.-~,~..‘,M."‘ ”

\.:\.“..
' ~~ \.‘.._
u,

':’lt’.

'ii

'--<-,’_'(_‘.[__

' I

"‘~..
‘0‘
v...,

 

     
   

  

LIB?“ AR Y .
Michigan Spam
University

This is to certify that the

thesis entitled

THE THEORY OF MEANING AS INTENTION

presented by

Peter Arthur Facione

has been accepted towards fulfillment
of the requirements for

JILL—degree in lbilasaphy

Major professor

 

Date M

0-7639

— f 11
fi‘ a ’mp‘-

Margy 3"“

 

 

 

 

 

ABSTRACT

THE THEORY OF MEANING AS INTENTION

BY

Peter Arthur Facione

This theory of meaning was originated by H. Paul
Grice and Henry S. Leonard simultaneously and independently.
These philosophers propose to analyze meaning in terms of
the intentions or purposes of speakers. There are two
major issues that such a theory of meaning must resolve.
First, it must analyze a person's meaning something, what I

call "meaninglf in terms of the intentions or purposes of

that person. Second, it must reduce or relate utterance or

inscription meaning, which I call "meaningz, to meaningl.
Several distinctions are presented in the first
chapter which aid in determining exactly what the problems
are that this theory must resolve, and exactly what the
objections are that other philosophers have raised against
this theory. Besides the meaningl/meaning2 distinction,
there is a survey of several of the well known senses of
'mean'. There is a discussion of the illocutionary/perlo—
cutionary distinction so as to distinguish intentions,

forces, results and responses of these two kinds. Both

Grice's and Leonard's programs are outlined and some

Peter Arthur Facione

attention is given to how each man would account for the

conventional meaning2 of utterance-types in terms of

 

meaningl. Several prima facie objections to this theory
are discussed, but these turn out to be without merit given
the distinctions made earlier.

The main attention of this work is on the first
issue, and especially Grice's attempts to provide an
adequate analysis of meaningl. In Chapter II Grice's 1957
article "Meaning" is outlined and the debate it engendered
is discussed. This chapter aims at recording the various
objections to this theory, and especially this analysis.
The objections are summarized and evaluated in this chapter.

Grice's reply to his critics is recorded and dis—
cussed in Chapter III. His analysis is traced through its
encounter with several counter-examples, including some new
ones. The analysis is found to be both too weak and too
strong. But this matter is overshadowed by arguments that
Grice's analysis is irreparably incomplete and misguided.
It is not possible to account for a speaker's illocutionary
acts, those with which we should wish to associate meaningl,
or his illocutionary intentions given an analysis like
Grice's. Such an analysis seeks to explicate meaningl in
terms of the speaker's perlocutionary intentions. Further,
a theory of meaningl that restricts itself to listing only
intentions in its analysans must fail. Such a theory

neglects the fact that intentions alone do not always

 

Peter Arthur Facione

suffice for the performance of an illocutionary act, whether
in this act communication is to take place or not.
Chapter III culminates in an analysis of meaningl

for those cases when communication is involved:

U utters x at time t meaning to * A ...

 

if and only if there is some f, g, and 2; (later

or the same as E) such that:

g utters x at time E

(l) believing that A would think, at EL, that
x has E which E correlates to *-ing ...

(2) intending to * A ... in uttering x

(3) intending that E think, at EL, by virtue of
believing that x has E which E correlates to
*-ing ..., that U intended to * A ... in
uttering x at E —

Explanation of notation:

'U'--variable for speakers

'x'--variable for utterance-tokens

't'--variable for moments of time (as is 't")

'*'--variable for illocutionary forces

'f'--variable for features of utterance-tokens

'c'--variable for modes of correlating the values
of 'f' with the values of '*'

'A'--schematic letter (not a variable) to be
replaced by a specification of 2's conception
of the audience he intends to communicate with

'...'--schematic device to be replaced by a specifi-
cation of the propositional content of x_where
appropriate

Chapter III also provides an analysis of the notion of a
person meaningl what he says.

The appendix to this work provides biographical
and bibliographical data on the two philosophers, Henry

Leonard and Herbert Grice.

 

THE THEORY OF MEANING AS INTENTION

BY

Peter Arthur Facione

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Philosophy

1971

 

@ Copyright by
PETER ARTHUR FACIONE

1971

 

For Noreen

ii

 

ACKNOWLEDGMENTS

I wish to express my deep gratitude and admiration
to Professor Herbert Hendry of the Department of Philosophy
of Michigan State University for his wise guidance and
continued support. I also wish to thank Professors William
Callaghan, Rhoda Kotzin, Richard Hall, and Herbert Bohnert
as well as all those people, both faculty and students, who

by their aid and encouragement helped me in this work.

iii

Chapter

TABLE OF CONTENTS

I. TWO SENSES OF MEANING . .

Introduction
Distinctions
The Theory .
Grice's Program . . .
Leonard's Program . .
Difficulties
Footnotes to Chapter I .

II. GRICE'S

"MEANING" . . .

Introduction
"Meaning"

Group
Group
Group
Group
Group
Group

One .
Two .
Three
Four .
Five .
Six .

Summary .
Footnotes to

III. THE DEFINITION

Introduction
Senses of Meaningé n1 .
Defining Meaning .
Preliminaries . . .
Redefinition I . . .

Redefinition II,
Redefinition II,
Redefinition III,
Redefinition III,

Chapter II

OF MEANING

Defining Meaning4, II .
Redefinition IV . .
Redefinition V . . .

Problems. .

A Definition of Meaningl

Conclusion .

Footnotes to Chapter III

iv

Version A .
Version B

Version A
Version B

Page

17
20
35
47
6O

66

66
7O
81
83
86
92
96
111
119
124

129

129
131
135
135
141
146
150
159
168
173
174
180
184
197
203
205

Chapter Page
BIBLIOGRAPHY . . . . . . . . . . . . . . 211

APPENDIX . . . . . . . . . . . . . . . 216

 

Of the many other consequences of the view . . . I
will say nothing. I will only beg the reader not to make
up his mind against the view--as he might be tempted to do,
on account of its apparently excessive complication-~until
he has attempted to construct a theory of his own on the
subject. . . . This attempt, I believe, will convince him
that, whatever the true theory may be, it cannot have such
a simplicity as one might have expected beforehand.

Lord Russell
--On Denoting--

vi

CHAPTER I

TWO SENSES OF MEANING

Introduction

 

A survey of the philosophical terrain reveals a
significant concern in recent times with the philosophy of
language and theories of meaning. Philosophers have found
such theories interesting both in their own right and also
as clues to the resolution of problems in other aspects of
philosophy.1 In recent years several theories of meaning
have been presented, elaborated, examined, criticized,
rethought, recriticized, and laid to rest. The theory of
meaning as intention or purpose has not undergone so com-
plete an examination as yet. Generally anthologies and
studies in the philosophy of language merely mention this
theory, often classifying it in misleading ways. William
P. Alston calls the theory of meaning as intention a modern
refinement of the "ideational" theory.2 G. H. R. Parkinson
calls it a "causal" theory.3 Thomas Olshewsky, however,

devotes a section of his anthology, Problems in the Philos-

 

ophy of Language to this theory and to the relationship

 

between meaning and the intentions or purposes of speakers.4
I shall examine that theory of meaning that proposes

to explicate a speaker's meaning something by listing only

intentions or purposes of the speaker. The intentions in
such a list would be jointly sufficient and individually
necessary for meaningful speaking.

Given an adequate explication of what might be
called "speaker's meaning" this theory would proceed to
explicate other aspects of "standard meaning" or meaning as
applied to elements of a language. I shall concentrate my
attention on the issue of the definition or explication of
meaning in terms of intentions. My primary aim is to show
that such a definition encounters insurmountable problems
and is, thus, a misadventure. My secondary aim is to offer
an adequate explication of "speaker's meaning." In this
chapter I shall offer preliminary distinctions and assump-
tions as well as brief sketches of two versions of this
theory. In the next chapter I shall provide a commentary
on the history of H. P. Grice's first definition of
"speaker's meaning." In the final chapter the adequacy of
this theory will be challenged and an analysis of "speaker's
meaning" suggested. The analysis will be in the spirit of
the many explications discussed in this work, but it is not
bound by the requirement to list only intentions of the
speaker in its analysans.

The earliest published versions of the theory of
meaning as intention or purpose, discounting mimeographed
texts and book reviews, were available in 1957. Thus this

theory competed with the work of Wittgenstein and Austin for

the attention of philosophers. These two luminaries were
not to be overshadowed by the respected but not widely
published H. P. Grice, nor by the lesser known H. S.
Leonard. Though this theory received little attention at
first, it has engendered an interesting debate that ranges
through a number of years and journals. In the last few
years several philosophers have discussed this theory.
Alston tried to indicate some problems with it; P. F.
Strawson tried to reviseéfit. Paul Ziff and N. L. Wilson
tried to destroy it; T. E. Patton and D. W. Stampe tried to
defend it.5 Max Black, regarding Grice's and Leonard's
views as highly important, is preparing a criticism of

their position.6 John Searle, in his recent Speech Acts:

 

An Essay in the Philosophy of Language, directs his atten-

 

tion to this theory trying to both criticize it and borrow
its insights.7 D. S. Clarke also borrows from the work of
Grice.8

Many of these philosophers have failed to notice a
great part of the original primary source literature pub-
lished by the two authors of this theory. Apparently
everyone except Clarke and Wilson based their criticism on
the single, rather incomplete, article "Meaning" that Grice

published in Philosophical Review in 1957. They all

 

neglected to examine Leonard's version of the theory which

was presented in Principles of Right Reason in 1957, as

 

well as in two articles, published in 1959 issues of

Philosophy of Science, "Interrogatives, Imperatives, Truth,

 

Falsity, and Lies," and "Authorship and Purpose." Grice's
earlier article, the center of over a decade of contro-
versy, has been supplanted by two more recent articles,
"Utterer's Meaning and Intention" which Grice published in

Foundations of Language in 1968, and "Utterer's Meaning,

 

Sentence—Meaning, and Word-Meaning," published in Philo-

sophical Review in 1969. Leonard also published one final

 

article on his version of the theory, "Authorship of Signs";

this appeared in the 1960 issue of the Papers of the

 

Michigan Academy of Science, Arts and Letters. Biographies

 

of Leonard and Grice, as well as lists of their publica—
tions, are in the appendix.

Assumptions, distinctions, and working definitions
are presented below. A section offering a rough outline of
this theory of meaning and the issues involved in its
presentation is followed by sections outlining Grice's and
Leonard's versions of the theory. The final section

examines some prima facie difficulties with the theory.

 

Distinctions

 

On February 20, 1971 teletype machines throughout
the nation printed out the message that the United States
was in immediate danger of nuclear attack. The message
originated from the Oklahoma offices of the emergency
defense warning system. Upon receiving this message, which

did contain the authenticating codeword "hatefulness," radio

stations were to follow a predetermined civil defense plan.
Most did not. Although the broadcasters knew what the
message meant, they did not believe that its authors in
Oklahoma intended to deliver that particular message to
them. They inferred that the message was a mistake, that
its authors did not mean it. Other avenues of information
did not corroborate the warning. The station operators
noticed that the message was received at a particular time
on a Saturday morning, a day and time normally used for

test messages. These operators made, in practice, the
distinction between "speaker's meaning" and "inscription
meaning." Let 'meaningl' designate the former and 'meaningz'
the latter. Meaningl seems to be related in some indefinite
way with the intentions of speakers or authors. Meaning2
seems to be associated with the standard (conventional,
literal) meaning of the particular inscription, gesture, or
signal the author uses.

I shall assume that the meaningl/meaning2 distinction
reveals a genuine ambiguity. Thus, 'mean' in these respects
is not to be thought of as a generic term subsuming meaningl
and meaningz. If it were, the distinction between what a
person meansl and what an utterance—type, §, or utterance-
token, x, means2 would become simply the difference between
speakers and utterances.lo How do these two senses relate

to the many other senses of 'mean'?

'Mean' has several senses, some of which are
clearly not under consideration. According to the Oxford

English Dictionary the adjective 'mean' means "common or

 

inferior." The noun has two meanings, "that which is in
the middle" and "a lament." But it is the verb which is
our main concern.

