the
intermediate level of illocution. However, his lack of detail and argument as regards the
effects of the illocutionary act, together with the lack of definiteness of the notion of
conventionality he employs, made it difficult for subsequent authors in speech act
theory to see clearly what role illocution was designed by him to play, so that some
precious suggestions of his (especially as regards the core effect of the illocutionary act
and its conventional character) have almost been lost. Philosophers and linguists
interested in analysing illocutionary acts (from assertion to greetings, from request to
promise, from the naming of a ship to congratulations) found additional sources of
inspiration in the Wittgensteinian notion of language game and the related conception
of speech as rule governed behaviour, or in Grice’s analysis of meaning as speaker
intention.
It should be noticed, incidentally, that a large part of the literature concerning illocution
makes hardly any use of the term illocutionary act, but finds it enough to speak of
speech acts in general. This attitude can be traced back to John Searle’s identification of
the illocutionary act with the whole speech act, based on the remark that the expression
of a proposition can only occur within a speech act which, in turn, need have
illocutionary force (Searle 1969: 25).12 While Searle’s proposal was, perhaps, only a way
of emphasizing the significance of illocution, its long term effects involved a partial
underrating of the specificity of the level of illocution, embodied in the feeling that the
expression illocutionary act could be dismissed without loss. We will disregard these
terminological variations and use the term illocutionary on every occasion in which we
will be speaking of some element (act, force, indicator …) pertaining to the level of
illocution.
In the study of illocutionary acts, or, if you prefer, of speech acts qua illocutionary acts,
one can distinguish two main trends: the conventionalist and the intentionalist. The
former stresses the role of conventions and rules in making it possible for a speaker to
perform an illocutionary act. The latter stresses the role of speaker intention at the core
of the speaker’s performance. Conventionalism may be associated with some form of
externalism (if conventions or rules refer to external circumstances and actual
behaviour), but may also be made compatible with an internalist perspective (if the
content of conventions or rules refers to mental states and attitudes, as happens at
least in part in Searle 1969; 1979a). Intentionalism is essentially internalist (cf. Harnish
2009). In the following sections we will examine the accounts of illocutionary acts as rule
governed uses of language put forward by John Searle and William Alston. These views
can be considered as conventionalist (but we will see that both accounts also leave
room for intentions and other mental states). On the intentionalist side, we will take
into consideration various steps and aspects of what may be called the “Gricean
tradition” in speech act theory, stemming not from Grice himself (who only marginally
took illocutionary acts, or forces, into consideration), but from other philosophers who
used his analysis of speaker meaning to redefine the illocutionary act. In the conclusion
to this Section, we will consider some of the ways in which the role of conventionality in
illocution has recently been reassessed.

4.1. Illocution as rule governed behaviour

In his analysis of performative utterances, Austin proposed a non exhaustive set of rules
the violation of which hinders the utterance from happily or felicitously bringing about
what it is designed to. One prescribes the very existence of an accepted conventional
procedure for bringing about a certain effect; another requires speaker, other
participants and circumstances of the utterance to be appropriate to the procedure, that
is, such as the procedure requires them to be. Two other rules express the requirement
that the procedure be executed correctly and completely (one might understand:
correctly and completely enough). The violation of rules belonging to these four kinds
may make the performative utterance ineffective, and the envisaged act infelicitous and,
specifically, “null and void”. Two further rules require the speaker to entertain
appropriate mental states and attitudes and to behave consequentially. The infelicities
generated by violations of these two last rules do not make the act null and void, but
make its performance an abuse (because of insincerity or break of commitment) (Austin
1975: 15–18). Austin applies the same notions of procedure, conditions, and infelicity to
illocutionary acts (1975:105–106)13, the kind of acts that (successful) performative
utterances perform.
This aspect of illocution is focused upon and reinterpreted by both John R. Searle and
William P. Alston. For Searle, speech is a rule governed activity (cf. 1969: 16). Attention
to rules or conditions which make it possible for a speaker to perform a certain
illocutionary act, partly inherited from Austin, characterizes his early work, but does not
disappear in its subsequent developments, where “counts as” rules play a central role in
the explanation of social reality (Searle 1995, 2010). Alston started working on “linguistic
acts” in the 1960s (see his 1964), largely under the influence of Wittgenstein’s
Philosophical Investigations (1953). The recent developments of his approach (see
Alston 2000) further elaborate upon the rules governing illocutionary acts and the role
they play in their actual performance.
4.1.1. Searle’s conditions for the performance of illocutionary acts

Searle’s investigation of illocutionary acts starts from the question: what conditions are
necessary and sufficient for an illocutionary act to have been successfully and non
defectively performed in the utterance of a given sentence? (1969: 54). The search for
what have later on been called the felicity conditions of illocution ary acts (the term is
not used in this precise form either in Searle 1969 or in Austin 1975) is for Searle a
search for “necessary and sufficient” conditions and involves therefore a claim to
exhaustiveness that Austin would never raise. Thus, Searle requires that “normal input
and output conditions” obtain, that the speaker has the intention to perform a certain
illocutionary act, that the speaker intends to make the hearer understand the meaning
and force of her utterance, that the sentence uttered is appropriate to the
performance of the intended illocutionary act. Beyond these conditions having a general
character, each illocutionary act type has got its propositional content conditions
(requiring that a proposition be expressed and that it be of a certain kind), preparatory
conditions (requiring speaker and hearer to possess certain capacities and attitudes),
and a sincerity condition (requiring the speaker to entertain the attitude or
psychological state expressed by her illocutionary act) (Searle 1969: 54–71). The
correspondence of these kinds of conditions to some of Austin’s rules (those concerned
with appropriateness to the procedure, completeness of execution, and the mental
states or attitudes that the participants are required to entertain) is intuitive. Searle
firstly illustrates both the general and the type specific conditions by reference to the
act of promising and then generalizes them to other types of illocutionary act in a
synoptic table (1969: 66–67). Observance of all of the conditions of a certain
illocutionary act type has the effect of making the utterance of a sentence count as the
successful and non defective performance of an illocutionary act of that type.
Among the many comments that can be made about Searle’s so called felicity
conditions, I would like to highlight the following ones.
(i) The most general conditions, common to all illocutionary act types, are very
demanding. Consider the first one. What are the “normal” input and output conditions
of speech? If the conditions posed are too strict, few actual situations will meet them,
while, if they are too loose, they might admit unwanted cases. Moreover, could we ever
be sure whether the identifying features of “normal” situations actually obtain in a given
case? And, finally, if we know how to make sure whether the identifying features of
normal speech situations obtain, any speech activity occurring in a situation that fails to
display these features is denied the status of the performance of illocutionary acts.
Thus, Searle’s first condition excludes non serious speech from the core object of speech
act theory (which indeed was Derrida’s charge against speech act theory) much more
radically than Austin ever did (see above, Subsection 2.2.1).
(ii) Another general condition is concerned with the intention to make the meaning and
force of one’s utterance understood. This is the “illocutionary effect” in Searle’s sense
(1969: 47), which is different from Austin’s. In Austin, the effect taken by an illocutionary
act is, as we have seen above in 2.2.3, its conventional effect, that is, the effect that the
procedure, to which the illocutionary act belongs or in which it consists, is designed to
bring about. Searle’s “illocutionary effect” is, rather, inspired by Grice’s analysis of
speaker meaning.
(iii) Most conditions are concerned with attitudes or mental states of speaker and
hearer, such as intentions, wishes, or beliefs. It is particularly relevant to notice that
Searle’s essential condition consists of an intention of the speaker: shortly said, the
intention to perform a certain kind of illocutionary act. Since all conditions are
necessary, it follows that in order to perform an illocutionary act a speaker must
have the intention to perform it. This is less trivial than it may seem at first sight. While
most actions may be performed unintentionally, one cannot, in this view, perform an
illocutionary act unintentionally. The type of illocutionary act performed (if one is
performed at all) depends on the speaker’s intention to perform an act of that type.
(iv) Finally, the inclusion of the sincerity conditions among the “necessary and sufficient”
conditions for the performance of an illocutionary act is problematic. Searle realizes
this is so and (following Austin) concedes that actual sincerity is not necessary for the
successful performance of the illocutionary act, but only for its non defectiveness. He
thus concedes that there are such things as lies and insincere promises. However, he
maintains that even in such anomalous cases, the speaker has to entertain at least the
intention to be made responsible, by her utterance, for having the appropriate belief or
intention (or other mental state or attitude)(1969: 62).14 Lack of this intention would not
make the illocutionary act merely defective, but undermine its performance.
An important complement to Searle’s analysis of felicity conditions is his claim that the
linguistic devices we use in order to characterize an utterance as having a certain
illocutionary force are used appropriately only if the conditions for the performance of
the corresponding illocutionary act are satisfied (1969: 62–64). This makes felicity
conditions become part of the codified meaning of illocutionary force indicating devices
and stengthens the relationship between speech acts and the syntactic and semantic
dimensions of natural languages (cf. also Searle [1975] 1979b: 20–27).15
In perfectioning his speech act theory, Searle (1979b) sets out to describe five main
classes of illocutionary acts (cf. Kissine, this volume, Table 1). In so doing, he derives
some main dimensions for the characterization of types of illocutionary acts from his
previous analysis of felicity conditions. To the sincerity conditions there corresponds the
expressed mental state: that is, illocutionary act types can be described according to
whether they express a mental state or not and, if they do, the kind of mental state they
express. To the essential condition, that is, the condition requiring that the speaker have
the intention to perform just that type of illocutionary act, there corresponds the
illocutionary point. The kind of things the speaker may aim at are thus specified by the
five main illocutionary points. To these two dimensions of characterization of
illocutionary act types, Searle adds what he calls direction of fit. This feature of the
illocutionary act is related to its having a propositional content, which may fit or fail to fit
the way the world is. But the direction in which such a fit is to be sought and aimed at
varies with illocutionary point. Illocutionary acts aiming to direct the addressee’s
behaviour have worldto word direction of fit: the world has to fit the propositional
content of the illocutionary act (the same holds for illocutionary acts aiming to commit
the speaker to do something). Illocutionary acts stating how things are have a word
to world direction of fit: it is the words that have to fit the way the world is. Illocutionary
acts whose whole point is to express a mental state or attitude of the speaker’s have no
direction of fit and illocutionary acts bringing about new states of the world, such as
appointing, resigning, or declaring a session open, have, according to Searle, both
directions of fit at once.16
In Searle and Vanderveken’s Foundations of Illocutionary Logic (1985), a study
developing Searle’s speech act theory in the direction of a general axiomatics of
illocutionary acts, among the components of illocutionary force there appear (besides
propositional content conditions and preparatory conditions, as in Searle 1969, and
illocutionary point and expression of a mental state, as in Searle 1979b), also mode of
achievement (that is, the particular way, if any, in which the illocutionary point has to be
achieved), degree of strength of the illocutionary point (e.g. differences such as that
between requesting and insisting), and degree of strength of the expression of the
relevant psychological state (e.g. differences such as that between requesting and
imploring) (1985: 12–20). Searle and Vanderveken claim that every possible illocutionary
force can be constructed by specifying these seven elements (some of which, in the case
of elementary illocutionary forces, may take zero as a value) or recursively operating on
them. Since illocutionary points are finite in number, there is a finite number of
elementary illocutionary forces in which the propositional content conditions, the
preparatory conditions, and the expressed psychological state are determined only by
the illocutionary point, while mode of achievement and degrees of strength take value
zero. Searle and Vanderveken’s illocutionary logic is concerned with studying how all
other illocutionary forces can be obtained from these elementary ones.17