The verb 'mean' has four general senses. It means
(i) "to mediate or moderate"; (ii) "to lament, pity, or
complain of"; (iii) "to occupy the middle"; and (iv) "to
intend," "to signify." There are seven senses of this

fourth general sense of 'mean' listed in the Oxford English

 

Dictionary. Some of these seven have still more minute

 

distinctions made. The first three of the seven concern us.
As a transitive verb 'mean' means, (a) "to have in
mind as a purpose or intention; to propose, design."ll

Also as a transitive verb it can mean, (b.1) "to intend to

indicate (a certain object), or to convey (a certain sense)

when using some word, sentence, significant action, etc."ll
Sense (b.1) is exemplified in:

(1) What did John mean when he said "Sleep is fun"?

(2) The document does not mean literally what it says.11

There is another, slightly different, b-sense of 'mean'.
Sense (b.2) is revealed in a "question of the form what does
(a person) mean (by certain conduct)?"11 The response to

this question is, normally, a revelation of the person's

"motive or justification." The third of the seven specific

senses of 'mean', (c), is "of things, words, statements; to
have a certain signification; to signify, or import, or

portend."ll Such a sense is found in

(3) But sayi what mean those coloured streakes in
Heavn.

Clearly, Grice and Leonard are justified in thinking that a
theory of meaning should take into account the intentions
or purposes of speakers.

One might think of meaningl as sense (a). Likewise
it is possible to regard sense (c) as the sense of meaningz.
However, as the examples indicate, both meaningl and
meaning2 can be discerned in sense (b.1). The reference to
the speaker's intention in the definiens suits meaningl
well. But note that the speaker intending to convey a
certain sense must use a "significant" action. This suggests
that the intention alone is not sufficient for successfully
meaningl something. By the end of Chapter III it will be
evident that this suggestion is true. Note also, that
example (2) does not fit the definition, since documents
cannot intend. This example, which is the dictionary's,
employs 'mean' in the sense of meaningz. Perhaps an example
that reveals both of the senses of 'mean' discernible in

(b.1) is needed.

(4) The authors of the document did not meanl what the
document literally meansz.

Other ambiguities offer slight difficulties. The
process-product ambiguity of 'utterance' should be little
trouble to the careful. The term 'say' is ambiguous in a
slightly different way. We can, for the sake of clarity
distinguish three senses of 'say'. One sense of 'say' is
roughly synonymous with assert. This is, perhaps, the
normal use of the word. To "say" something, in this sense
of the word, is to perform some illocutionary act. Thus,
we can, where necessary, denote this sense of 'say' by
using 'say.'. A necessary condition for sayingi something

1

seems to be that the speaker utter something which he

 

believes to be a sign. The speaker must, that is, "say"

 

something or perform a locutionary act. Where necessary we
shall use 'sayl' to denote this sense of 'say'. Notice
that a person might sayl something but not sayi it; as, for
example, a technician might do in testing a sound system.
Thus, although the speaker believes that what he saidl is a
sign, he may not be using it as a sign.

In Austin's How to Do Things with Words we find that

 

"to perform a locutionary act is in general, we may say,

"12 I

also and 39 ipsg to perform an illocutionary act.
believe that my technician example is a genuine exception to
Austin's general but not universal rule. However, the
exception may turn on my characterization of the illocu-

tionary/locutionary distinction, which is slightly different

than Austin's.

I am not sure whether Austin would be sympathetic
to the technician example as an exception. In one place he
counts pronouncing a sentence as an example of an illocu-
tionary act.12 However, in View of his later characteriza-
tion of illocutions I believe pronouncing a sentence to be
an unfortunate and unharmonious example. In another place
Austin says "every genuine speech act is both illocutionary

and locutionary."l3

This passage is a puzzle, for how are
we to determine what Austin wished to exclude by using
'genuine'?

It seems, however, that we can employ as a defini-
tion of "sayingl something" the notion of uttering something
which one believes to be a sign. The act of making this
utterance is what I shall be referring to when speaking of
a locutionary act. This notion is slightly different than
Austin's concept of a locutionary act. He characterized a
locutionary act as the act of uttering something that in
fact has a certain meaning, that is, for him, sense and

14 Austin had some reservations about his

15

reference.
illocutionary/locutionary distinction. L. Jonathan Cohen
has argued in "Do Illocutionary Forces Exist" that indeed
one cannot distinguish between the two. He argues, roughly,
that a specification of the "meaning" of the utterance
involved in the locutionary act is the same as the specifi—

cation of the force (illocutionary) of making the utter—

ance.l6 However, it does not seem that Cohen is consistent

10

in his use of 'meaning', nor that his use of the word is
the same as Austin's. In spite of Austin's reservations
and Cohen's views, it seems that one can maintain some
useful distinction between illocutions and locutions.
There is a difference between sayingl something and sayingi
something. As we shall see in the final chapter, the
former involves certain beliefs on the speaker's part con-
cerning the features of his utterance. The latter involves
not only sayingl something, but also certain intentions on
the speaker's part concerning what he wishes to do in
sayingl something. These intentions indicate his motivation
for sayingl what he saidl rather than something else.

There is still another sense of 'say' to be noticed.
I might accidentally "say" something. For example, if I
mispronounce 'shut the door' someone who hears might think
that I had saidi "je t'adore." I would have accidentally

or inadvertently uttered something that happened to be a

 

sign. Let us use 'saya' to denote this accidental sense of
'say' when necessary.

We will also rely on the illocutionary/perlocutionary
distinction. In saying1 something a person might do any of
the following: report, announce, predict, admit, ask,
reprimand, pledge, request, order, propose, name, congratu-
late, promise, thank, or exhort. These are examples of
illocutionary acts. On the other hand, by sayingl something

a person might persuade, deceive, encourage, irritate,

11

amuse, frighten, get one to believe that such-and-such,
bore, inspire, impress, distract, get one to intend to do
such-and—such, or embarrass. These are examples of perlo-
cutionary acts.

To perform an illocutionary act is to sayl something
with a certain force in accord with convention.18 In one
place Austin says that to perform an illocutionary act is

necessarily to perform a locutionary act.19

(Perhaps Austin
has a very broad conception of locution or utterance.) But
in other places in latter lectures he says that what is
essential is that the illocution be done in accord with or
on the basis of conventions. Generally these conventions
are the conventions of a language, thus the connection to
locutions. But it is possible to perform an illocutionary
act non-verbally.20 Actually, both perlocutions and illo-
cutions can be brought off non—verbally. But one can
distinguish between the two in that conventions are essential
to illocutions; whereas one can bring off a perlocution by
non-conventional means.
Strictly speaking, there cannot be an illocutionary act
unless the means employed are conventional, and so the
means for achieving its ends non-verbally must be con-
ventional. But it is difficult to say when conventions
begin and end.
Austin's last remark about conventions will be born out when,
in later discussions, we will debate the extent of conven-

tions, those curious unagreed-upon—agreements, that make

communication possible.

12

The illocutionary force of an utterance is the
function that that utterance has by virtue of the conven-
tions of the language in which it is cast and the uses to
which it is generally put by people speaking in accord with
the conventions of that language. What illocutionary act
is being performed, that is, the illocutionary force of
one's utterance, is determined by the way in which people
generally use the utterance. That is, what people who
speak the language generally do, by Virtue of the conven-
tions of that language, in issuing that utterance.22 In
one place Austin says that we find out the force of an utter-
ance not by looking to the speaker's intentions, or to the
circumstances of the utterance, but to those conventions
which constitute the act.23 To say x is to do y. To deter-
mine what illocutionary act y is one should not look so much
to the speaker's intentions, that is to what act he intended
to perform, but to linguistic conventions. That is, to what
wg_do in saying x. The speaker's illocutionary intention
may or may not be in accord with these conventions, never-
theless, what he has done is determined by the conventions.

In the above paragraph I have extended the illocu-
tionary/perlocutionary distinction to intentions as well as
acts. It will become very useful to speak of the speaker's
illocutionary or perlocutionary intentions. These intentions
are, simply, what he intended to do in or by saying some—

thing.

13

Perlocutionary acts are acts done in order to
achieve certain effects or induce certain responses.18
Such acts are often done by means of language. At times an
illocutionary act can be a means to a perlocutionary act.
For example, I can warn you in order to frighten you.24 We
can distinguish acts from their consequences. We can, thus,
distinguish an illocutionary act from its intended or unin-
tended consequences. In the case of illocutions no conse—
quences need be intended. But, in the case of perlocutions
we have acts done in order to achieve certain consequences.
One may not intend every consequence of one's perlocutionary
acts, but in performing a perlocutionary act one does
intend some consequence.25 We can characterize the perlo-
cutionary effect of what we are saying as the result we

14

intend to bring about by what we say. Austin calls this

intended effect or consequence the perlocutionary object of

the act.26

When unintended the effect is called a perlocu-
tionary sequel. It seems that one cannot be said to have
performed a perlocutionary act unless one has achieved the
perlocutionary object. I have not amused you unless you
respond appropriately. But, I can warn you no matter how
you respond.

One's perlocutionary intentions must include the
intention to achieve some effect or produce some response in

one's hearer. One must be careful, however, for there is

also an "effect" or result intended in the case of all

14

illocutionary acts as well. This intended result is common
to all illocutions, although the speaker's intended perlo-
cutionary effects can differ from act to act. This result
or response is that the audience understand what the
speaker is doing. That is, that he understand the illocu-
tionary force of the speaker's utterance. Austin says,
. . . unless a certain effect is achieved, the illocu-
tionary act will not have been happily, successfully
performed. This is to be distinguished from saying
that the illocutionary act is the achieving of a
certain effect.
One's illocutionary intentions would include the intention
to "secure uptake." The audience is intended to understand
both the "propositional content" of the utterance and its
intended force.27 But no further "effect" need be
intended.

As we shall see, Grice's and Leonard's theory of
meaning focuses on perlocutionary acts, or better, perlo-
cutionary intentions. Both philosophers note that sayingl
something may produce certain consequential effects on one's
audience. The speaker may intend to achieve these effects.
The act done by sayingl something with the intention to
achieve some effect is characterized by Austin as perlocu-
tionary.28 Leonard and Grice will suggest that in such
cases it is appropriate to say that the speaker meant some-
thing. It seems, however, that it is better to say that

the speaker intended to do or achieve something by sayingl

whatever he saidl. It seems that we might not wish to claim

15

that he meantl what he saidl in such a case; we may not
even wish to say that he meant1 anything at all in such a
case, unless an illocutionary act were the means employed
to achieve the perlocutionary effect.

It seems that the intention to perform an illocu—
tionary act is not the same as the intention to perform a
perlocutionary act, just as the two kinds of acts are not
the same. Moreover it seems that meaningl something is
better associated with illocutionary acts or illocutionary
intentions rather than with perlocutionary acts or inten-
tions. However, it is not obvious that the two classes of
intentions cannot be reduced to some common class of inten-
tions, or that illocutionary intentions cannot be reduced
to some, perhaps very complex, perlocutionary intentions.
Indeed, at least one passage in Austin suggests that there
is a great deal of similarity involved here. Austin, for
example, speaks of possible perlocutionary objects and
sequels of illocutionary acts.26

Summarizing, we have illocutionary acts and perlo-
cutionary acts. We have a person's illocutionary intentions
and perlocutionary intentions. We can speak of the perlo-
cutionary effect of a speaker's utterance and ask whether it
was intended or not. We can also speak of the illocutionary
force of an utterance (which it generally has by virtue of
the conventions of some language). And we can speak of

response intended in the case of all illocutionary acts,

 

’ ‘ 1.i‘.,.i..||....

16

that of securing uptake. In all this let us assume that
the slippery distinctions can be maintained in some non-
pernicious way.

Some theories of meaning seek to analyze meaning in
terms of propositions, some in terms of sense and reference,
and some in terms of causes and effects. The theory of
meaning as intention is unlike these since it takes serious
note not of what is saidl but of the speaker's intentions
in sayingl it. It is not the sense, reference, proposi-
tional content, or effect of the utterance-type or token
that is of primary interest. Thus, this theory of meaning,
like those that attend to the "use" of an utterance, or its
part in a "language game," must be distinguished from
theories that remove the consideration of speakers from the
study of meaning.

Both of these approaches to meaning deserve philo-
sophical scrutiny. For the one is suited to meaning2 and
the other to meaningl. As reference and predication cannot
be ignored by an adequate theory of meaningz; one cannot
ignore the role of speakers in an analysis of meaningl.
Perhaps also, a complete approach to meaning would combine
the two. For how can speakers hope to be successful unless
there is some fixed sense to what they say1 and meanl?
Likewise, how shall we account for the origins of the
references and predications we examine if we exclude all

consideration of the speakers of a language? Moreover,

17

both aspects to the problem of meaning are interesting in
their own rights.

Having made the preliminary distinctions and noted
the place of this kind of theory of meaning relative to
other kinds, let us examine the rough outlines of the theory

of meaning as intention or purpose.

The Theory

 

This theory of meaning seeks an explication of

sayingl something and meaningl something by it. Since a
person may not meanl what his utterance means2 the analysis
of meaningl is thought to be independent of the meaning2 of
what is saidl. Moreover, the speaker need not have uttered
something he believes to be a sign, but only deliberately
produced something as a sign. Strictly speaking, what this
theory seeks to explicate is a person's meaningl something

by that which he uttered as a Sign. It is not necessary

 

that what is signified be itself a purpose, although the
speaker or author must intend or purport to signify some-
thing. The utterance itself is generally thought to "mean"
what the author intended to signify by it. Clearly, it may
not mean2 this at all. However, one can shift attention
from the intentions of particular speakers on particular
occasions. Advocates of this theory can hold the weaker
View that the meaning2 of an utterance-type is in some way
dependent upon the sense that the speakers of the language

"generally" intend to convey when using one of its tokens.

18

They need not maintain a direct reduction of the meaning2

of each token to what the speaker meant by uttering it.

1
Two paramount issues emerge: I, to provide an adequate
analysis of meaningl in terms of the intentions or purposes
of speakers; and II, to provide an adequate analysis of
meaning2 and its relationship to meaningl.

Issue I calls for analysis of the notion of
intending to signify something. What have I done when I
have intended to signify something by something that I have
uttered? If I utter something and meanl something by it
have I saidl it with more sincerity, or with the intention
to evoke some response, or with the purpose of achieving
some perlocutionary effect, or with the purpose of delivering
evidence to my audience that I believe something or wish
something done? The guiding hypothesis of this theory is
that meaningl can be analyzed in terms of some relevant kind
of intentions or purposes on the part of the speaker. This
hypothesis is suggested by the fact that 'mean' is, in some
contexts, synonymous with 'purpose' and 'intend' as senses
(a) and (b.1) indicate.

It appears that to mean something is to do some—

1
thing mental, or to do something akin to intending or pur-
posing, whatever these may be. To sayl something and meanl
it, or to utter something and meanl something by it, involves

a locutionary act. If idiomatic expressions lend any sup—

port to philosophical views, then there is some evidence

19

that meaningl something can be distinguished from

uttering something. Consider
(5) I said it, but I didn't mean it.

(5) also suggests that one might separate the intended per-
locutionary effect or illocutionary force of an utterance

from the meaning2 of the token employed. Other expressions
suggest that these effects and forces can be distinguished

from the hearer's understanding of the utterance.

(6) I meant it as a threat, not a promise.

In an appropriate context (5) is synonymous with

(7) I said it, but I didn't intend it literally.
And in an appropriate context one might replace (6) by

(8) It was not my purpose to promise but to threaten
when I said that.

Thus there is some evidence for the hypothesis that to meanl
is to intend or to purpose in a certain way (however these
may be thought of or classified). The evidence adduced here
can lend only some initial plausibility to the hypothesis.
This theory of meaning derives plausibility from
other sources as well. First there is a common association
of meaning with intention and purpose as is evidenced by
the dictionary definitions. (This might be a result of the
fact that speaking is deliberate purposive behavior.)
Second, the theory allows its authors to account for a wide

range of phenomena. Given the identification of meaning

20

and purpose Leonard is able to incorporate his philosophy
of language with a theory of signs (natural and conven—
tional) and sign-events (natural and deliberate) as well as
with a theory of purposeful behavior. Leonard offers an
account of the general functions of language and is also
able to generate a theory that attributes truth-values to
interrogatives and imperatives as well as declarative
sentences. After accounting for the truth or falsity of
what is saidl Leonard discusses the concepts of honesty,
dishonesty, and candor in sayingi it. His theory also
includes a discussion of the nature of purposes and propo-
sitions, which is, in effect, a discussion of the nature of
meaningl. Grice offers a detailed analysis of meaningl in
terms of intentions. He shows no lack of zeal in trying to
provide us with a list of intentions that constitute the
"Meaning-intention" (he sometimes refers to this as the
"M-intention" or the "Meaningnn-intention"). Grice also
offers a sketch of a theory of meaning2 for linguistic as
well as non-linguistic signs. Unlike Leonard, Grice devotes
most of his effort to the analysis of meaningl. This issue
is the central concern of the present work. Before con—
fronting the issue, let us examine in more detail the pro-

grams followed by Grice and Leonard.

Grice's Program

 

Grice's goal is the explication of his favored sense

of 'say' and the relation of this notion to meaning2.29 Let

21

us use 'sayG' for Grice's sense. His sparse comments leave
us with no idea how to characterize sayingG something. He
tells us that we can distinguish sayingG something from what
has been "implicated" by what was saidi. According to
Grice, if I sayi "p and q, therefore £5 I would have saidG

that p and that q and that E! but only implicated that E

0 There is no

follows from the conjunction of p and q13
evidence of how to make the distinction. If I saidi "3 even
though q" or "B because q" or "p yet q" I should not know
how to determine what has been saidG and what has been
implicated. Actually, if I saidi "p or q" I should be more
prepared to argue that I have asserted the disjunction
(saidi "either p or q") rather than that I have saidi that p
or saidi that q and only "implicated" that the truth lies,
after all, in their disjunction.

Not only is Grice's favored explicandum not cleared
up by this distinction, but matters are made more difficult.
The new, and equally bewildering notion of "implicature"
has been added. Grice will go on at times to speak of the
"conventional implicature" of a sentence.31 He does not
reveal if we are to regard this as the same as the meaning2
of the sentence or not. Unfortunately Grice's intolerably
bad start is an omen.

Presumably there is a definite relationship between
the "conventional implicature" of a sentence, or any such

complete utterance, and what the speaker meant by uttering

1

22

the sentence. The conventional implicature of the sentence
is, in normal cases, the speaker's means of revealing his
intentions ("his meaning") and the hearer's means of recog-
nizing what the speaker intends. Grice, in noting this
conventional mode of correlation between sentences and
speaker's intentions, thinks that he has discovered one of
possibly many modes of correlating sentences with "responses"
of the bearer. This unfounded jump from intentions of the
speaker to responses of the hearer is made because Grice
claims that part of meaningl something is to intend some
response in one's hearers. Grice's move from the speaker's
intention of producing some response in his hearer to that
response itself is clearly mistaken.

The program to explicate sayingG and its relation
to meaning2 is to have seven major stages. Grice in the
first stage distinguishes between several relevant senses
of meaning. Grice uses 'meaningn' to designate the natural
sense of 'meaning', and 'meaningnn' to designate its "non-

natural" sense.

(9) These marks mean that an animal was confined here.

(10) Those clouds mean rain.

Both (9) and (10) exemplify meaningn. (11) and (12)

exemplify meaningnn, which is the relevant sense of the word.

(11) When I say "stop" I mean stop!

(12) 'Stop' means desist.

23

Grice does not, especially in his earlier work, take full
note of meaning1 and meaning2 which are both discernible
within his meaning .
nn
In his later work Grice claims that there are four
senses of meaningnn. Two of these seem to be related to
meaningl and two to meaningz. I shall use superscripts to
mark off Grice's senses of meaningnn and drop the subscript

'nn' where the context makes it unnecessary.

Meaningl Grice calls this the "timeless meaning of an

 

utterance-type." If the utterance-type were a
word one could give its meaning1 by listing aII of
its senses, as was partially done above in the
case of 'mean'. Similarly one can give the
meaning1 of a gesture or a sentence by listing
everything it might meanz.

Meaning2 Grice calls this the "applied timeless meaning" of

 

an utterance-type. If I said that the sense of
'meaning' in the previous sentence is meaningnn and
not meaningn or meaningl I would have given the
applied timeless meaning of the word 'meaning'.
That is, the meaning of the utterance-type here

used as it is derived from the meaning1 of that

type.

Meaning4 What a speaker meansl by uttering x at a particular

time is called the "utterer's occasion meaning."

24

Meaning3 We might think of the utterance-token as having
a meaning by virtue of what its producer meant4
by it. This meaning might not be a part of the
meaning1 of the utterance-type I to which the
token x belongs. We would be claiming that the
token "means" whatever its author meant4 by it.
We might go on to attribute this "meaning" to the
utterance-type I as well, since it is the "mean-
ing" of one of its tokens. We might thus have
created what Grice calls the "utterance-type

occasion meaning."

Surprisingly Grice predicates meaning3 of types, not
tokens. He would have us believe the odd-sounding view that
an utterance-type means (means3) that E if and only if some-
one meant (meant4) that p by uttering a token of it.32
Clearly this violates the meaningl/meaning2 distinction
unless we say that talk of "what an utterance-token meansl"
is only short for talk of "what the producer of the token
meantl by it."

Since Grice takes meaning4, utterer's occasion
meaningnn, to be more primitive than the other three senses
of meaningnn, the second stage of his program is the explica-
tion of meaning4 in terms of the intentions of the utterer.

In Chapter III I present in detail the execution of the

first two stages of Grice's program.

25

Stage three elucidates meaningl; stage four,
meaningz, and stage five, meaning3. Stages three and four
are partially outlined below. Grice hopes to show that
given an adequate analysis of meaningl (strictly speaking,
meaning4), one can formulate an adequate analysis of con-
ventional meaning for both linguistic and non-linguistic
utterance—types. Moreover Grice seeks analyses for both
complete utterance—types, like sentences, and incomplete
utterance-types, like words or expressions. His analysis
of meaning2 often relies on such intuitive concepts as
"having it in one's repertoire" and "being a resultant
procedure." Often these are left woefully vague.

Stage six is the specification of the conditions
under which the meaning3 of the speaker's utterance is what
he saidG. The final stage is a supplementation of this with
an account of the elements that make up the conventional
meaning of an utterance but are not elements of what the
speaker saidG. Grice outlines his program in "Utterer's

29 He has

Meaning, Sentence-Meaning, and Word-Meaning."
published the details of the first two stages of his pro-
gram, but offers only a few pages of notes about the next
three stages and a brief sketch of how one might treat of
the final two stages.

Let us assume that Grice has an adequate analysis of

meaning4 in terms of the intentions of the speaker, U. When

I utters a token I of type I he has meant4 something by

26

uttering I if and only if he has a certain set of intentions;
call this set the "M—intention." At least one of the inten-
tions in this set is to achieve some perlocutionary result.
Grice seems to believe that these results are variations of
the following: U may intend to have his audience, I, come
to realize U's attitude toward proposition p. Or, U may
intend that A come to have the same attitude (or some atti-
tude) toward p'by virtue of realizing what U's attitude
toward p is. If U.M-intends one of these kinds of results,
as opposed to simply intending it, then U is trying to
achieve the result as the perlocutionary effect of uttering
something. Although one ordinarily would make reference to
the intended results in a specification of the intentions in
the M-intention, we can, following the above, speak of "M-
intending the result." Let us confine our attention to
assertions offered by U to bring about beliefs in A. We can
characterize meaning4 by the following rough definition,
leaving the exact content of the M—intention for Chapter III.
Notice that, following Grice, U always intends something like
clause (1) and may intend (2) as well. (Notice, also, how

oddly restricted is the avenue of A's coming to believe that
2.)