4.1.2. Alston: illocutionary rules and speaker’s responsibility

According to William P. Alston, when a speaker utters a sentence she may bring about
one or more results, all of which are effects of her utterance (such as getting the hearer
to open the door, irritating the hearer, distracting someone who is reading). At the
source of all this, there is the act of uttering a sentence, which Alston calls a sentential
act (roughly corresponding to Austin’s phatic act), while among all the things done, there
are the illocutionary and the perlocutionary act, roughly coextensive with Austin’s but
characterized differently (Alston 2000: 25). The illocutionary act, in particular, is
characterized by Alston as the kind of act that has “content” and that is specified by
oratio obliqua reports (2000: 26). In the case of an utterance of:
(5) Please, open the door.
we can therefore say that the speaker is asking someone to open a certain door, and
thus apply to her an illocutionary act concept. Differently from Austin, Alston does not
think that there is any particular kind of effect that the utterance of (5) must have in
order for the speaker, in issuing it, to have asked someone to open a door.
Against the early proposers of intention based conceptions of illocution (such as Schiffer
1972: see below, Subsection 4.2.1), Alston refuses to define the illocutionary act as a
sentential act with perlocutionary intentions (2000: 40–45). He admits that illocutionary
acts are rule governed, but, unlike Searle, does not envisage their rules as necessary and
sufficient conditions for their performance and describes a subtler, but no less
important, connection between the performance of the illocutionary act and the rules
governing it (2000: 51–80). A speaker may perform a certain illocutionary act even if she
violates one or more of its conditions. Suppose that the door referred to in issuing (5) is
already open:the speaker issuing (5) can nevertheless be said to have issued a request,
albeit, perhaps, an idle one. Or suppose the speaker of (5) is not in a position to request
anything of her addressee. Her act might be wrong and subject to sanctions as
inappropriate or impolite, but she can still be said to have made a request. However,
there is something in an illocutionary act’s being rule governed that makes it possible for
a speaker to perform it: the speaker, in issuing the related utterance, has to recognize
that the illocutionary act it is designed to perform is governed by certain rules and that
therefore certain conditions should hold (Alston 2000: 54–55). It is this recognition that
enables the speaker to perform the illocutionary act. If the speaker were to admit that
one of these conditions did not hold, it would become impossible for her to perform the
illocutionary act. So, for example, in order for a speaker, in issuing (5), to have asked
someone to open a certain door, she must recognize that there are rules governing the
act of making a request and that the conditions they require hold. So, according to
Alston, in performing an illocutionary act the speaker takes responsibility for certain
conditions being the case.This analysis enables Alston to recognize the conventional
effects of certain illocutionary act types somewhat indirectly, that is, as something the
speaker takes responsibility for. However, in his framework there seems to be no
difference between the conventional effects of illocutionary acts (such as speaker’s
commitment in promising) and the conditions their rules require to hold, since the
speaker has to take responsibility for both in order to perform her illocutionary act.
Since the performance of the illocutionary act depends on the speaker’s taking
responsibility for certain conditions to hold, as distinct from the satisfaction of those
conditions (or the actual carrying out of a certain procedure), Alston’s conception of the
illocutionary act turns out to be in its own way intention based and even internalist:
conceiving of illocutionary acts as rule governed does not in principle exclude
intentionalism.

4. 2. Illocution as speaker intention

Speaker intention was brought to the fore in the analytic philosophy of language by Paul
Grice, an Oxford philosopher who participated in ordinary language philosophy, but
developed a perspective of his own on language and meaning. As is well known, he
proposed an analysis of what it is for a speaker to mean something by a linguistic
expression (or gesture or other sign), cast in terms of the kinds of intentions that the
speaker has in using it (see above, Section 3.3, and Kemmerling, this volume). Grice’s
analysis of speaker meaning was turned into a model of the illocutionary act by Peter F.
Strawson, inspired Stephen Schiffer’s view of meaning and speech acts, influenced John
Searle’s conception of the speech act,18 and later on inspired the view of linguistic
communication of Kent Bach and R.M. Harnish. It also inspired Kemmerling’s view of
illocutionary acts as “Gricy” or “quasiGricy” actions (2001), which, however, is no longer,
properly speaking, an intention based view of illocution (see below, 4.3).

4.2.1. Speaker meaning as a model of illocution

Paul Grice does not mention nor discuss Austin’s notion of illocution. He, though, is
sensitive to the fact that language has different uses and this awareness is reflected in
the distinctions he makes among the possible contents of core intentions in his analysis
of speaker meaning. For example, the response that the speaker intends to produce in
the hearer may consist of the belief that p (in the case of indicative type utterances), or
the action p (in the case of imperative type utterances) ([1957] 1989b). In a later version
of the theory ([1968] 1989c: 123), the intended response may consist of the belief that
the speaker believes that p (in the case of exhibitive indicative type utterances;
protreptic indicative type utterances are made with the further intention that the hearer
himself believes that p), or of the intention to do p. These distinctions apparently belong
to the speech act theoretic level of illocution, since they are tied up with what Austin
called the inviting of a response, and even more with Searle’s illocutionary points and
directions of fit. It would seem that in fact Grice’s speaker meaning always comes with
some illocutionary force, or at least some (sketchy) indication of it, mainly concerning
the speaker’s expressed psychological states. This feature of Grice’s analysis is resumed
by Stephen Schiffer (1972: 95–99) in his analysis of illocutionary acts. He highlights two
main kinds of core intentions within the speaker’s complex meaning intention,
corresponding to assertive and imperative speech acts, and distinguishes between them
basically by means of a relationship with the world corresponding to Searle’s direction of
fit. Schiffer also considers the different means by which the speaker may intend her
core intention to be fulfilled and attempts some correlation with a main linguistic
illocutionary force indicator such as sentence type.19
A decisive contribution to shaping the relationship beween illocution and speaker
meaning was made by Peter F. Strawson in his ([1964] 1969). Strawson proposed a
distinction between two main groups of illocutionary acts: those which are the acts they
are because of some convention that the speaker follows, and those which depend on a
particular kind of intention of the speaker. This intention, often called communicative
intention, may be described as the intention to achieve an effect on the audience by
means of the audience’s recognition of the intention to achieve that effect and
intending also that the intention that the audience recognize the speaker’s intention to
achieve that effect be recognized by the audience (Strawson 1969: 386–389). Most
everyday illocutionary acts belong to this second kind, while conventional speech acts
occur mostly in the context of institutions and, according to some speech act theorists,
may even turn out not to be real illocutionary acts, since language is not indispensable
to their performance (Urmson 1977; see also Schiffer 1972: 93).
Strawson argues that Austin, in claiming that illocutionary acts are not successfully
performed unless the hearer’s uptake is secured, and in identifying uptake with “the
understanding of the meaning and of the force of the locution” (Austin 1975: 117),
adumbrated a conception of the illocutionary act to which the audience’s understanding
of speaker intention is central. According to Strawson, the speaker performing an
illocutionary act must have a complex, reflexive intention of the kind proposed by Grice
in his conception of speaker meaning, and Austin’s illocutionary acts can (at least for
those cases that do not depend on the following of conventions), be explicated as cases
of Gricean speaker meaning (as reformulated by Strawson 1969). This application of a
Grice inspired conception of speaker meaning to illocution diverges from Grice’s original
notion insofar as the relevant effect on the audience, to which the speaker is taken to
aim, is not the production of a belief or the elicitation of an intention, but the mere
understanding of the speech act. One might say that also the understanding of a speech
act can be represented as a belief, or a set of beliefs, of the hearer (cf., in Grice, the
abovementioned intended effect of the exhibitive use of indicative type utterances). But
even if we take understanding to be a belief, there is a restriction to the kind of content
of the belief to be elicited, since this belief has to be about what the speaker means and
intends to do with her utterance. Moreover, Strawson proposes uptake, or
understanding, as the intended effect of a speech act also in cases of illocutionary acts
performed in uttering non indicative sentences.
In his (1969), Searle largely agrees with Strawson as regards the relationship of
illocutionary act, speaker meaning and uptake. He too shows critical appreciation of
Grice’s notion of speaker meaning. In his reformulation of it (1969: 50), he makes two
main points: first, that the speaker’s core intention should be seen as directed merely at
making the hearer understand the speaker’s utterance (anything more demanding
would, according to him, amount to perlocution); second, that this understanding
should rely upon the hearer’s knowledge of the rules governing the uttered sentence.
Thus, Searle includes a Gricean suggestion in his analysis of speech acts as rule
governed, taking intentions of the kind that characterize speaker meaning (in his version
of this concept) as a necessary condition for the performance of illocutionary acts.
Both Searle and Strawson deem the uptake effect to be the only effect that is
necessarily connected with the successful performance of an illocutionary act: Searle
even calls it “illocutionary effect”. His terminological choice is largely responsible for the
disappearance of what Austin called “conventional effects” (the effects illocutionary acts
“take” provided they are successful in achieving uptake and not otherwise gravely
infelicitous, see above, 2.2.3) from the field of speech act theory and from the study of
illocution in particular. Indeed, readers of Strawson (1969) and Searle (1969) can but
assume that the only effect produced as integral to the performance of the illocutionary
act, whenever a speaker succeeds in performing an illocutionary act, is its uptake, that
is, the understanding of the meaning and force of the speaker’s utterance.