(Def. 1) By uttering x U meant that p if and only if for
some A‘U uttered x M—intending that A should:
(1) think that U believes that p, and in some
cases also: (2) come to believe that p_himself
by virtue of thinking that A believes that p.33

 

27

Although issue II is not the main concern of the
present work, it would be instructive to follow Grice's
path to see how difficult and complex a matter it is to
provide an analysis of meaning2 and its relationship to
meaningl. This complexity is evident even apart from the
idiosyncrasies that distract us in Grice's own approach.
Grice's analysis of meaning2 is intended to account for
both non-structured utterances, like gestures, and the syn-
tactically structured utterances of a language. Let us
begin with a non-structured utterance. Suppose that a
particular U uses gesture g to let A_know that 2. Thus, for

U, or in U's idiolect, g means that p. This might be

2
defined by saying that it is U's policy to utter q if, for

4 .
If on an occa81on when

some I, U wants A,to think that p.3
U does utter g he has what Grice refers to as "the simple
intention" just mentioned, then according to Grice, we can
infer that U_M-intended to affect I in the specified way.
That is, by uttering g U_meant4 something.

However, it is possible that g has still another
sense in its meaning1 within U's idiolect. Grice offers the
concept of "having a certain procedure in one's repertoire"
to provide for this. For example, U_may have it in his
repertoire that he lights a cigarette when he is nervous, but
he may, at times, play with a paper clip instead, and he may

light a cigarette when he is not nervous. Relying on this

notion Grice suggests a definition like the following:

28

(Def. 2) For U utterance —type X means (has as one of its
""""" "pif and only if U has in his reper-
toire the procedure to utter a token of I if he
intends some A to believe that p. 35

 

I n u l o o I o -
Gr1ce general1zes from mean1ng w1th1n an 1nd1v1d-

ual's idiolect to the meaning1 of I for a group, G.

(Def. 3) For group G, utterance-type X meansglp" if and
only if at least some (? many) members of group
G have in their repertoires the procedure of
uttering a token of X if, for some A, they want
A to believe that p. 35

According to Grice a member of the group retains this pro-
cedure in his repertoire because he believes that some other
members of U have, or have had, this procedure in their own
repertoires. The members of group U "conform" to the prac-
tice or habit of, in general, using such-and-such a device
to accomplish such—and-such an end. Where there is conform—
ity there is more than a "usual" or "unusual" use, there is
a "correct" and an "incorrect" use. However, Grice may be
requiring too much here. It would be sufficient if the
members of U merely knew of the procedure. They need not
actually be willing or able to do it themselves--have it in
their own personal repertoire of actions.

Anticipating this objection Grice alters the original
understanding of having a procedure in one's repertoire. It
is sufficient not only that U have the readiness to employ
procedure I, but also that U "is equipped to use" I even if
he would never intend to use it. Grice soon abandons

attempts to clarify these concepts.36

29

In spite of the problem of providing clearer under-
standings for these essential concepts, Grice's position
does have virtue. One can now speak of what U should meanl
when he utters I. If U performs I before a member of his
own group, and if the group recognizes I—ing as a sufficient
condition for U—ing, by virtue of the conventions of their
community, they can rebuke U for I—ing without the proper
intentions. That is, for not realizing that or what he had
signaled. U should have meant what Imeans1 for them. Thus
one can make an individual's meaningl dependent upon the
meaning2 of a particular action for a particular group.
Officially, however, the dependency is to run the other
way. The ultimate source of meaning, in this kind of a
theory, is to be meaningl not meaningz. Grice's views pre-
serve this dependency, for the necessity mentioned above is
a practical or moral necessity, not a logical necessity. I
shall have more to say about this matter later.

Grice turns to the definition of applied timeless

meaning for unstructured utterance-types, meaningz.

(Def. 4) When U uttered token x of type X, X meant "pf
if and only if, for some A, U intended that A
recognize (and perhaps to recognize that U
intended A.to recognize) what U_meant4 by
uttering x, on the basis of A's belief that,
for U, Imeansl "p."3 —

 

To know what U_means4 one must know how U acts. One must,

as Grice's definition represents, know what is in U's reper-

toire. On this basis one can reasonably infer what U_meant4

30

knowing what I_meansl. But I may be ambiguous, thus U's
intentions as indicated in (Def. 4) may be frustrated. Afs
success in coming to know what U_meant4 depends on selecting
the appropriate member of I's possibly many—faceted mean-
ingl. The "appropriate" selection is known as the applied
timeless meaning, the meaningz, of I. To learn what U
meant4 A must know what I meantz. To know this A_must know
what U_meant4 by I on this occasion, assuming that U is
relying on the meaning1 of I in selecting I as his utter-
ance. Unless A can learn how U acts, what he says when and
why, he cannot break the circle. This circle is a more
narrow, thus more vicious, version of a generally not too
troublesome circle involving a particular language and those
who share it. Unless one knows the language it is difficult
to find out what its speakers meanl and what its utterance-
types meanz. The difference is only in that one cracks the
first circle on the side of the speaker, but the broader
circle is often entered on the side of the language.

Grice turns next to syntactically structured
utterance-types, intending to provide definitions for both
the meaning1 and meaning2 of complete utterance-types,
sentences, and incomplete utterance-types, words, and expres—
sions. The standard meaning2 of a sentence or phrase is
consequential or resultant upon the meanings2 of the words
or elements that enter into its construction, says Grice.38

Surely Grice should include word order, punctuation or

 

 

31

inflection, and perhaps context as well, since these also
play a role in determining the meaning2 of a sentence or
phrase. But perhaps this widely accepted claim is still

not perfectly accurate. The meanings of the words in

2
(13) He's a son of a stickleback fish

are known. But, most likely, what the sentence means2 is
not a function of solely the meanings2 of its words and its
grammatical structure.

Within the broad notion of standard-meaning2 one
finds buried both literal-meaning2 and idiomatic—meaningz.
The former is a function of the common, or standard meaning2
of words and grammatical structure. The idiomatic-meaning2
of an expression is a meaning2 associated with it that may
be the same as its literal-meaning2 but need not be. No
doubt anyone who has attempted intralinguistic translation

has encountered the problem of providing an idiomatic trans-

lation given a dictionary recording only literal meanings

2.
In Polish, for example, a double negative may be used for
emphasis. It is idiomatically correct to render "Nic nie
mamy" as "We have none!" rather than as the literal "We do

not have nothing." Idiomatic-meaning2 is a standard—
meaning2 that applies to an entire expression and which may
be independent of the literal—meaning2 of the expression.

A complete analysis of meaning2 would provide for both of

these kinds of standard-meaningz.

32

The above distinctions give pause to the ready
acceptance of Grice's view that the meaning2 of a sentence
depends in a completely straightforward way upon the mean—
ings2 of its component words. A further difficulty for this
principle is multivocality. The word '1ove' has, for
example, a literal-meaning2 that is, within vague limits,
rather constant throughout nearly all of the uses of the
word. However, in scoring a tennis match the word takes on
a different meaningz. Perhaps then, the literal-meaning2
of a particular word-token depends upon the context of the
utterance of the sentence-token. One must know first what
the sentence means2 before one can know what a particular
word in the sentence meansz, or at least so it seems. Even
if we restrict Grice's claim to the literal aspect of
standard-meaningz, the selection of the "appropriate"
meaning2 for a particular word or expression is not an
entirely straightforward matter.

Grice introduces a concept which is designed to
capture those essentially recursive elements of linguistic
communication which Paul Ziff accuses him of overlooking in

39 It is the concept of a

his first analysis of meaningl.
resultant procedure. Roughly, a resultant procedure is some
procedure for producing an utterance which is itself deter-
mined by still more primitive procedures.38 Grice says no

more, but offers us an example. (As has happened before,

Grice's method of explanation-by-single-example fails us

 

33

when we seek more than an intuitive notion.) Suppose that
U has the procedure, II, in his repertoire of asserting p
when he wants some A_to think that he believes that 2.
(According to Grice II identifies the indicative mood with
the grammatical structure for assertions, and it correlates
this mood with the propositional attitude, believing. Grice
here assumes that p is an indicative expression. The mis—
taken suggestion that certain moods are used for specific
purposes and correlated with specific propositional atti-
tudes is not essential to Grice's position, although he
frequently repeats it.)

By an "R-correlation" Grice understands a refer-
ential correlation which associates a particular object with
a nominal word or phrase, I. A "D—correlation" is a denota—
tional correlation which associates a class with an adjec-
tival word or phrase, U. For the moment let us follow Grice
in this implausible extensionalist direction by letting an
"M-correlation" be a mood correlation which associates a
particular grammatical structure with a particular proposi-
tional attitude. For example, the indicative mood is M-
correlated with believing by II,

Grice continues his example. Suppose that U also
has procedure II in his repertoire. He utters an M-correlate
predication of U on E when he wants some §.t° have a partic-
ular attitude toward the proposition that the R—correlate of

E is a member of the D-correlate of U. Given that U has II

34

and II we might assume that he has the resultant procedure
III. He utters an indicative predication of U on E if he
intends that some A think that he believes that the object
R—correlated with U is a member of the D—correlate of U.
We can expand the example by assuming that U believes that
Jones' dog is an R—correlate of 'Fido' and that the set of
hairy—coated things is the D—correlate of 'shaggy'. Thus,
if U utters "Fido is shaggy" and if he has III then,
assumes Grice, U is acting on the basis of another resultant
procedure, III: U utters "Fido is shaggy" if he wants some
A to think that he believes that Jones' dog is hairy-
coated.40
Here again Grice has moved beyond the bounds of
logic. We can, at best, infer that A, unless he is Jones,
will come to think that something or other, not necessarily
Jones' dog, is the R—correlate of 'Fido', and that whatever
this something is, U believes that it is shaggy. The refer-
ential opacity of the context 'U believes that U is shaggy'
prevents one from inferring that A knows what the R—correlate
of i is. If the English language were a purely extensional
language, if it had no referentially opaque contexts, if it
were used solely for the purpose of displaying sentences
which could be used only to make assertions and commands and
the like without the use of metaphor and irony; then, per-
haps, the program of producing X-correlates would provide an
adequate basis for deriving the "resultant procedures" of

English-speaking people.

35

We have followed Grice long enough in his pursuit
of an analysis of meaningz. This record has provided a
launching point for several more general comments on some
of the issues involved in the analysis of meaningz. Two
things are evident from the discussions: issue II is not
lacking in complexity, and Grice's approach contains
serious, although perhaps not insurmountable, flaws.
Besides his mistakes in logic, his notions like that of a
resultant procedure are problematic and jeopardize later

developments in his program.

Leonard's Program

 

Leonard's philosophy of language coherently accounts
for a wide range of phenomena given the assumption that

meaning is purpose. I say assumption" because Leonard does
not present a complete theory of meaningl although he does
discuss meaningl in several places. Let us look at some of
his most original claims. In "Interrogatives, Imperatives,
Truth, Falsity and Lies" he comments in passing on the
"nature and grounds" of sentence meaning, urging that what
he has to say does not apply to "all kinds of meaning nor
all kinds of signs."41
Since "only concrete particulars have meaning"
meaning attaches to sentence—tokens, not sentence-types.

Notice Leonard's apparent confusion of the meaningl/

meaning2 distinction in:

36

The meaning attaches to the utterance or token despite
the fact that the identity of the meaning which attaches
to the utterance or token is generally to be discovered
by referring to the form or forms which the utterance
exhibits or illustrates.

We shall see that the "meaning" that attaches to the token
is not meaning2 but something like Grice's meaning3, which

is derived from what the producer of the token meant by it.

1
Yet the identity of this token's meaning3 is learned "by
referring to the form." This is, it seems, only a disguise

for claiming that we look to the meaning2 of the sentence-

type to help in determining what the author meantl (i.e.,
what the token "means"). But, then types have meaning!

Leonard's way out might be to claim that it is a "different
kind" of meaning; not the kind referred to in the claim
"only concrete particulars have meaning."

Leonard distinguishes between natural and conven—
tional signs, and between natural and deliberate sign-
events. A sign—event is the deliberate production or the
interpretive reading of a sign. Since spokenl utterances
are deliberately produced signs we can ask what they sig—
nify. An utterance-token can mean certain things by virtue
of being a natural sign of them. The token might be read
as a sign that its producer is English-speaking, or that he
is insecure, or that he is highly educated.42 But "to ask
what an utterance means as a deliberate sign is to ask what
its speaker was intending to accomplish by making that

43

utterance." Or again: the meanings of deliberately

37

produced conventional signs are "whatever was the purpose

"43 A discussion of

of their author when he produced them.
the nature of the kind of meaning Leonard is trying to
isolate here would involve a discussion of the nature of
purposes. A discussion of its grounds would be a discussion
of authorship. Let us turn to authorship first.