4.2.2. Speaker intention and audience’s inference

Once illocutionary acts (at least those that do not belong to the institutional,
conventional group) are seen as based in intentions that are intended to be recognized,
a further step arises quite naturally. How are these intentions to be detected by the
audience? Searle’s revised version of the Gricean scheme sticks to the idea that it is
thanks to the rules governing the linguistic expressions used, that the audience can
recognize the speaker’s intentions. But from the way in which Grice (see e.g. [1969]
1989d: 103) reconstructs step by step the way in which the hearer is intended by the
speaker to recognize her intention(s), all the way up from noticing the features
possessed by the utterance and the correlation of those features to certain types of
response, one is easily led to conceive of this process as inferential. Searle too admits of
an inferential retrieval of illocutionary force in cases that cannot be explained by resort
to conventional illocutionary force indicators ([1975] 1979c). In Kent Bach and Robert M.
Harnish’s work Linguistic Communication and Speech Acts (1979), the recognition of
both meaning and illocutionary force is a matter of hearer’s inferences.
4.2.1.1. Searle’s distinction between direct and indirect speech acts
On the way towards a general characterization of the understanding of illocution ary
force as inferential, a half way position is represented by Searle’s distinction between
direct speech acts, recognizable in virtue of the conventions they conform to, and
indirect ones, whose understanding is based on inferential processes (Searle 1979c).
Searle notices that utterances of certain illocutionary act types, for example requests,
are often not cast into the linguistic forms conventionally assigned to the illocutionary
act type to which they belong. In the case of requests, which are directive speech acts
whose conventionally associated linguistic form is the imperative sentence type, the
utterances performing them may take interrogative or even indicative forms. Inspired by
Grice’s view of conversation (Grice [1975] 1989e, see Bianchi, this volume), Searle
(1979c) claims that the performance of an illocutionary act, in case of mismatch
between the intended illocutionary act type and the linguistic form of the utterance, is
made possible by the activation in the hearer of an inferential path involving premises
comprising: the assumption that the speaker, being cooperative, is not attempting to
perform patently pointless speech acts; assumptions belonging to speech act theory
such as those concerning the preparatory conditions of a certain illocutionary act; and a
variable number of contextual assumptions. Consider:
(6) Can you pass me the salt?
As a question, this utterance would be pointless. It is of no use or interest to the speaker
to be told whether or not the hearer is able to do such a simple, effortless thing as pass
the salt across the table. Moreover, very likely, the reply to such a question cannot but
be affirmative: the speaker sees how the hearer behaves, how he moves his hands and
arms in eating and doing other things, knows that the salt is within his reach, and that
the salt’s container is a small object quite easy to handle. So why should the speaker ask
such a question? If it is assumed that the speaker is conversationally cooperative and,
therefore, that her utterance has some aim or point, one will conclude that the speaker
is not asking a question, but doing something else, related in some way to the form and
content of her utterance. Searle notices that the addressee’s ability to comply with the
request is among the preparatory conditions of the request as an illocutionary act. From
the fact that the speaker is asking whether a preparatory condition of the request to
pass the salt is satisfied, the hearer can infer that she actually intends to request him to
pass the salt. So (6) has as its real or primary illocutionary force that of a request and the
force of a question, which it appears to have, is only its secondary illocutionary force.The
linguistic form taken by indirect speech acts may be closer or farther with respect to the
standard linguistic form of the primary act performed. In (6), the propositional content
of the secondary illocutionary act is very close to that of the primary: in both cases
(apart from what is added by the modal can) it is a matter of the addressee passing the
salt to the speaker. But now consider:
(7) I like soup with a pinch of salt in it.
Here, the proposition expressed is not about the addressee and does not provide a
representation of his passing the salt to anybody. It is not even about the salt in the
container on the table. Nevertheless, in particular contexts and with the help of
appropriate contextual assumptions, (7) too can be uttered, and understood, as an
indirect request to pass the salt.
Searle (1979c) takes into consideration various forms of indirect speech acts, most of
which display a close connection with one or other preparatory condition of their
primary illocutionary force. A wide literature in linguistics and socio linguistics has
tackled indirect speech acts of even more disparate kinds. Considerations about the
different degrees and ways in which illocutionary acts may be indirect can be found in
Bach and Harnish (1979:173–202), who pay great attentionto the “standardization”
processes which indirect speech acts undergo, in the vast literature on politeness based
on Brown and Levinson (1987)(cf. Locher and Graham eds. 2011), and in research
concerned with mitigation (e.g. Blum Kulka, House and Kasper (eds.) 1989; see Caffi,
this volume; Walker, this volume).
It should be noted that the basic assumption underlying the distinction between direct
and indirect speech acts is that there is, for each illocutionary act type, a conventionally
associated linguistic illocutionary force indicator, which expresses the speaker’s
communicative intention directly. In comparison to this, any other way of making one’s
illocutionary act recognizable becomes indirect. But this picture of the functioning of
illocutionary force indicators is a simplification: indeed, they often come in clusters, may
be distributed at various places in the utterance, and may fail to be completely
consistent with one another, generating hybrid or even ambiguous configurations. A
way to challenge such a picture could be to say (along the lines set by Hornsby 1988, see
above, 3.1) that it is the (locutionary) “interrogatively saying” whether the addressee
can pass the salt which amounts, in context, to the (illocutionary) request to pass the
salt.
Direct speech acts, moreover, are usually not explicit performatives: for instance, we do
not usually preface our requests with “I request that” and we rarely thank someone by
saying “I thank you”. Thus, by accepting the direct vs indirect distinction, one might be
led to conclude that explicit performatives are indirect speech acts. This has been
argued, in different ways, by Bach and Harnish (1979: 203–208, 1992) and by Searle
(1989) (see Doerge, this volume).

4.2.2.2. Bach and Harnish on linguistic communication

The most decisive turn towards an inference based conception of illocution is due to
Kent Bach and Robert M. Harnish, who, in their volume Linguistic Communication and
Speech Acts (1979), outline an inferential view of linguistic communication. The
inspiration for such a view is broadly speaking Gricean, since it focuses on complex
intentions of the speaker, called “reflexive” intentions because part of their content is
that they be recognized and be recognized partly on the basis that this is intended, and
on the hearer’s retrieval of them. As to the illocutionary act, Bach and Harnish follow
Strawson and Searle in limiting the content of the core intention of the speaker (in their
terms, “illocutionary intent”) to getting her utterance understood (cf. 1979: 154).
Their inferential analysis of the process of producing and understanding a speech act,
dubbed “Speech Act Schema”, starts from the speaker’s utterance of a linguistic
expression and leads step by step, by means of inferences, to the locutionary act (see
above, Section 3.1) and to the illocutionary. The premises that are required for the
inferential process to go through include the initial remark to the effect that a speaker S
is uttering a sentence, the “linguistic presumption” that the hearer shares the speaker’s
language and can use his knowledge of it to identify “communicative presumption”
that the speaker is saying what she says with some recognizable illocutionary intent
(1979: 7), and a set of mutual contextual beliefs. The step from the recognition of the
locutionary act to that of the illocutionary act avails itself of the communicative
presumption and of the relevant mutual contextual beliefs. Further inferences are
needed to distinguish between “literal” illocutionary acts (in which what the speaker
says determines what she does therein: 1979: 10–11) and “nonliteral” ones (1979: 65–
76). The Schema invites fine grained descriptions of expressed mental states and
attitudes: after all, what is to be inferred in order to identify the illocutionary force of
the utterance is the attitude expressed by the speaker (1979: 15). This general analysis is
reflected in a detailed classification of communicative illocutionary acts (Bach and
Harnish 1979: 39–55).
In Bach and Harnish’s terminology, the notion of successfulness of an illocutionary act
undergoes a change. For Austin, successfulness coincides with actual performance,
which involves the bringing about of a conventional effect (once uptake is secured and
provided no misfire or other fatal flaw makes the illocutionary act null and void). For
Searle, it depends on the satisfaction, by the utterance and its context, of the necessary
and sufficient conditions for the designed illocutionary act type (Searle 1969, Searle and
Vanderveken 1985: 21–22) and again coincides with actual performance. In both cases,
the successfulness of an illocutionary act is conceived as internal to the act’s
performance: what is at issue is whether the speaker succeeds in performing her
illocutionary act. For Bach and Harnish, performing an illocutionary act is an easier
matter: it may be enough for the speaker to issue an utterance with a certain
illocutionary intent and express a suitable proposition. Depending on the illocutionary
act type, the satisfaction of some (but not all) of Searle’s preparatory conditions may be
required (cf. Bach and Harnish 1979: 56). But in general, when Bach and Harnish speak
of “successfulness”, they refer to communicative successfulness, which is external to the
act’s performance and depends on the actual recognition of the illocutionary intention
by the hearer. Thus, it seems, an illocutionary act that has in fact been performed (since
the speaker had a certain illocutionary intent and her utterance expressed a suitable
proposition, etc.) may achieve, or fail to achieve, communicative success depending on
whether the hearer actually recognizes, or fails to recognize, the illocutionary intent.20
Besides communicative illocutionary acts, Bach and Harnish also admit of a group of
conventional illocutionary acts (1979: 108–119), which affect institutional states of
affairs. These are illocutionary acts, but their production and understanding does not
follow the Speech Act Schema: they are animated not by a communicative intention, but
by a conventional one. A “conventional intention” is fulfilled not by means of its
recognition, but by means of satisfying a convention, that is, in Bach and Harnish’s
acceptation of that term, a counts as rule establishing the conditions at which in a
certain community and in a certain kind of context, an utterance of a certain type
counts as doing such and such.

4.3. Reassessing conventionality

Ruth Millikan (1998) argued for the conventionality of illocutionary acts from a new
viewpoint: what makes the act “conventional” is that there is a fixed routine we go
through in each of its occurrences. The most interesting contribution of this idea of
Millikan’s is that it makes us see the performance of illocutionary acts as an execution of
procedures or routines that can be, and in fact are, repeated. In this light, Austin’s and
Searle’s conventional rules or conditions are turned into features of the (prototypical)
routine or script which the performance of the illocutionary act is to carry out. This
opens the way to a view of illocutionary force as recognizable not in virtue of one, or
few, linguistic illocutionary force indicators, but in virtue of the physiognomy or
configuration formed by the whole speech act in its situation. But like most speech act
theorists, Millikan too is concerned with the conventionality of the means by which the
illocutionary act is performed, rather than with the nature of the action itself and
therefore does not pay attention to the bringing about of the (conventional)
illocutionary effect.
Some impulse towards a reassessment of the conventionality of illocution, more clearly
aimed at rethinking the illocutionary effect, comes from a debate concerning the
actional nature of the discourse of pornography, which did not originarily belong to the
philosophy of language or to pragmatics, but to political theory and the philosophy of
law. To defences of pornography in the name of the right to “free speech”, it was replied
that pornographic discourse too is action because it is “performative” (MacKinnon 1987)
and, more specifically, that it performs illocutionary acts having conventional effects
upon the status of women as members of a society and as speakers, classing them as
inferior and as not in a position to refuse to have sex (which amounts to a legitimation
of rape) (Langton 1993, Hornsby and Langton 1998). In the attempt to explain what is
wrong with a certain way of representing women and addressing them, that very side of
illocution has emerged which had been neglected and nearly forgotten for decades.21 It
is not by chance that an author involved in the pornography debate, Mary Kate
McGowan, also rediscovered the Austinian category of exercitive illocutionary acts in
order to apply it to the exercise of power that may occur within conversational
exchanges (McGowan 2004).
After long term work discussing and applying a view of illocutionary acts focused on
their conventional effects, mostly conducted in Italian (cf. Sbisà 1989), I too have
recently insisted on the conventionality of illocution as the key to understanding a
number of social and interactional phenomena related with roles and statuses,22 both
informal and institutional. In this view, as suggested by Austin (in his 1975 but also in his
manuscript notes, preparatory to his lectures: see Sbisà 2007), the nature of
illocutionary effects is conventional insofar as they do not consist of material or
psychological states of individuals, which once brought about cannot be cancelled or
annulled (if they are further modified, it remains a fact they were brought about), but
of states of agents whose occurrence depends on inter subjective agreement,
whether interpersonal or social, whether ad hoc or governed by conventions or rules.23
We constantly act upon such “conventional” states, which – since intersubjective
agreement upon the uptake of an illocutionary act is presumed by default (at least in
cases in which illocutionary force is made available by the speaker by means of socially
recognized illocutionary force indi cators) – might at any moment turn out, because of
some newly discovered fatal flaw in the procedure designed to put them into being, to
have never been brought about, but are real enough nevertheless, at least in the sense
that their by default existence is causally effective. In this perspective, the conventional
nature of illocutionary acts resides in their becoming effective only if uptake is secured
(precisely as claimed by Strawson 1969), but their becoming effective consists (contrary
to Strawson 1969 and many others) in bringing about or changing conventional states of
affairs.
A view of illocution that does not accept conventionality as an essential feature of all
illocutionary acts, but neither defines them in terms of intentions, was put forward by
Andreas Kemmerling (2001). He analyses at least certain types of illocutionary acts as
actions which are performed, by conceptual necessity, if what the agent does (or her
doing so) makes it clear that in so doing, she wants to perform an action of that kind.
Kemmerling calls these actions “Gricy actions”, since he credits Grice with suggesting the
idea that the recognition of a desire may provoke its satisfaction. So, for example, if her
uttering certain words makes it clear that the speaker thereby wants to issue a
command, then this is enough to make her utterance a command. This core or essential
structure may be accompanied by other specific (and optional) components both on the
side of conventions and of intentions. By conventions, Kemmerling means conventions
governing how the act is to be performed; he does not consider the issue of effects,
which, after all, is not excluded by his analysis (but only neglected or marginalized). But
most importantly, in discussing the role of intention, he argues convincingly that
illocutionary acts can be performed unintentionally if the speaker’s behaviour is such as
to count as the performance of a certain illocutionary act. In this context, Kemmerling
admits that as Austin said, a judge should be able to decide whether a certain
illocutionary act was performed, possibly checking the participants’ behaviour against
rules and conventions or in a weaker way, as I have suggested above, routines or scripts
connected with lexical elements belonging to the semantic field of terms for
illocutionary acts.

5. Mysteries of the perlocutionary act

Austin appears to have dealt with perlocution only in order to contrast it with illocution,
not for its own sake. He tends to explore the general issue of what it is for a speaker to
perform a perlocutionary act (as opposed to an illocutionary act) but neglects
possible specific issues concerning kinds of perlocution. This has perhaps contributed to
making his notion of perlocution difficult to understand – when not simply not pertinent
to speech act theory. As a consequence, there is little literature specifically involved with
perlocution, and there too, one can find misunderstandings of Austin’s original
definition as well as shifts in the meaning of the relevant speech act theoretic terms.
We will consider three problems about perlocution: the first concerns the relationship
between perlocutionary goals (or, in Austin’s terms, perlocutionary objects) and sentence
type or mood (indeed, it seems that sentence type or mood, beyond its contribution to
illocutionary force indication and its possible involvement in the locutionary or rhetic
level of the speech act, goes hand in hand with kind of perlocutionary object); the
second concerns the delimitation of perlocutionary effects, and the third concerns the
very possibility of considering perlocution as an action of the speaker’s.

5.1. Perlocutionary goal and mood

Grice (in his analysis of meaning [1957] 1989b), and following him Schiffer (1972), put at
the core of meaning the intention to achieve an effect on the audience. This intention is
aimed at eliciting either a belief or an action, which can both be classed as
perlocutionary goals. Indeed, the achievement of these goals involves the actual
production of a psychological state in the audience, or even of a behavioral response.
Thus, Grice’s speaker meaning appears to be connected with perlocution, and insofar as
it is also used as a model to explicate illocution, illocution too is made to appear as in
some way dependent on perlocution. This has led most speech act theorists to criticize
Grice’s analysis of meaning and reject at least some aspects of it (see e.g. Searle 1969:
42–50; Alston 2000: 42–50). But other authors found it interesting to emphasize the
emerging connection between perlocutionary goals (or some of them at least) and
illocution. Thus, Ted Cohen (1973) argues that certain perlocutionary goals should be
recognized as directly associated with certain illocutionary forces, while, for others, the
connection is weaker, and argues in favour of the acknowledgement of the role of
associated perlocutions with respect to illocutionary acts.24 A connection between
illocutionary force and perlocutionary goal also emerges when the distinction between
assertive and imperative speech acts is considered. This distinction should be a matter
of illocution (inso far as mood is an illocutionary force indicator), but at the same time,
or perhaps instead, of locution or rheme (insofar as the linguistic meaning of syntactic
and morphological features is involved). However, there are theoretical frameworks and
contexts in which a relationship between mood or sentence type and perlocution can be
noticed. Schiffer’s (1972) analysis of illocutionary acts makes the distinction between
assertive and imperative speech acts descend from the distinction between the two
main kinds of response that the core intention of the speaker may aim at, that is,
producing a belief and eliciting an action, and at the same time stresses its connection to
the contrast between the world to word and word toworld directions of fit. It might
therefore seem that there is a connection between kind of perlocutionary goal
(producing belief vs eliciting action) and the so called direction of fit of the illocutionary
act. In this light, Searle’s directive illocutionary point, that is, directing someone’s
behaviour, and its associated world to word direction of fit might turn out to be not so
much a matter of illocution, but of perlocution (whether in Austin’s sense of
perlocutionary object or in Cohen’s sense of associated perlocution).
Certainly making illocution depend on (attempted) perlocution runs contrary to the
original spirit of speech act theory, in which illocution was viewed as a new concept,
independent from those of saying and of producing consequential effects by one’s
speech, and designed to explain how these two levels are connected. Indeed, this is why
Grice’s analysis of meaning and Schiffer’s application of it to illocutionary acts were
criticized. But the fact that a perlocution is attempted may be part of an action script
belonging to the illocutionary level, and there might well be (indeed there are) linguistic
forms dedicated to indicating broad kinds of possible perlocutionary intent. This aspect
of the illocution perlocution relationship has been thematized by Klaus Petrus (2010),
who has called the act of performing a perlocutionary act by means of an illocutionary
act a perillocutionary act. Petrus maintains that it is perillocutionary acts, rather than
illocutionary acts, that best account for communication.25