Authorship is the bringing about of a candid pro-
ductive sign-event; this is, in turn, a kind of deliberate
sign—event. Not every deliberate act issues in the delib-
erate production of a sign. The agent must not only act
deliberately, but also intend that the state of affairs he
brings about should function as a Sign. Each "deliberate
UIgU is dictated by at least two purposes. The one of these
is the purpose of signifying. The other is the purpose
signified."44 This latter purpose is the "deliberate
meaning" of the deliberate sign, provided that the two

purposes relate as follows: Purpose I is the deliberately

signified purpose of sign I if and only if the speaker U

 

deliberately produced I believing that some audience I
would read I as a sign of I and believing that thereby I
would be achieved.45 Notice how crucial are U's beliefs
concerning the actual results of his use of U. In Chapter
III I argue that U's beliefs concerning the token employed
play a crucial role in meaningl. If this is so, then a
theory of meaningl that lists only U's intentions or pur—

poses must be inadequate.

 

38

A deliberate sign signifies "every one of its pur—
poses" and many other things as well by virtue of being a
natural sign. The deliberately signified purpose (the
deliberate meaning of a deliberate sign II that purpose
as an instrument for the accomplishment of which the sign
was selected, provided that this selection was dictated
by the conviction or hope that the purpose will be served
because the sign will be taken to be a sign of that
purpose.44 If the agent engaged in a deliberate sign-event
intends that his act of bringing about a certain state of
affairs, rather than simply the state of affairs itself,
function as the sign of his purpose, he is engaged in a
productive sign—event. Leonard uses 'sign ' to refer to

1

the sign that U relies on in a productive sign-event. In

 

"Authorship and Purpose" this is defined roughly as follows:

(Def. 5) A person, U, produces x as a sign1 to another
person, H, that p if and only if U brings I
about in order that H read U's bringing it about
that I as a sign thaE p.46 _

 

(Notice that here Leonard is using the propositional
variable 'p' whereas from what has been said above we might
expect that he would use 'P', to range over U's purposes.)
Signsl need not be either truthful nor honest. If
p is false they are not truthful; if U believes that p is

false, they are not honest. Moreover, a sign may not be

1

candid. If U produces a sign to some I but intends that U

1

not realize that U is deliberately producing the signl then

 

39

U is not being candid. When U is not being candid he is
hoping that (or perhaps, intending that) his hearer be
ignorant of the fact that U's bringing it about that I was
a deliberate (as opposed to natural) sign—event.47 Leonard
defines authorship, the production of a candid signl,
roughly as follows:

(Def. 6) An author, A, produces x as a candid sign to H
that p if and only if A brings x about In order
that H will read A's bringing I —about as a sign 4
that A is producing I to be a signl to I that p.

 

8

In this A intends a double reading. First I is intended to
read A's bringing I about as a productive sign—event. Then
I must read A's bringing I about as a sign (natural or con—
ventional) that p. Leonard is trying to capture the fact
that the audience cannot tell what the author intends him
to understand by the deliberate sign unless he first recog-
nizes the author's behavior as a sign addressed to him.

On the basis of this we might conjecture how Leonard
would define meaningl. The following definition is designed
to resemble the Gricean definitions that follow in later
chapters. Notice that in Leonard's case it is clear that
U's bringing it about that I is to function as the sign;
Grice will obscure matters later by employing the ambiguous

'utterance' in his formulations.

(Def. 7) By uttering x U meant that P if and only if

1

For some H, U brought about x intending that:

(1) his bringing about x can _be read by H as a
sign (either natural or conventional) of U' s
purpose I;

 

40

(2) I think, by virtue of U's bringing I about,
that U intends (1).

(Def. 7) tortures Leonard's concepts somewhat, but it does,
when compared to Grice's work, reveal some of the similarity
of their two approaches to meaningl. Note that I have sup—
plied the definition with the purpose variable 'P'. Leonard
claims that when the sign—making is candid the signification
of the Sign produced is always a feigned or actual purpose
of the author.49 Because it may be feigned, it is not
necessary to require that U purpose I in the definition.

Although the defects and virtues of (Def. 7) should
be discussed, they will not be. Why? First, because I can
only conjecture that Leonard might have accepted something
like (Def. 7); second, because I believe that whatever
Virtues it possesses are captured in my analysis of meaningl
in the final chapter; and finally because I hope that its
defects are there avoided. Some of these defects and vir-
tues will become more obvious as the argument of the present
work proceeds.

There is a crucial transition in Leonard's theory of
meaning from the grounds of meaning to the nature of meaning.
Does it follow that if to meanl something is to act purpose-
fully, the signification of the utterance employed must be
a purpose? The answer is "No." (Thus, one defect in
[Def. 7] is the use of 'P' for 'p'.) But it is unclear how
Leonard stands with regard to this. At one point he claims

a theory of sign—reading and a theory of purposeful behavior

41

can be "appealed to in order to show that when the sign—

making is candid . . . the deliberate meaning or delib—
erate signification of the sign employed . . . must be a
49

feigned or actual purpose of the author." But when we
get into the actual demonstration we find some weaker claims

being made. There Leonard maintains that "when the sign—

event is candid the author cannot expect the offered sign

to be read otherwise than as a sign of some purpose enter— ‘

50 . .
" The cons1derations become more

tained by the author.
practical than logical. If the author wishes his sign to be
interpreted correctly he must, if he is practical, offer it
as a sign of a purpose. Leonard argues that a reasonably
sophisticated person, I, will interpret A's candid signsl
by reading A's production of I as a sign of A's purpose
relative to him and a certain proposition, e.g., that he, I,
think that g. If this is U's interpretation, "the author
may as well capitulate and intend that kind of signification
by it."51
But Leonard's claim is still unnecessarily strong.
It is sufficient for this theory of meaning to show that to

mean is to purpose in a certain way or to purpose certain

1
things; one need not argue that the token "means" a purpose
of its author. We need not argue that what is signified be
itself a purpose, nor that any practical author will intend
to signify only purposes by his candid signsl. But suppose

he does, what then is the nature of this purpose, i.e., of

meaning?

42

It is generally, according to Leonard, the author's

purpose that I stand in some relation towards some proposi—

tion, p.52 (It is most obscure in the Leonard text, whether

or not this proposition p is the same as the proposition p
referred to in (Def. 5) and (Def. 6). If it is, then
Leonard's theory is in more trouble. If p can be said to

be what A means in (Def. 6), then we can say that, in some

1

sense, the signification of a candid sign is a proposition,

1
not a purpose.) But these elements of the author's purpose

are not immediately apparent. Every purpose must first be

 

analyzed into its primitive concern and its primitive topic
of concern. The latter is always a proposition, that is,
some envisaged state of affairs or possible situation. The
primitive concern is always an attitude of the speaker, pro
or contra. The purpose is the speaker's attitude toward the
proposition.

The primitive topic of concern may be a proposition
to the effect that someone act in a certain way. The pur—
poser may move toward achieving or preventing this by com—
municating to the agent, I. For example:

It might be my primitive concern to make it true, and
be my primitive topic of concern, that you telephone me

 

tomorrow. . . . The communication would consist in
addressing to the agent a deliberate sign of this pur-
pose, . . . "Phone me tomorrow."

An author's primitive topic of concern is often a

proposition to the effect that the agent, I, act in some way

relative to some nested or contained proposition, p.52 The

43

nested proposition is the ultimate topic of concern. The

 

ultimate concern becomes the fusion of the primitive con-

 

cern with what is left of the primitive topic of concern
after the ultimate topic of concern is removed. When no
contained proposition is discovered in the primitive topic
of concern, then the primitive tOpic of concern becomes, as
it stands, the ultimate topic of concern.

A sentential utterance is said to signify a purpose

of its author by eIpressing his ultimate concern and indi-

 

cating his ultimate topic of concern. Leonard argues that
if the proposition a sentential utterance indicates is a
fact, the sentence is true. If not, it is false. Thus,
EXEEX sentential utterance has a truth-value.53 Generally
changes in the author's ultimate concern are manifest in
changes of grammatical form. The author may express one
ultimate concern toward p by using the declarative form, and
express another by using the imperative or interrogative
form. Changes in the indicated topic of concern are
generally manifested in variations in the subject and predi-

cate of the sentential utterance.54

A token signifies or means a purpose, the identity

 

of which is to be learned by reference to the, let us say,
meaning2 of its type. But where does meaning2 come from?
Leonard offers us one hint when he says

The "meaning" of a sentence is thence viewed as an

abstraction from the signified meaning (always a
purpose) of the uttering.49

44

Sign-types are aids in determining the significance of their
tokens only by virtue of the intentions of the producers of
these tokens. If, in general, the producers of tokens of a
certain type did not meanl such-and-such by their tokens,
these tokens would not signify what they do. We can claim
that the meaning2 of the type can be abstracted from this.
The utterance-type means2 what people generally meanl by

its tokens. Meaning2 is derived from the common will,

intentions, or purposes of the producers of conventional

signs.55 Leonard is suggesting that meaning2 is dependent
upon the purposes of speakers insofar as they implicitly
agree to speak in accord with those conventions that con-
stitute their language. Further, this is the practical
thing to do.

The hearer, U, can move from the meaning2 of a token

(as this is derived from the meaning2 of its type) to what

its author meantl by it. But this move is only a gUUUI.

The token is only a UIUU to its author's meaningl. The fact
. . . that the chosen tokens do signify the author's
purpose, will be of little value--either to the author
who is attempting to communicate his purpose or to the
receiver who is trying to discover the author's purpose--
unless the receiver is actually aided by the sign-tokens
to identify correctly the author's purpose.

The author might be foolish and choose an inappropriate

token, thus defeating his own purpose and confounding his

receiver. A responsible author will choose appropriate

tokens to facilitate the receiver's "guessing game." The

token which best indicates the author's topic of concern

45

and best expresses his concern with that topic would be the
best token he could choose to signify his purpose. What a
token indicates and expresses is largely a function of its
conventional meaningz. However, conventionally determined
meaning2 is independent of the particular use to which this
token is being put by this particular author. And so the
meaning2 is only a clue to the author's meaningl.

In his earliest formulations Leonard claimed that
the meaning of a sign-token was whatever its author intended
it to be. We know, now, that he was speaking of its sig-
nification, not its meaningz. An author can be mistaken
about what his token means2 to his audience. If this occurs
it is likely that the audience will be in error about the
signification of the token. He might take it to be a clue
of one thing when in fact it was intended to be a clue of
something else. The author, by his choice of that token,
cannot be wrong about "its meaning" (its signification).

He gives it "its meaning." He can only be unwise in his
choice, thus making his audience's task of guessing the
author's meaningl more difficult. (These statements would
be more informative were it not for the notion of significa-
tion that Leonard is working with.) In this Leonard finds
the "arbitrariness of language"--that the "meaning" (sig-
nification) of each sign-token is derived from the intention

of the author who uses it.57

46

Herein lies a virtue of Leonard's theory: We can
mean1 whatever we wish by what we say; and we must mean1
what we say. Less aphoristically, we do not have to be
speaking in accord with conventions for our utterances to
have meaningl. Nevertheless, to be successful in communica—
tion, to get someone to understand what we meanl, we must,
if we are practical, abide by conventions. Practically
speaking, we must meanl what our utterances meanz.

Perhaps this is not enough. Perhaps we have more
than a merely prudential constraint on what we can mean by

1

what we sayl. Can a jury sayl 'Not guilty' and meanl that

the accused is guilty? Can a bride sayl 'I do' and meanl
she does not? Can a bridge player sayl 'One heart' and
thereby bid three spades? There is a certain practical
necessity about the way we speak; it is more than a coinci-
dence but less than a logical necessity. Sayingl certain
things in certain circumstances makes us accountable or
responsible, given the conventions of the language, for
certain things. For example, if I ask you to stoke the fire,
I can be held responsible for indicating that I believe that
there is a fire, that you have the power to stoke it, and
that I want it stoked. (But, of course, the sentence 'Stoke
the fire' or 'Would that the fire be stoked' does not entail
any sentence like 'There is some fire such that. . . .) To

violate the conventions of language is more than simply to

act in a silly and impractical way--it is to create and to

47

speak a new language. I shall return to this point again

in the next section and in the final chapter.