5.2. Borders of perlocution

Perlocution has no upper border: any consequential effect of a speech act may be
considered as perlocutionary. If breaking news surprises you so that you trip and fall,
my announcement has not only been believed true by you (which is already a
perlocutionary effect) and thus surprised you, but has also made you trip, fall, and (say)
injure your ankle. This aspect of the so called “accordion effect” concerning actions and
speech actions in particular (see Austin 1975: 110–115; Feinberg 1964) meets general
consent, apart from those speech act theorists who prefer to limit the notion of
perlocutionary effect to intended perlocutionary effects (cf. for example Bach and
Harnish 1979: 81). The lower border of perlocution is more problematic. Is making you
hear my voice properly called a “perlocutionary” effect? Are manifestations of uptake
(that is, of the understanding of the meaning and force of the locution) themselves
perlocutionary effects? And what about those effects of speech acts that go beyond the
mere making the meaning and force of the utterance understood, among which many
conventional effects concerning obligations, commitments, entitlements and licenses?
As to the first two questions, a good answer is provided by Steven Davis (1979), who
proposes that not only the consequential effects of illocutionary acts, but also those
of locutionary acts should be taken as perlocutionary, provided the exercise of linguistic
competence is involved in their production. Indeed, if my particular pitch in shouting a
warning offends your ears, this is a consequential effect of an aspect of a speech act, the
phonetic act, which does not involve linguistic competence and should not therefore be
counted as perlocution. If my mentioning our old friend Tom (with whom I have
seriously quarreled) surprises you, it is the use of a certain name to refer to a certain
individual that elicits your reaction, and this is a perlocutionary effect depending on a
rhetic feature of the speech act. As to the third question, the answer is more
complicated.
As we have seen above, Strawson and Searle have read Austin’s distinction of three
kinds of effects connected with the illocutionary act in a reductive way, taking only the
first of the three, the securing of uptake, as genuine and general. Searle includes the
third, inviting a response or sequel, in the definition of the directive illocutionary point
(counting as an attempt to direct the audience’s behaviour) and the second, taking
conventional effects, in the definition of the declarative illocutionary point (changing
states of affairs in the world). I have claimed above (Section 2.2.3) that the second effect
too was meant by Austin to be general. Recognizing this makes the real difference
between claiming that illocutionary acts are conventional and claiming that they are not,
or not essentially, such (cf. Sbisà 2007). I think that it is fair to say that, in any case, if
there is a conventional effect the illocutionary act is designed to obtain, and this
conventional effect is subject to cancellation or weakening in correspondence to various
kinds of flaws the performance may turn out to have (and which make it “infelicitous”),
that effect is an illocutionary effect as opposed to a perlocutionary one. But if
illocutionary acts are defined, with Bach and Harnish, as “communicative” acts, then
they have the illocutionary effect of making the hearer understand the speaker’s intent,
that is, for instance (in commissives), that she intends to undertake a certain obligation,
and the obligation itself appears as an additional effect, even if it should be clear that it
is the conventional effect that the procedure of promising is designed to bring about. So
what was, for Austin, the effect taken by the illocutionary act may turn out to be dealt
with as perlocutionary.
The tendency not to recognize conventional illocutionary effects, shifting them outside
illocution towards perlocution, is widespread. It can be noticed also in Michael Geis’s
Dynamic Speech Act Theory, an attempt to embed speech act theory within a general
theory of conversational competence (Geis 1995). Geis reacts against the reductive
understanding of the effects of speech acts, but, perhaps because the standard picture
of the illocutionary act he confronts himself with is Searle’s (where the notion of
illocutionary act is associated with the sole effect of the hearer understanding the
meaning and force of an utterance), he does not vindicate illocution as the proper place
of interpersonal effects of the conventional kind and prefers to propose the new
categories of “transactional effects” and “interactional effects” (cf. e.g. Geis 1995: 11,
64).26 His examples of transactional effects make it clear that he is thinking of effects of
the deontic modal kind (see e.g. Geis 1995: 67–94), while for some aspects at least his
interactional effects (tied to face work and conversational sequencing) appear to be
classifiable as perlocutionary, but the choice of the new labels prevents his proposal
from throwing light upon the distinction between illocution and perlocution.

5.3. Are perlocutionary acts real actions of the speaker?

A further problem with perlocution, directly arising from its definition and the
description of its core cases (such as convincing, persuading, alerting, getting someone
to do something), is whether the perlocutionary act can really be considered an act of
the speaker (Gu 1993). The problem here is that, in order to perform the perlocutionary
act, the speaker need not add any gesture or psychological activity to those that she
needs to perform in order to perform the illocutionary act. It is the receiver’s reaction
(psychological or behavioral) that makes the difference. Many philosophers, following
Donald Davidson ([1969] 2001), accept the idea that every action is identical with a
bodily movement and this is obviously incompatible with Austin’s views on action and
his analysis of the speech act. According to a Davidson inspired view of action, the
perlocutionary act cannot exist as a genuine action of the speaker, but only as a way in
which the utterance act can be described in consideration of its consequences.
Obviously, it might correspond to an act of the receiver, which is however itself not the
perlocutionary act in the speech act theoretic sense, but the reaction elicited by the
speaker’s utterance (and therefore, a perlocutionary effect).
One can accept both that, in perlocution, it is the receiver’s reaction that makes the
difference, and that the bringing about of that reaction is an act (or action) of the
speaker, but this commits one to the claim that what is central in the identification of an
action is not the agent’s bodily movement per se, but whether a certain change in a
state of affairs can be traced back to the responsibility of that agent. If one wants to
stick to a philosophy of action tracing actions back to bodily movements, one should
reject the existence of perlocutionary acts altogether.
6. Theoretical implications

What is at stake in the proposals and debates about the locution illocution perlocution
distinction? The most obvious answer is that it is the existence of the intermediate level
of illocution between the act of saying and the consequential effects of a psychological
or behavioral kind it may have on the hearers. That people speak is a trivial truth and so
is that speech may produce effects on its receivers. But is there an intermediate level,
illocution, and what does it consist of? In our overview concerning illocution (Section 4),
we have seen that there are different conceptions of it. Illocution may be viewed as the
production of conventional effects, as the carrying out of a rule governed activity, or as
the expression of a communicative intention and therefore of a psychological attitude of
the speaker. Affirming the existence of illocution as an intermediate level between
saying and achieving perlocutionary effects makes more sense when illocution is viewed
as consisting of full blown actions that bring about effects of their own. When activity
(as opposed to action) is focused upon, effects may be neglected in comparison to ways
to achieve them (verbal behaviour and the associated mental states and attitudes).
These aspects too are relevant to the study of illocution, but with an exclusive focus
upon them, illocution may reduce to an attempt to achieve the psychological or
behavioral effects which are constitutive of perlocution, or if this fallacy is avoided, to
the mere expression of communicative intentions. This may not be enough to qualify
illocution as an autonomous, intermediate level of speech action between locution and
perlocution. So certain developments of the notion of illocution (indeed, some
mainstream ones!) appear to miss, at least in part, the point of having that notion
altogether.
A second problem concerns the absence of any mention of “propositions” in Austin’s
analysis of the speech act. Propositions are held to be essential in the philosophy of
language, whenever one wants to distinguish between sentences and what is expressed
by their occurrences in utterances. Same saying is usually accounted for in terms of the
expression of one and the same proposition by different sentences or different
occurrences of a sentence. Likewise, communication is usually accounted for as the
speaker’s making the addressee grasp the same proposition she has in mind. Since
Austin’s locution illocution perlocution distinction does not mention propositions, it
leaves open whether we should use them in speech act theory. The neglect of
propositions might even be taken to suggest that the conception of communication as
expression and transfer of something expressed (propositions) should be replaced with
a conception of communication as action (the bringing about of changes in the context)
or, more specifically, interaction (the bringing about of such changes in cooperation or
negotiation with other participants). These possible implication of Austin’s outline of
speech act theory were overwhelmed by John Searle’s reintroduction of propositions
and then gradually forgotten. Thus, to accept a proposition–force distinction is to stick
to the received view of linguistic communication, while the locution illocution
perlocution distinction may allow for a neutral (if not critical) standpoint as to the use of
the notion of proposition within the description of a speech act, that is, in a theory of
speech as action.
A third problem concerns truth and correctness. Truth, in Austin’s perspective, is
predicated of “the accomplished utterance” (1975: 140; cf. 1979b: 120), that is, it would
seem, what results from the “total speech act in the total speech situation” (1975: 148)
when the force of this act qualifies it as aimed at saying something true. It is therefore
the specific kind of correctness as regards the world that is typical of statements and
other verdictives of the factual kind (or perhaps of expositives involving a factual
judgement). Austin is not perfectly clear about this,27 but his idea seems to be that we
deem a statement (or similar speech act) true or false only once the utterance has been
taken, or recognized, as an occurrence of that kind of a speech act in conformity to its
conventional rules (1975: 145). Subsequent pragmatics, mainly under Grice’s influence,
has adopted a different view, according to which the truth/falsity judgement and the
judgement concerning the successfulness or felicity of the purported illocutionary act
are independent of one another, so that any indicative utterance may be subject to the
former, even when it is more or less gravely inappropriate or infelicitous (in such cases,
in a Gricean perspective, it is not the case that the utterance is a misfire or other failure
to perform anything illocutionary, it merely triggers implicatures that happen to be
false). The Grice inspired conception fits better a view of saying, or of the locutionary or
the rhetic act, as itself subject to the truth/falsity judgement in abstraction from
illocutionary force and its alleged apparatus of conventional rules and effects. It may
also fit a view of asserting as the expression of a communicative intention, which does
not depend on the observance of conventional rules or conditions. The Austin inspired
conception, while obviously in need of precisation and revisions, refuses to take
indicative utterances as such to be subject to the truth/falsity judgement and may fit a
contextualist conception of truth such as that defended by Charles Travis (see e.g. his
1997, 2000). The former conception may extend to non indicative utterances by sticking
to truth as satisfaction while changing the “direction of fit” of the truth/falsity
judgement (which involves some notion of force, after all, unless one assigns mood to
the rhetic act: see Section 3.1); the latter conception extends to non assertive kinds of
speech acts according to the standards of correctness as regards the world that are
proper to each, which may aim at fairness or righteousness rather than at truth.
I will not discuss any of these problems further here. But it should be clear to the reader
that choices about how to interpret or refine the notions of locution, illocution and
perlocution are not merely technical and philosophically neutral. They have implications
concerning various philosophical issues about language, communication, and social
relationships. Therefore, conversely, one’s preferred philosophical conception of
language or communication or social relationships might turn out to be consistent with
certain interpretations of the locution illocution perlocution distinction, but not with
others.
7. Potentialities of application

Is the distinction between locution, illocution, and perlocution useful in the analysis of
discourse or conversation? Does it help throw light upon the functioning of social groups
and cultures, or upon the dynamics of politics?
I would like to mention at least two contexts of social and political relevance in which
the notions of illocution and perlocution have been employed by leading authors.
The philosopher and sociologist Jurgen Habermas used these notions in his theory of
communicative action (Habermas [1981] 1985: Ch.3) in order to contrast
communicative actions aiming at coming to an understanding in an ideal framework
of freedom and transparency (which he identifies with illocutionary acts) from verbal
actions by which the speaker agent pursues her own ends strategically (which are
instantiated by perlocutionary acts). Judith Butler, a philosopher engaged in political
and gender issues, used the notions of illocution and perlocution under the influence of
Derrida’s reading of Austin to comment upon social and political problems such as how
to deal with racist “hate speech” (Butler 1997). In both these authors, notions such as
illocution and perlocution liberate some political and critical potential, but at the
price of subtle transformations that make them different or even incompatible with
any one of the corresponding notions as defined and discussed in the various trends
of speech act theory. As to Butler, she takes illocution and perlocution as two forms
of performativity, which might be true in a context where a non technical notion of
performativity (directly tied to that of performance) holds, but is quite misleading in
the context of anything like an interpretation of Austin’s distinctions. She also seems
to conceive of illocutionary and perlocutionary acts as embodied in distinct
utterances (while one of the characteristic points of speech act theory is that one
and the same utterance may run locution, illocution, and perlocution).28 Habermas
introduces into the definition of the illocutionary act the idea that illocutionary acts
raise claims related to diverse kinds of validity, namely truth, sincerity, and
normative righteousness. To measure the distance of Habermas’s conception from
standard speech act theoretic views, consider that this amounts to saying that
assertions aim to be judged as true or false while other illocutionary acts aim to be
judged as to the extent to which they satisfy their felicity conditions or their
sincerity. While most speech act theorists took one or more dimensions of success
(all of which distinct from truth or correctness) to be common to all illocutionary act
types, Habermas’s proposal recalls to mind the alleged contrast of constatives and
performatives.29
But what about the locution illocution perlocution distinction in the large world of
research on discourse, conversation, social interaction, cultural diversity of linguistic and
communicative practices? In general, in the fields of discourse analysis, conversation
analysis, and linguistic anthropology, the situation is not exciting.
Discourse analysis has not always noticed the opportunities offered by the locution
illocution perlocution distinction. For example, Critical Discourse Analysis does not count
speech act theoretic notions among its tools, even if it admits that uttering is acting
(Fairclough 1989: 9) and deals, among other matters, with what are in fact matters of
illocution (e.g. assertion, command, offer, speaker’s commitment) and perlocution (e.g.
persuading) (Fairclough 1989, 2003; see also Wodak 2011). Linguistic anthropology,
while recognizing that people do things (and culturally relevant ones) with their
words, finds the tie of illocutionary acts to explicit performatives and conventional rules
too rigid and the redefinition of illocution as communicative intention too individualistic
(cf. Kuipers, this volume; Richland, this volume). Conversation analysts have maintained
that the locution illocutionperlocution distinction and speech act theory in general are
of little use to the aim of understanding conversation as a social and cognitive activity.
For a conversation analyst, for example, what is worth recognition in a
conversational turn such as
(8) Yeah. Well get on your clothes and get out and collect some of that free food and we’ll
make it some other time Judy then (Schegloff 1984: 30)
is not so much the presence of verbs in the imperative mood or the alleged directive
aim, but that it acts in the conversational sequence as the closing initiation of the phone
call in which it occurs (the interlocutor replies “Okay then Jack” and then greetings
follow). But it should be clear that positive as well as negative evaluations of the
applicative potential of notions such as locution, illocution, and perlocution largely
depend on the specific way in which these notions are defined and interpreted. Van Rees
(1992) replied to Schegloff (1984) that (8) is a directive expressing the speaker’s wish
that the interlocutor start getting dressed for her dinner party, which amounts to a
positive attitude towards her and her well being. According to van Rees, this explains
why that turn can also initiate a closing sequence. But this is clearly enough a weak
defense. Whatever the theoretical significance of explanations of verbal behaviour in
terms of intentional states, are we as analysts really interested in what the speaker of
(8) is expressing in his conversational turn? Is it so sure that he is expressing only and
exactly a wish that his interlocutor start getting dressed for her dinner party, and
therefore an interest in her well being? And even if it were so, does this tell us anything
illuminating about what is happening between the two participants, what they are doing
to each other? There might be more to say about a conversational turn like this and the
sequence containing it, but to do so, we have to bring into the picture certain
“conventional” aspects which are typical of illocution, what von Savigny (1988: 45–51)
called the “conventional make up” of a situation or what I call the deontic modal
competence of the participants. By issuing an exercitive (such are commands and acts of
exhortation), the speaker makes it possible for the addressee to close the conversation
without violating what might be his right to keep on talking to her. Rather, she now
owes to him to close the conversation, so when she does so he can feel he has still not
lost some kind of influential position upon her which he is possibly interested in
occupying. By the same token, he acquires the obligation not to hinder her from closing
the conversation. This explains why the turn can be a closing initiation much better than
is done by resort to the expressed intentional states: the turn’s expressive value appears
now as something epiphenomenal with respect to an underlying game of rights and
obligations. We have therefore seen that the replies to our initial questions in this
Section depend largely on the specific way in which the notions of locution, illocution
and perlocution are interpreted. Moreover, both the applications of the locution
illocution perlocution distinction and the refusals to apply them may be found to express
misconceptions of the very notions at issue or, at least, partial perspectives on their
potentialities. Therefore, in attempting to apply (or refusing to apply) speech act
theoretic notions to conversation, discourse, culture and society, we should first decide
which definitions of locution, illocution, and perlocution we are going to pick up, and
confine our subsequent verdict to these.
In this vein, let me conclude by suggesting that, under the definition of illocution as
bringing about conventional effects that can be traced back to Austin, the notion of
illocution (within the locution illocution perlocution distinction) can provide a starting
point or perhaps a unifying background to reflections and hypotheses about how
discourse structures interpersonal and social relationships, how conversational turns
relate to one another, and how language participates in or contributes to the life of a
culture. An Austin inspired apparatus comprising socially accepted procedures (named
by illocutionary verbs), illocutionary force indicators shaping the physiognomy of
utterances, sequentially manifest uptake and conventional (defeasible) effects can be of
use in the analysis of what a conversation achieves, what a stretch of discourse
proposes or imposes on its audience, how the rules of a culture are enacted, exploited
and modified. Unfortunately, this analytical potential has so far not been developed
systematically enough in the direction of a stable methodology. But historical distance
from the main theoretical proposals about the locution illocution perlocution distinction
might now help to understand which lessons are to be taken from which of them, and
how some lessons from the original Austinian proposal can be put to work effectively.