Difficulties

 

Several difficulties with this theory have become
more or less apparent in these outlines and discussions.
For one, there is a problem concerning the "arbitrariness"

Leonard spoke of. This theory seems to allow either

(Thesis 1) A speaker may meanl whatever he wishes by what-
ever he utters.

or

(Thesis 2) Whatever a person utters means whatever he

wishes it to meanz. 2

Were it not for the distinction between the two senses of
meaning, the first would entail the second. Whereas the
first is a correct interpretation of this theory, the second
is an unfortunate, and obviously false, view often attrib-
uted to Grice and Leonard.

Can we sayl whatever we please and meanl anything
we wish by it? The answer to the practical question is,
rather obviously, "No." Since both Grice and Leonard sup-
pose that one uses discourse to achieve one's purposes, this
theory actually provides a reason to employ the conventional
meaning2 of one's utterance. We need not, however, fall
back on Leonard's discussion of clues and guesses to realize
this. People will, unless we advise them of some idiosyn-

cratic or esoteric peculiarity of our discourse, interpret

48

what we meanl on the basis of the meanings of our utter-

2
ances. The fact is we look first to meanings2 and only
then to purposes or intentions. At times our audience may
not know our intentions, nor even be intended to know them.
But we know that it is impractical to assume that people do
not meanl what they sayl, in normal cases.

Stanley Cavell in "Must We Mean What We Say?" makes
the strong claim that "there is a relation between what one

"58 On the other extreme we

says and what one IU§I_mean.
have the false (Thesis 2). In the middle, apparently
sensitive to both views, we find J. L. Austin. In "Plea
for Excuses" he said that we should study ordinary language
by examining "what we should say when, and so why and what
we should mean by it."59 Is Austin's 'should' merely a
caution to be prudent, or is it more? The answer is not
easily discovered for no methods of resolution present
themselves. Even citing cases, as was done above, is not
definitive since it is not clear how one should interpret
them. There are legal restrictions on the freedom of
speech. We know that one who shouts "Fire" in a crowded
theater may be prosecuted. But what does the existence of
such a law indicate?

It seems that one can meanl whatever one wishes by
the words or sentences he speaks. But this logical possi—

bility has dramatic consequences if acted upon. A group of

people acting with all their imagination upon this option

 

49

would soon totally and effectively undermine the very con—
ventions which they were suspending. Communication would
be in a dismal state, if it were possible at all. Every
word and gesture would be plagued with the imminent use-
lessness.

Our linguistic conventions, though not logically
necessary, are not violated without grave results. They
constitute 92E language, ranging not only over its grammar
but also its use. They indicate when, why, and how we are
to act or avoid acting whenever language is a part of our
social intercourse. They do not, generally, prescribe the
specific act to perform, but they limit the range of
alternatives by indicating which are out of place. Insofar
as we are members of a community the linguistic conventions
of that community permeate our lives, although our accep—
tance of them is only implicit. Some see this theory of
meaning as a threat to our language. They think that it
entails (Thesis 2) which is an invitation to "relativism"
and "subjectivism." But this theory does not represent such
a threat. While it indicates that people may meanl whatever
they wish, it also urges that, if practical, people will
meanl what their utterances meanz.

A second set of problems that both versions of this
theory of meaning share is inherent in the devices Grice
and Leonard have chosen to use in its presentation. Both

versions of the theory seem committed to the use of

50

propositions. This, of course, imports familiar logical
and metaphysical problems. There are problems about the
identification and individuation of propositions, problems
about their ontological status, as well as their relation-
ship to specific linguistic constructions like sentences
and 'that'-clauses.

Other problems plague both versions. Its authors
are trying to analyze the pragmatic aspects of meaning;
this imports problems associated with the use of inten-
sional notions like believing, intending, etc. We can
expect to encounter problems with quantification, the break-
down of the substitutivity of identity, and referential
opacity.

Another difficulty is posed by the use of numeri-
cally indefinite quantifiers. In (Def. 3), for example,
we find Grice using the curious quantifier 'at least some
(? many)‘. Such quantifications are common, but difficult
to formalize. If I say "Horses are swift," I have not said
"All horses are swift," nor "At least one horse is swift,"
nor "Your average horse is swift," nor "Most horses are
swift." Numerically indefinite quantifiers are present in
'A horse is a faster animal than a dog'. What shall we do
with these problems? It is wise in the context of the
present work to put them aside in the hope that they will
eventually be resolved in a way that does not undermine

the theory.60

Ii

51

William Alston is critical of both Leonard and
Grice. He regards their theory as a modern revision of the
traditional ideational theory. Alston's critique of the
ideational theory is that in identifying the meaning of an
utterance as an object in the mind of its author, it has
rendered communication impossible, granting that only the

61 But sometimes we do know

author can know his own ideas.
what other people are thinking, what they intend, and what
their purposes are, in an untechnical way of speaking. And,
in any case, the meaning of an utterance need not be thought
of as a mental entity, even if to mean1 is to do something
mental.

For Leonard the meaning is the signified purpose.
For Grice "to ask what U meant is to ask for a specifica-
tion of the intended response, E" that U would have A
make.62 (That is the response U intends to induce in A by
uttering I.) It is not clear that in either case we are
dealing with "mental entities." A purpose is in part a
proposition for Leonard. And, the response that Grice
speaks of need not be mental. Perhaps, then, Alston's
critique simply does not apply.

However, is it not curious that a purpose is to be
thought of simply as an attitude toward a proposition,
toward some possible state of affairs? Later we shall find
that Grice's list of intentions includes some odd members

as well. Grice will mention the intention that an

52

‘utterance-token should have a certain curious feature.63
To have a purpose is more than just to have an attitude
toward.some possible situation. Further, there are limits
to what we can reasonably intend. We can intend to induce
responses. But can we intend that objects already existing
have certain properties? At times we can. I can, for
example, intend to polish my car so that it will shine.

But at times we cannot. I cannot intend that the sun shine
tomorrow. Unless we are divine such "intentions" are, if
not impossible, at times pointless. I can hope, believe,
or doubt that a particular utterance-token I has property
I, but often I cannot intend it. I can intend to bring
about only what I believe to be within my power to bring
about.

Whatever their official positions are, both philos-
ophers seem to be aware that there is something more active
and directive to purposes and intentions. Their examples
reveal this. Purposes and intentions involve commitments
to doing something that will, in the speaker's estimation,
achieve whatever the speaker desires to accomplish. Let us
rely on this more ordinary and intuitive notion of intending
to see if these theories are plausible.

It seems, first, that they are not, as Alston sug-
gested, basically identical. The two philosophers do not
attempt to cover the same ground, as the outlines indicated.

They do, however, share some fundamental points. Both agree

53

that the analysis of meaningl can be made in terms of pur-
poses or intentions of speakers. But a purpose is not
exactly an intention. Some human acts are, it seems,
intentional but not purposeful. That is, they are done
deliberately but without a conscious purpose or end in view.
Often purposes and intentions look in different directions.
One seeks intentions when validating excuses, but one seeks
purposes when evaluating justifications and rationales.

Can meaningl be thought of as intending or pur-

posing? We can say
(14) I did it and I meant it.

This, generally, means that I acted deliberately or inten—
tionally, that I intended to do what I did. However, the
sense of 'meant' in (14) is not the sense of meaningl. I

can also say
(15) It was a meaningful event.

This might mean that the event has some personal importance,
perhaps that it has given me a sense of purpose. But the
sense of 'meaningful' in (15) is not the sense of meaningl.

Yet we would not allow
(16) This is the intention the author intended
as a proper paraphrase of

(17) This is the meaning the author intended.

54

When we tell someone to say they are sorry and mean
it, we are telling them to be serious about their apology.
We are not telling them to speak a certain phrase and have
a certain intention. The command 'Say it, and mean it!‘
is not so much a command to have certain purposes as it is
a command not to have certain other intentions or purposes.
It is a command not to speak in jest. Perhaps then, to
meanl is not exactly the same as intending or purposing.

But this theory need only maintain that to meanl is to
intend or to purpose in a certain way; not that every
intending or purposing is meaningl, nor that every sense of
'mean' can be explicated in terms of 'intend' or 'purpose'.

This theory requires, if issue I is to be success—
fully resolved, that meaningl be akin to purposing or
intending. Either they are all "mental acts" or dispositions
or propositional attitudes or whatever. To suggest, as I
have above in comparing (l6) and (17), that meaningl is not
exactly intending is hardly conclusive. The value of the
evidence presented by certain idiomatic expressions is
questionable. Also, the intending or purposing that mean—
ingl is akin to may be far more complex than first supposed.

Let us consider the hypothesis that to meanl is to
perform some act. What kind? Above we urged that the act
of meaning1 something must be distinguished from the act of
speaking or sayinga something. Likewise one can distinguish

speech acts, like those listed as examples of illocutionary

 

55

21nd perlocutionary acts, from the act of meaningl something.
IPerhaps this is why we do not answer the question "What are
you doing?" with 'I am meaning'. But we do not reply with
'I am intending' either. There is something odd about both
replies. We might answer with 'I have the intention to do
such-and-such'. But we would not use, generally speaking,
'I have the meaning to do such-and-such'. There again
appear to be dissimilarities between meaning1 and intending.

Perhaps it is wise to claim that neither meaningl
nor intending are kinds of acts. (Recall that Leonard
defined purposing as having an attitude toward some possible
situation.) One philosopher who does seem to think that
this theory of meaning entails that to mean1 is to perform
a mental act is Dennis Stampe. Stampe, who has defended
Grice's theory against Paul Ziff's criticisms, takes up
the question of whether or not to meanl is to perform a
mental act in his article "Toward a Grammar of Meaning."64
He concludes that meaningl is not any kind of a doing, thus,
not a mental act.

Stampe accepts a distinction between meaningn and
meaningnn which he falsely characterizes as Grice's.
Stampe's distinction is between the "factive" and the
"agentive" senses of 'mean'.65 Indeed, (10) and (11) tend
to support his View, but (12) simply does not fit his molds.
Curiously Stampe continues on to discuss 'mean' as an

intransitive verb.66

56

Eventually he argues that meaningl is not a kind of
doing even though the verb 'mean' may take agentive sub-
jects. He offers two reasons: We cannot say that the sub-
ject did something even though we have said that, by I, U
meant that p. We cannot answer the question 'What did U
do?‘ by saying 'U meant p' or 'U was meaning.‘ Further, we
do not use adverbs of manner to modify 'mean', except in
the case in which 'mean' is being used in some other sense,
as in 'He meant well.‘ U cannot be said to have meant p
rudely or angrily or vaguely or softly, although p may be
vague or rude, or U may have been angry or have spoken
softly.67

Stampe argues, on Grice's behalf I believe, that
sentences like 'Smith meantnn that p' do not report actions
but they do "explicate something" Smith did or said. They
iflElX.that Smith did something and perhaps say with what
intentions it was done or said. Stampe believes this is so

because sentences like
(18) 'Covert' means veiled

might be thought of as grammatical transformations from

agentive contexts like
(19) by 'covert' one means veiled.

Stampe claims that 'meannn' takes agentive subjects,

whereas 'meann takes factive subjects. Grice, on the other

hand, has used both kinds of subject with 'meannn' (as I

57

iruiicated above and will indicate in the next chapter). We
kxunn that both meaningl and meaning2 are discernible in
(Irice's meaningnn, meaningl being agentive and meaning2
lxaing factive. Thus it would not be surprising for Grice
supporters to argue that one can derive (18) from (19).
'Fhey would think that both rely on the same sense of 'mean',
viz., meannn. But the move cannot be made, in this I agree
imith Mr. Stampe. The move from (19) to (18) involves a
jump from one sense of 'mean' to another. We both see it
as a move from the agentive to the factive sense, I regard-
ing it as a move from meaningl to meaning2 having nothing
to do with meaningn, he as a move from meaningnn to meaningn.
Stampe presents the crucial derivations:

(20) 'Jim meansnn y.by I' goes into 'y is what is meant

by I' and finally into 'y is what is meantnn'.56

nn

It should be an easy jump for Grice to arrive at 'y is what
I meansnn' given (20). Stampe does not, however, regard

'y is what I_means' as a legitimate transformation of 'y is
what is meant'. But we did not expect that it would be
given that 'meant' in 'y is what is meant' is used in the
sense of 'meanl' and not in the sense of 'meanz'. Yet his
is the sense of 'means' in 'y is what I means'. On the
other hand Stampe is prepared to call it a supportable
hypothesis that 'y meansnn I' can be derived from 'By I

agent meansnn y'. (He thus further confuses the senses of

'mean', now allowing the "agentive 'meannn'" to take a

 

58

factive subject. But if the transformation can be made,
this is what will result.) He argues that the hypothesis
is supported by the derivations cited in (20) as well as
that similar derivations fail when factive expressions are
used with 'meann'.68 I must confess that I fail to see how
either contention supports the hypothesis, especially given
our two senses of 'mean' discernible in Grice's 'meannn'.