Acknowledgements

I am grateful to many colleagues and friends for exchanges and discussions on the topics
I am dealing with here. Above all, I am indebted to Andreas Kemmerling, Paolo Leonardi,
and Ken Turner for their remarks on a draft of this chapter.
Notes

1. Loqui (from which locution) is the Latin word for ‘to speak’.
2. A similar claim is made (in other terms) by Geis (1995: 9).
3. Austin uses the word aetiolation metaphorically, taking it from botanics where it
refers to the modifications in the growth of plants that are induced by keeping them
away from sunlight.
4. Urmson (1977) raises the opposite claim that performatives and illocutionary acts
belong to two different kinds: conventional acts and speech acts respectively.
5. For discussion of the securing of uptake, see Petrus (2006); Sbisà (2009).
6. This explains why, as Holdcroft puts it (1978: 20), the perlocutionary act “can be
redescribed as the performance of an illocutionary act with certain consequences”. It is
indeed the same agent (the speaker) who in issuing one and the same utterance both
achieves conventional effects at the illocutionary level and produces further
consequential effects at the perlocutionary level.
7. Another problem facing contextualism, which has received no satisfactory treatment
up to now, is how much context affects what is said and how much it affects the
evaluation (in terms of truth and falsity) of its assertion.
8. For the imperative sentence type, for instance, Bach and Harnish propose: “S is saying
that! (… p…)” as equivalent to “S is saying that H is to make it the case that (… p…)”
(1979: 25).
9. Use theories of meaning partly influenced by Dummett’s philosophy have been
developed by Dag Prawitz and Cesare Cozzo (see Cozzo 1994; Prawitz 2006).
10. Another use theory of meaning, paying great attention to illocution, was put forward
by Eike von Savigny (1988).
11. A speech act theoretic approach to meaning, aimed at remedying the shortcomings
of use theories, has been proposed by Barker (2004). At first sight, it may appear to be a
further attempt to explain meaning in terms of force, since one of its key explanatory
concepts is that of “proto-illocutionary act”. This is, however, a merely verbal matter,
because what the author is speaking of is the joint expression (or “advertising”) of two
intentions, the former “representational”, the latter “communicative”, where the
former is not reducible to the latter. But to give substance to his notion of
representation (and moreover, to account for compositionality, one of the hardest
points in developing a usetheory of meaning), Barker resorts to substantive assumptions
concerning the psychological nature of semantics, which are quite heterogeneous with
respect to the speech act theoretic framework.
12. See also Searle (1969: 18): “Austin baptized these complete speech acts with the
name ‘illocutionary acts’” [my emphasis].
13. That illocution is liable to infelicity is further confirmed by the fact that Austin takes
the liability of statements to infelicity as evidence of their illocutionary nature (1975:
137–139).
14. This aspect of Searle’s felicity conditions is further discussed, in its relationship to
promising, by Ambroise, this volume.
15. The relationship of illocutionary forces to linguistic devices and structures was a hot
topic in the 1970s in linguistic approaches to speech acts (cf. e.g. Sadock 1974). For the
debate on the so called “performative hypothesis” (an attempt to trace illocutionary
force back to performative verbs in the deep structure of sentences), see Holdcroft
(1978), Levinson (1983: 246 263), and Doerge, this volume, Section 7.3.
16. The image of the double direction of fit is used by Searle (1989) to account not only
for institutional illocutionary acts, but also for explicit performatives: see Kissine, this
volume, Section 3.1 and Doerge, this volume, Section 5.4.
17. Another approach connecting speech acts to logic is that of Nicholas Asher and Alex
Lascarides (2003). But these authors are not so much concerned with speech actions, as
with dynamic semantics as applied to discourse.
18. As to Searle, it might be interesting to recall that in his philosophy of mind he
affirmed the primacy of intentional states over speech acts, claiming that language
is derived from intentionality and not conversely (Searle 1983: 5). He also suggested
that the basic categories into which (according to his own theory) illocutionary acts fall,
derive from some fundamental features of the mind, whose intentionality not only
creates the possibility of meaning, but also limits its forms (1983: 166). It is difficult to
avoid the conclusion that his view of speech acts is far more permeated by
intentionalism than might seem at first sight.
19. Schiffer (1972) raises also important objections to the completeness of Grice’s
analysis of the notion of meaning (see Kemmerling, this volume).
20. I have tried to distinguish two different uses of the notion of successfulness as
regards illocutionary acts. I do not mention sincerity (or other forms of non
defectiveness), because neither the notion of succeeding in performing an illocutionary
act, in Austin and Searle, nor the notion of communicative successfulness in Bach and
Harnish include them. There is general agreement, it seems, to the effect that an
insincere illocutionary act is not therefore null and void, and that insincerely performed
illocutionary acts can be communicatively successful.
21. The debate on pornography has also dealt with the extent to which illocution is
reshaped by context (see e.g. Saul 2006, Bianchi 2008) and the role of presupposition or
conversational scorekeeping (Langton and West 1999).
22. Among these phenomena, I also include the temporal (or “narrative”) development
of verbal interactions (Sbisà 2002).
23. A seemingly analogous acknowledgement is made by Searle (1995, 2010). But Searle,
while recognizing the role of language in bringing about institutional facts, endows with
this task solely the class of “declarative” illocutionary acts (whether performed by the
aid of a performative verb in the first person present indicative active, or by means of ad
hoc conventional formulas). Moreover, his attention focuses on deontic states in
institutional contexts. My claim is broader. It is that relationships among participants in
social and interactional situations always involve attribution and detention or loss of
states of power, duty and knowledge, or, to put it differently, states that can be
described by means of modal deontic predicates such as can and ought to and their
negations, all of which are grounded in illocutionary uptake.
24. Cohen’s distinctions, albeit more articulated, are similar to Austin’s distinction
between perlocutionary object and perlocutionary sequel (see above, 2.3). It should be
noted that Cohen, writing in 1973, could not refer to the distinction between
perlocutionary objects and sequels in the form in which it appears in the second edition
of Austin’s How to Do Things with Words (1975).
25. Petrus (2010), however, focuses on perillocutionary acts that are overt but
strategically indirect.
26. Geis (1995) is also critical of the mapping from individual utterances to illocutionary
forces and argues for a more complex mapping of elements of interactional structures
onto utterances.
27. Pages 142 145 of Austin (1975) comprise passages edited by J.O. Urmson in 1962
from students’ notes.
28. It is a merit of Butler (1997) that she gives an interesting reading of Derrida’s notion
of iterability, in which iterability does not undermine the suitability of speech to be
action, but contributes to it.
29. On both Habermas and Butler, see Leezenberg, this volume.