These considerations emerge: First, if Stampe is
right, meaningl is not a kind of doing, not a mental act.69
Second, his views are interesting in their syntactical com-
parisons of 'meaningl' and 'meaningz' but irrelevant. This
theory asserts only that meaningl is akin to intending or
purposing. It need not hold that any of these are mental
acts. Third, even those who might be regarded as sympa-
thetic to this theory find it hard to move from expressions
like 'U means I by I' to expressions like 'I means y'.

This indicates that Grice and Leonard may not succeed in
issue II. The relationship between meaning2 and meaningl
will not be easy to determine; a complete reduction may be
impossible.

We have before us a number of useful distinctions
and a breakdown of the issues involved. The two known
versions of this theory have been outlined and the theory
itself briefly sketched. Some prima facie problems have

been exposed and discussed. In several cases they were not

of immediate concern, inconclusive, or not to the point.

59

Some have cast doubt on the theory, but this can be over-
come provided the philosophers can treat the two major
issues adequately. The second issue would be a major under-
taking, as the outlines above indicated. But we shall not
require a discussion of the relationship of meaning2 to
meaningl. In the following chapters there is a critique

of Grice's efforts to adequately analyze meaningl by listing
only intentions of a speaker. Since Leonard offers no
defense for his assumption that such an analysis is possible,
we have only Grice's work to review. In doing this problems
are revealed which become insurmountable obstacles to the
successful resolution of issue I. Let us begin our examina—
tion of this issue with a presentation of Grice's first

analysis.

 

FOOTNOTES TO CHAPTER I

1Note for example George C. Kerner's The Revolution

in Ethical Theo y (New York: Oxford University Press,
1966); as well as Ludwig Wittgenstein, Philosophical Inves-
'tigations (New York: Macmillan, 1953); V. C. Chappell, ed.,
Ordinary Language (Englewood Cliffs, N.J.: Prentice-Hall
Publishers, 1964); and A. J. Ayer, ed., Logical Positivism
(New York: The Free Press, 1959).

 

 

 

2William P. Alston, "Meaning," in Encyclopaedia of
Philosophy, ed. by Paul Edwards (New York: Macmillan and
The Free Press), V, pp. 233-41; Philosophy of Languagg
(Englewood Cliffs, N.J.: Prentice-Hall Publishers, 1964),
p. 107.

 

3G. H. R. Parkinson, ed., The Theory of Meaning

(New York: Oxford University Press, 1968?, p. 184.

4Thomas M. Olshewsky, ed., Problems in the Philos-
ophy of Language (New York: Holt, Rinehart and Winston,
Inc., 1969), chapter 3.

5William P. Alston, "Linguistic Acts," American
IUIlosUphical Quarterly, I (1964), 138-46; P. F. Strawson,
"Intention and Convention in Speech Acts," Philosophical
Review, LXXIII (1964), 439-60; Paul Ziff, "On H. P. Grice's
Account of Meaning," Analysis, XXVIII (1967), 1-8;
T. E. Patton and D. W. Stampe, "The Rudiments of Meaning:
On Ziff on Grice," Foundations of Language, V (1969), 2-16;
N. L. Wilson, "Grice on Meaning: The Ultimate Counter-
example," Nous, IV (1970), 295-303.

6Max Black in correspondence, May 4, 1970.

7John R. Searle, Speech Acts: An Essay in the
Philosophy of Languagg (Cambridge: Cambridge University
Press, 1969).

8D. S. Clarke, "Meaning, Force, and Rhetorical

Effect," abstracted in The Journal of Philosophy, LXVII
(1970), 828-29.

 

 

 

9Henry S. Leonard, Principles of Right Reason
(New York: Henry Holt and Company, 1957); "Interrogatives,
Imperatives, Truth, Falsity, and Lies," Philosophy of Science,

 

60

61

XXVI (1959), 172—86; “Authorship and Purpose," Philosophy
of Science, XXVI (1959), 277-94; "Authorship of S1gns,"
Papers of the Michigan Academy of Science, Arts, and Letters,
XLV (1960), 329-40. H. P. Grice: "Meaning," Philosophical
Review, LXVI (1957), 377—88; "Utterer's Meaning, Sentence—
Meaning, and Word-Meaning," Foundations of Language, IV
(1968), 225-42; "Utterer's Meaning and Intentions," Philo-
sophical Review, LXXVIII (1969), 147-77.

loCf. W. V. O. Quine, Word and Object (Cambridge:
M. I. T. Press, 1960), 129-34.

 

11James A. H. Murray, Henry Bradley, W. A. Craigie,
C. T. Onions, eds., The Oxford English Dictionary_(New York:
Oxford University Press, corrected and reissued 1961), VI,
273. Entries concerning 'mean' are on pages 269—274.

12J. L. Austin, How to Do Things with Words (New
York: Oxford University Press, 1965), p. 98.

13

 

Ibid., p. 146.

l4Ibid., p. 108.

lSIbid., p. 148.

16L. Jonathan Cohen, "Do Illocutionary Forces
Exist?" Philosophical QuarteIIy, XIV (1964), 118—37.

17For a more complete list of illocutionary acts, or
verbs that denote illocutionary forces, see Lecture XII of
Austin's How to Do Things with Words, op. UII. William P.
Alston also provides lists of’illocutians and perlocutions
in his Philosophy of Language, Up. UIU., p. 35.

18J. L. Austin, Up. cit., p. 120.

 

19Ibid., p. 113.

201bid., pp. 120, 121.

21lbid., p. 118.

22Ibid., pp. 99, 149.

231bid., p. 127.

24Ibid., p. 121.

251bid., p. 115.

26Ibid., p. 117.

62
27Ibid., p. 116.

28Ibid., p. 101.

29H. P. Grice, "Utterer's Meaning, Sentence—Meaning,

and Word-Meaning" ("UMSMWM"), Up. cit., p. 225. This
section outlining Grice's program is drawn from pages 225-

229 of

"UMSMWM."

30H. P. Grice, "UMSMWM," 9E: cit., p. 228.

31H. P. Grice, "Utterer's Meaning, and Intentions"

("UMI"), gp. cit., p. l60f.

32H. P. Grice, "UMI," Up. cit., p. 150.

33H. P. Grice, "UMSMWM," Up. cit., p. 230.

34Ibid., p. 229f.

35Ibid., p. 233.

36Ibid., p. 234.

37Ibid., p. 234f.

381bid., p. 235.

39See "Group Three," in Chapter II, infra.

40H. P. Grice, "UMSMWM," 9p. cit., p. 236f. (My

RP2 is Grice's RP4.)

41H. S. Leonard, "Interrogatives, Imperatives,

Truth, Falsity, and Lies" ("IITFL"), Philosophy of Science,
XXVI (1959), 173.

tion.

 

42Ibid., p. 174.

43Ibid., p. 175.

44Ibid., p. 176. My italics.

45IUIU. This is a paraphrase of Leonard's defini-

46H. S. Leonard, "Authorship and Purpose" ("AP"),

SE: cit., pp. 286-88. And "Authorship of Signs" ("AS"),
Up. c1t., pp. 334-35. This definition, like the next, is
only a rough approximation of Leonard's.

47H. S. Leonard, "AP," 92: cit., p. 288f.

63

48£§19., p. 289; and "AS," Up. cit., p. 236.

49H. S. Leonard, "AP," Up. cit., p. 277.

5°Ibid., p. 291.

51Ibid., p. 293.

52H. S. Leonard, "IITFL," Up. cit., p. 178.

53In "IITFL" Leonard claims that every sentence has
a truth—value. Several philosophers have commented on this
position. In general the major problem seems to be that
all indirect questions have the same truth-value which they
derive by virtue of being indirect questions, and not by
virtue of the proposition (or propositional function) that
they indicate. There are other comments as well. The
articles of primary relevance are: (1) H. S. Leonard,
"IITFL," Up. SEE: (2) J. M. O. Wheatley, "Note on Professor
Leonard's Analysis of Interrogatives, etc.," Philosophy of
Science, XXVIII (1961), 52-55. (3) H. S. Leonard, "Reply
to Professor Wheatley," Philosophy of Science, XXVIII
(1961), 55—64. (4) C. L. Hamblin, "Questions Aren't State-
ments," Philosophy of Science, XXX (1963), 62-63. And (5)
Gerald Stabl, TrA Review of the Leonard-Wheatley-Leonard-
Hamblin Debate Over the Attribution of Truth-values to
Interrogatives," Journal of Symbolic Logic, XXXI (1966),
666-68. Two other articles of interest are: (6) C. L.
Hamblin, "Questions," Encyclopaedia of Philosophy, Up. UIE.,
VII, 49-53; and (7) Nuel D. Belnap, Jr., "Questions: Their
Presuppositions, and How They Can Fail to Arise," in IUU_
Logical Way of Doing Things, ed. by Karel Lambert (New
Haven: Yale Press, 1969), (essays in honor of Henry
Leonard). The general thrust of the discussion is that
Leonard was mistaken. It is unfortunate that Leonard's
commentators focused on this minor aspect of his philosophy.

54

 

 

 

 

 

 

H. S. Leonard, "IITFL," Up. cit., p. 172.

55H. S. Leonard, Principles of Reasonipg (New York:
Dover Publications, 1967). (This work is a revision of
H. S. Leonard's Principles of Right Reason, Up. cit.),
p. 157.

 

 

56Ibid., p. 146.

57H. S. Leonard, Principles of Right Reason, Up.
cit., p. 138. See also Principles of Reasoning, Up. cit.,
p. 138.

 

 

58Stanley Cavell, "Must We Mean What We Say?"
Ordinary Language, 923 cit., p. 102. Of course, it is

 

64

;possibde that there is really no mistake here, for Cavell
(and Austin too below) may be using the ambiguous word

Vmeanf in quite a different sense. He might be using it in
the sense of 'imply'.

59J. L. Austin, "A Plea for Excuses" (the Presi-
dential Address to the Aristotelian Society, 1956),
IProceedings of the Aristotelian Society, XLII (1956-1957),
1-30. Reprinted in Ordinary Language, op. cit., from.which
this citation is taken, p. 36. __ '__—

60Leonard and Grice are comfortable putting aside

these problems in this way. Leonard, for example, says
concerning the introduction of 'Tap' (a formal device read
"U_thinks that p"): "To be sure the introduction of this
formula raises all sorts of fundamental epistemological
issues that cannot even be touched upon in this paper. Its
use must be acknowledged as a temporary device, aiding in
the solution of certain problems, but doomed to replacement
by a more adequate symbolism when the problems of episte-
mology are directly treated. It is only to be hoped--and it
is expected--that its eventual replacement will not invali-
date conclusions reached in this paper by its use." ("AP,"
22: UIU., 285.) See also the "Preface to the Dover Edi-
tion" (page v, especially), PrincipIes of ReasoniUg, Up.
cit., for Leonard's acknowledgement of certain metaphysical
and logical problems relating to propositions and senten-
tial utterances. Grice also is aware of some of these prob-
lems. He is, rightly, opposed to aprioristic restrictions
against the use of "any intensional notions or devices
which seem to be required in order to solve one's conceptual
problems, at least at a certain level, in which (metaphys—

ical bias apart) reason and intuition command." ("UMSMWM,"
Up. cit., 242.)

61William P. Alston, Philosgphy of Language, 923
cit., pp. 22-25.

62H. P. Grice, "Meaning," as reprinted in Philo-

sophical Logic, ed. by P. F. Strawson (New York: Oxford
University Press, 1967), p. 46.