References

Alston, William P.
1964 Philosophy of Language. Englewood Cliffs, N.J.: Prentice Hall.
Alston, William P.
2000 Illocutionary Acts and Sentence Meaning. Ithaca: Cornell University Press.
Asher, Nicholas and Alex Lascarides
2003 Logics of Conversation. Cambridge: Cambridge University Press.
Austin, John L.
1962 Sense and Sensibilia. Oxford: Oxford University Press.
Austin, John L.
1970 [1949] Intelligent behaviour. A critical review of The Concept of Mind. In: Oscar
P. Wood and George Pitcher (eds.), Ryle. A Collection of Critical Essays, 45–51. Garden
City, NY: Doubleday.
Austin, John L.
1975 How to Do Things with Words. The William James Lectures delivered at Harvard
University in 1955. Edited by James O. Urmson and Marina Sbisà. Oxford: Oxford
University Press. [First edition (edited by James O. Urmson). Oxford: Oxford University
Press. 1962.]
Austin, John L.
1979a Philosophical Papers. 3rd edition. Oxford: Oxford University Press.
Austin, John L.
1979b Truth. In: John L. Austin 1979a, 117–133. [First published in: Proceedings of the
Aristotelian Society Suppl. Vol. 24: 111–129. 1950.]
Austin, John L.
1979c How to talk. Some simple ways. In: John L. Austin 1979a, 134–153. [First
published in: Proceedings of the Aristotelian Society 53: 227–246. 1953.]
Bach, Kent
1994 Semantic slack: What is said and more. In: Savas L. Tsohatzidis (ed.), Foun dations
of Speech Act Theory: Philosophical and Linguistic Perspectives, 267–291. London:
Routledge.
Bach, Kent and Robert M. Harnish
1979 Linguistic Communication and Speech Acts. Cambridge, Mass.: Harvard Uni versity
Press.
Bach, Kent and Robert M. Harnish
1992 How performatives really work, Linguistics and Philosophy 15: 93–10.
Barker, Stephen J.
2004 Renewing Meaning: A Speech Act Theoretic Approach. Oxford: Oxford Uni versity
Press.
Bianchi, Claudia
2008 Indexicals, Speech Acts and Pornography. Analysis 68 (300): 310–316.
Blum Kulka, Shoshana, Juliane House and Gabriele Kasper (eds.)
1989 Cross cultural Pragmatics: Requests and Apologies. Ablex: Norwood.
Brandom, Robert
1994 Making It Explicit: Reasoning, Representing, and Discursive Commitment.
Cambridge, Mass.: Harvard University Press.
Brandom, Robert
2000 Articulating Reasons: An Introduction to Inferentialism. Cambridge, Mass.: Harvard
University Press.
Brown, Penelope and Stephen Levinson
1987 Politeness: Some Universals in Language Usage. Cambridge: Cambridge University
Press.
Butler, Judith
1997 Excitable Speech: A Politics of the Performative. London: Routledge.
Carston, Robyn
2002 Thoughts and Utterances: The Pragmatics of Explicit Communication. Oxford:
Blackwell.
Cohen, Ted
1973 Illocutions and perlocutions. Foundations of Language 9: 492–503.
Cozzo, Cesare
1994 Meaning and Argument: A Theory of Meaning Centred on Immediate Argu mental
Role. Stockholm: Almqvist & Wiksell.
Davidson, Donald
2001 [1969] The logical form of action sentences. In Donald Davidson, Essays on Actions
and Events. 2nd edition. Oxford: Oxford University Press.
Davis, Steven
1979 Perlocutions. In: J. R. Searle, M. Bierwisch and F. Kiefer (eds.), Speech Act Theory
and Pragmatics, 37–55. Dordrecht: Reidel.
Derrida, Jacques
1972 Signature événement contexte. In: Jaques Derrida, Marges de la Philosophie, 365–
393. Paris: Minuit.
Dummett, Michael A.E.
1975 What is a theory of meaning? In: Samuel Guttenplan (ed.), Mind and Lan guage,
97–138. Oxford: Oxford University Press.
Fairclough, Norman
1989 Language and Power. London/ New York: Longman.
Fairclough, Norman
2003 Analysing Discourse: Textual Analysis for Social Research. London/ New York:
Routledge.
Feinberg, Joel
1964 Action and responsibility. In: Max Black (ed.), Philosophy in America, 134–160.
Ithaca: Cornell University Press.
Frege, Gottlob
1956 The thought: A logical inquiry. English translation in: Mind 65: 289–311. [Original
edition: Der Gedanke. Beiträge zur Philosophie des Deutschen Idealismus. 1918.]
Gauker, Christopher
1998 What is a context of utterance? Philosophical Studies 9: 149–172.
Geis, Michael L.
1995 Speech Acts and Conversational Interaction. Cambridge: Cambridge Univer sity
Press.
Grice, H. Paul
1989a Studies in the Way of Words. Cambridge, Mass.: Harvard University Press.
Grice, H. Paul
1989b Meaning. In: H. Paul Grice 1989a, 213–223. [First published in: The Philosophical
Review 64: 377–388. 1957.]
Grice, H. Paul
1989c Utterer’s meaning, sentence meaning and word meaning. In: H. Paul Grice 1989a,
117–137. [First published in: Foundations of Language 4: 225–242. 1968.]
Grice, H. Paul
1989d Utterer’s meaning and intentions. In: H. Paul Grice 1989a, 86–116. [First
published in: Philosophical Review 68: 147–177. 1969.]
Grice, H. Paul
1989e Logic and conversation. Chapter 2. In: H. Paul Grice 1989a, 22–40. [First published
in: Peter Cole and Jerry L. Morgan (eds.), Syntax and Semantics 3: Speech Acts, 41–58.
New York: Academic Press. 1975.]
Gu, Yueguo
1993 The impasse of perlocution. Journal of Pragmatics 20: 405–432.
Habermas, Jürgen
1985 The Theory of Communicative Action. Vol. 1. English transl. Boston: Beacon Press.
[Original edition: Theorie des kommunikativen Handelns. Vol.1. Frankfurt: Suhrkamp.
1981.]
Harnish, Robert M.
2009 Internalism and externalism in speech act theory. Lodz Papers in Pragmatics 5: 9–
31.
Holdcroft, David
1978 Words and Deeds. Oxford: Oxford University Press.
Hornsby, Jennifer
1988 Things done with words. In: Jonathan Dancy, J.M.E. Moravcsik and C.C.W. Taylor
(eds.), Human agency. Language, duty and value, 27–46. Stanford: Stanford University
Press.
Hornsby, Jennifer and Langton, Rae
1998 Free speech and illocution. Legal theory 4: 21–37.
Kaplan, David
1989 Demonstratives: An essay on the semantics, logic, metaphysics, and episte mology
of demonstratives and other indexicals. In: Joseph Almog, John Perry, and Howard
Wettstein (eds.), Themes from Kaplan, 481–563. Oxford: Oxford University Press.
Kemmerling, Andreas
2001 Gricy actions. In: Giovanna Cosenza (ed.), Paul Grice’s Heritage, 69–95.
Turnhout: Brepols.
Langton, Rae
1993 Speech acts and unspeakable acts. Philosophy and Public Affairs 22: 293–330.
Langton, Rae and Caroline West
1999 Scorekeeping in a pornographic language game, Australasian Journal of Philosophy
77: 303–19.
Levinson, Stephen
1983 Pragmatics. Cambridge: Cambridge University Press.
Levinson, Stephen
2000 Presumptive Meanings. Cambridge: Cambridge University Press.
Locher, Miriam A. and Sage L. Graham (eds.)
2011 Interpersonal Pragmatics. (Handbooks of Pragmatics 6.) Berlin: de Gruyter.
MacKinnon, Catharine
1987 Feminism Unmodified. Cambridge, Mass.: Harvard University Press.
McGowan, Mary Kate
2004 Conversational exercitives: Something else we do with our words, Linguistics and
Philosophy 27: 93–111.
Millikan, Ruth
1998 Proper function and convention in speech acts. In: Lewis Edwin Kahn (ed.),
The Philosophy of P. F. Strawson, 25 43. Chicago, Ill.: Open Court.
Petrus, Klaus
2006 Illokution und Konvention, oder auch: was steckt nun wirklich hinter Austins
“securing of uptake”? Grazer Philosophische Studien 70: 101–126.
Petrus, Klaus
2010 Illocution, perillocution, and communication. In: Klaus Petrus (ed.), Meaning and
Analysis: New Essays on Grice, 220–234. Houndmills, Basingstoke: Palgrave.
Prawitz, Dag
2006 Meaning approached via proofs. Synthese 148: 507 524.
Recanati, François
2004 Literal Meaning. Cambridge: Cambridge University Press.
Recanati, François
2010 Truth Conditional Pragmatics. Oxford: Oxford University Press.
Sadock, Jerrold M.
1974 Toward a Linguistic Theory of Speech Acts. New York: Academic Press.
Saul, Jennifer
2006 Pornography, speech acts and context. Proceedings of the Aristotelian Society
106: 227–246.
Savigny, Eike von
1988 The Social Foundations of Meaning. Berlin/Heidelberg: Springer.
Sbisà, Marina
1984 On illocutionary types. Journal of Pragmatics 8: 93–112.
Sbisà, Marina
1989 Linguaggio, Ragione, Interazione. Bologna: Il Mulino.
Sbisà, Marina
2001 Illocutionary force and degrees of strength in language use. Journal of Prag matics
33: 1791–1814.
Sbisà, Marina
2002 Cognition and narrativity in speech act sequences. In: Anita Fetzer and Chris tiane
Meierkord (eds.), Rethinking Sequentiality, 71–97. Amsterdam: John Benjamins.
Sbisà, Marina
2006 Communicating citizenship in verbal interaction: principles of a speech act oriented
discourse analysis. In: Heiko Hausendorf, Alfons Bora (eds.), Analysing Citizenship Talk,
151–180. Amsterdam: John Benjamins.
Sbisà, Marina
2007 How to read Austin. Pragmatics 17: 461–473.
Sbisà, Marina
2009 Uptake and conventionality in illocution. Lodz Papers in Pragmatics 5: 33–52.
Schegloff, Emanuel
1984 On some questions and ambiguities in conversation. In: J. Maxwell Atkinson and
John Heritage (eds.), Structures of Social Action: Studies in Conversation Analysis, 28–52.
Cambridge: Cambridge University Press.
Schiffer, Stephen
1972 Meaning. Oxford: Oxford University Press.
Searle, John R.
1969 Speech Acts: An Essay in the Philosophy of Language. Cambridge: Cambridge
University Press.
Searle, John R.
1979a Expression and Meaning: Studies in the Theory of Speech Acts. Cambridge:
Cambridge University Press.
Searle, John R.
1979b A taxonomy of illocutionary acts. In: John R. Searle 1979a, 1–29. [First pub lished
in: K. Gunderson (ed.), Language, Mind and Knowledge, 344–369. Minneapolis:
University of Minnesota Press. 1975.]
Searle, John R.
1979c Indirect speech acts. In: John R. Searle 1979a, 30–57. [First published in: Peter
Cole and Jerry L. Morgan (eds.), Syntax and Semantics 3: Speech Acts, 59–82. New York:
Academic Press 1975.]
Searle, John R.
1983 Intentionality: An Essay in the Philosophy of Mind. Cambridge: Cambridge
University Press.
Searle, John R.
1989 How performatives work. Linguistic and Philosophy 12: 535–558.
Searle, John R.
1995 The Construction of Social Reality. New York: The Free Press.
Searle, John R.
2010 Making the Social World: The Structure of Human Civilization. Oxford: Ox ford
University Press.
Searle, John R. and Daniel Vanderveken
1985 Foundations of Illocutionary Logic. Cambridge: Cambridge UniversityPress.
Sperber, Dan and Deirdre Wilson
1995 Relevance: Communication and Cognition. Second edition. Oxford: Blackwell. [First
edition 1986.]
Strawson, Peter F.
1969 Intention and convention in speech acts. In: K.T. Fann (ed.), Symposium on Austin,
380–400. London: Routledge. [First published in: Philosophical Review 73: 439–460.
1964.]
Urmson, James O.
1977 Performative utterances. Midwest Studies in Philosophy 2: 120–127.
Travis, Charles
1997 Pragmatics. In: Bob Hale and Crispin Wright (eds.), A Companion to the Philosophy
of Language, 87–107. Oxford: Blackwell.
Travis, Charles
2000 Unshadowed Thought. Cambridge, Mass.: Harvard University Press. van Rees, M.
Agnes
1992 The adequacy of speech act theory for explaining conversational phenomena: A
response to some conversation analytical critics. Journal of Pragmatics 17: 31–47.
Wittgenstein, Ludwig
1953 Philosophische Untersuchungen/ Philosophical Investigations. Ed. by Eliza beth
Anscombe and Rush Rhees. Oxford: Blackwell.
Wodak, Ruth
2011 Critical Discourse Analysis: Overview, challenges, and perspectives. In: Gisle
Andersen and Karin Aijmer (eds.), Pragmatics of Society, 627 650. (Handbooks of
Pragmatics 5.) Berlin: de Gruyter Mouton.

View publication stats