63H. P. Grice, "UMI," Up. cit., p. 163.

64Dennis W. Stampe, "Toward a Grammar of Meaning,"
Philosophical Review, LXXVII (1968), 137-74. T. E. Patton
andfiD. W. Stampe, wThe Rudiments of Meaning: On Ziff on
Grice," Up. UIU.

65D. W. Stampe, "Toward a Grammar of Meaning," Up.

cit., p. 138.

 

65

66Ibid., pp. 147-48. The verb in its relevant sense

iAS'transitive in both the Oxford English‘DictionaIy_(Up,
<sit.) and the Webster's New Collegiate Dictionapy
(Springfield, Mass.: Merriam Co. Publishers, 1961 edition).
Stampe however claims "dictionaries generally mark the dis-
'tinction between the two senses of the verb by labeling
Fmeanf transitive in its non-natural sense and intransitive
111 its natural sense." Webster's New Collegiate Dictionary
inakes 'mean' intransitive when it means "to have a purpose
or intention" and "to be of a (specified) degree of impor-
'tance in influence or effect, as environment means much to
a child." The sense of 'mean' in the Webster example is
not.the sense of Grice's 'mean', yet the subject expression
is factive. In the Oxford English Dictionary 'mean' is
intransitive in the following senses: T'to be (well or ill)
intentioned or disposed," "to hold or entertain an opinion;
to think, imagine," and "to speak, tell." One sense of the
verb, the only one which would support Stampe's argument at
all, is, sadly, left unmarked. The dictionary does not
indicate whether 'mean' in sense (c) ("of things, words,
statements: to have a certain signification; to signify or
import; to portent") is intransitive or not. Perhaps this
is because in this sense it might be used either as transi-
tive or intransitive. In any case. Mr. Stampe's comments
about 'mean' do little to support his claims and distinc-
tions because the comments seem generally false.

67D. W. Stampe, "Toward a Grammar of Meaning,"

Up. cit., p. 142.

68Ibid., p. l65f.

69N. L. Wilson, 923 cit., pp. 300-2. Wilson accepts

Stampe's arguments that to mean is not to UU anything.

Wilson goes on to argue that on any occasion when a person
speaks he is speaking in accord with the meaning conventions
of a language. (This generalization seems, rather obviously,
false.) The words have already been bestowed with meaning(2)
He only acts in accord with these conventional bestowals.
Wilson neglects the question of how it came about that these
words mean what they do mean . Wilson's position is not
inconsistent with this theory of meaning (although Wilson
thinks that it is), for meaning might be reduced to meaning1
or derived in some way from meaning . These conventions
reflect what we mean by our words, thus determining what

the words meafig and 0, what an individual should mean1 by
these words.

 

CHAPTER II

GRICE'S "MEANING"

Introduction

 

H. P. Grice's "Meaning" is the best known presenta-
tion of the theory of meaning as intention.l It appeared

in 1957, the same year as Leonard's Principles of Right

 

Reason. In this article Grice offers his first attempt at
the analysis of meaningl in terms of intentions. Grice's
work on the problem of meaning spans two decades. As early
as 1952 we find hints that he was working on a theory of
meaning; his latest efforts appeared as recently as 1969.
The first clue that Grice was developing an original theory

is found in a review of John Holloway's Language and Intel—

 

ligence.2 This review was written by another Oxford philos—

3 .
Hart notes conversat1ons

opher, H. L. A. Hart in 1953.
with Grice and sketches a rough version of a theory of
meaning that is in the spirit of the analysis Grice would
present five years later.

Hart says that for one to claim that he understands

someone's statement it is both sufficient and necessary that

he recogpize from the utterance what the speaker intended

 

 

him to believe or do.4 According to Hart, one cannot
logically say "I understood what he meant by saying there

will be rain but I do not know what he intended me to

66

67

believe."4 Hart's claim seems true, but it does not entail
that meaningl can be defined by listing only intentions of
the speaker. It does indicate that whatever one lists in
the analysans, if it is adequate, it will mention some
intention of the speaker.

Hart offers an analysis of meaningl. In listing
three necessary conditions for a person's meaning something
he does not restrict himself to listing only that person's

intentions. According to Hart, A must:

(H1) utter noises which are in fact interpreted as signs,
(H2) intend that they be interpreted as signs, and
(H3) intend that the listener recognize from the utter-

ance that the speaker intended the listener
to believe or do something.

Hart's analysis, which deserves, but will not receive,
careful examination, will be reworked and presented by Grice
in "Meaning." Note, in passing, that Hart's analysis makes
"A's meaning something" depend in part on whether or not

A's noises are interpreted as signs. Not only does this
seem to require an interpreter other than the speaker, if

it is not to be a trivial clause; but it makes all utter-
ances meaningless which, for one reason or another, are not
thought of as signs.

There is some relationship between meaningl and
intentions. If I understand what a person, A, meansl by
uttering I, then I do know what he intends me to think. The
relationship seems to be a logical one, thus the possibility

of providing a definition of meaningl at least in part in

68

terms of A's intentions. Grice's problem is to provide the
analysis listing only intentions so that one can be said to

mean something if and only if one has the intentions in

1
this list.

In the case of his first analysis, Grice did not
take notice of the meaningl/meaning2 distinction. He seems
not to have, therefore, been prepared to take up issue II.
Furthermore Grice did not utilize any distinction like the
illocutionary/perlocutionary distinction. A's illocutionary
intentions must, in general, be in accord with the meaning2
of the utterance-type used, if they are to be realized.

Yet A's perlocutionary intentions can vary independently of
the conventionally determined illocutionary force of
uttering I. Also, the perlocutionary intentions can vary
independently of the illocutionary intentions. We can dis-
tinguish between what A did IU sayingi I and what results

he intended to accomplish Uy sayingEL I. Grice's analysis
seems to fit well the perlocutionary intentions A might have,
but it does not handle smoothly A's illocutionary intentions.
Yet meaningl is best associated with illocutionary inten-
tions. In trying to explicate 'By uttering I Ameantl some-
thing' in terms of some expression like 'By uttering I_A
intended . . .', Grice is committing himself to the reduction

of illocutionary intentions to perlocutionary intentions.

The two kinds are assimilated in the explicandum. Grice

69

offers no reason for his disregarding this distinction.
N. L. Wilson says:

At Oberlin I taxed Grice with this point. His reply,

if I remember it correctly, was: "But that's my thesis!

I may be mistaken, but I'm not confused."5

Grice's article, "Meaning," attracted much attention
from philosophers who attacked, defended, or revised Grice's
analysis of meaning. This chapter is a commentary on the
debate this article generated. A decade passed befOre
Grice responded to the wealth of criticism his work elicited.
"Utterer's Meaning and Intentions" replies to his critics
and offers a new analysis of meaningl in terms of intentions.
This article will be discussed in Chapter III, along with
relevant sections of "Utterer's Meaning, Sentence-Meaning,
and Word—Meaning."7
Grice's article is troublesome. It is often vague

and elliptical. The examples do not always warrant the
interpretation Grice gives them, nor foster the theory he
bases on them. After devoting a section to an exposition
of Grice's article, six sections will be devoted to its
repercussions. Each section treats of a class of problems
directed against this theory and specifically Grice's 1957
analysis. These sections do not follow an historical
sequence. Rather, each section represents a group of prob-
lems that are increasingly more serious or more pressing

with regards the adequacy of the analysis of meaning1 in

terms of intentions. The first group does not represent a

70

serious threat to this theory. Some of the difficulties
presented in Chapter I are catalogued with this group. The
second group is more serious, but the problems it contains
are more sweeping in scope. The third group contains prob-
lems that affect issue II. The fourth group contains dif-
ficulties which Grice may be able to overcome, but which
are serious difficulties for the present analysis. The
fifth group of problems is based on Grice's disregard of
the illocutionary/perlocutionary distinction and are, I
believe, insurmountable difficulties for the Gricean kind of
analysis of meaningl. The final group of problems suggests
sound reasons why one must include more than intentions in
one's analysis of meaningl. It will become obvious that in
some cases a problem should be classified under more than
one group. The summary section will list all the problems

classified according to merit and importance.

Meaning"

Grice distinguishes between two general senses of
the word 'mean'. 'Mean' is used in its "natural" sense in
"Those spots didn't mean anything to me, but to the doctor

8

they meant measles." The "non-natural" sense of 'mean'

(meaningnn) is exemplified in the following:

(1) Those three rings on the bell mean that the bus is
full.
(2) That remark, "Smith couldn't get on without his

trouble and strife," meant that Smith found his wife
indispensable.

71

'Means' in any sentence of the form "A means to do so-and—so

 

by . is a natural use of 'means'. Thus 'meant' in
"John meant to fool the defense by running on third down and
long yardage" is being used in its natural sense. Note that
here Grice's assimilation of illocutions and perlocutions
causes him to say something true of perlocutions but false
of illocutions. For, note that the "so-and-so" may be the

illocutionary force of uttering I or the perlocutionary

effect sought. "Any senses of 'mean' found in sentences" of

 

the pattern 'A means something by I' are non-natural senses

of 'mean'.9 (My italics.) Grice would allow that the
sense of 'meant' in (2') is non-natural.
(2') Smith meant that his wife is indispensable by "I

can't get on without my trouble and strife."

Meaning2 is exemplified in (l) and (2); meaningl in (2').
Both are senses within Grice's meaningnn. Grice claims that
given either (1) or (2) one may infer what somebody meantnn,
or, at any rate, should have meantnn by such-and-so.9 The
inference to what a person should have meant1 is possible
because of the relationship between illocutionary forces

and meaningz. When one intends to achieve a certain illo-
cutionary force one relies on conventional meaningz. Note
that one cannot infer from (2') what Smith should have
meantnn. Grice assumes, wrongly, that given the meaningnn

of I one can infer both what A should have meantn by

n

uttering I and what I meansnn. This assumption is true only

72

in the case of meaningz, but not necessarily in the case of

meaningl. Knowing "what Imeansl (where this is short for
"what Ameantl by I on this particular occasion"), one

cannot infer either what A should have meant or what I

1
meansz. Grice notices this problem and provides for it in
his recent analysis of meaningnn. However, even there the
relationship between the meaningnn of utterance-types to
the meaningnn of their tokens is not perfectly clear.

That Grice is committed to the reduction of mean—
ing2 to meaningl is made evident in his critique of causal

theories of meaning.10

At one point he comments that a
causal theory can give us, at best, only a statement about
the "standard meaning, or meaning in general" of a sign,
it cannot deal with what a particular speaker or writer
means by the sign on a given occasion." Grice adds the
interesting comment that "the causal theory ignores UUU

fact that the meaning (in general) [meaningz] of a sign

needs to be explained in terms of what users of the sign do

 

(or should) mean [meanl] by it on a particular occasion."11
The matter of this reduction is not our central concern.
We must examine, first, the analysis of meaningl.

To develop the definition of meaningnn Grice

examines the significance of certain paradigm expressions:

(G1) I meantnn something (on a particular occasion),
(G2) I meantnn that so-and—so (on a particular occasion),

(G3) A meantnn something by I (on a particular occasion),

73

(G4) A meant by I that so-and-so (on a particular
occasio ,

(G5) I meansnn (timeless) something (that so-and-so),
(G6) Ameansnn (timeless) by I something (that so-and-
so).

The goal is to specify a set of conditions for the truth of
(G6). Grice regards the elucidation of the first four
expressions as the key to understanding (G6).
Grice rejects the first suggestion that (G1) and

(G2) would be true merely because I was intended to produce
an effect. At first Grice confines his attention to cog-
nitive effects like producing beliefs. He offers an example
of what Leonard would call a "non-candid productive sign
event" to counter the suggestion.

Example 1 (Grice): I might leave I's handkerchief near

the scene of a murder in order to induce the detective

to believe that U_was the murderer; but we should not

want to sgy that the handkerchief (or my leaving it

thereTmeantnn anything or that I had meant n by leaving
it that I was the murderer.10 (My italics.?

 

 

Grice uses the italicized criterion frequently. It
is not a simple generalization for philosophers do not
survey speech behavior to determine what to say about their
language. As native speakers they behave in accord with,
and often can make explicit the conventions that constitute,
their language. Determining what we should or should not
want to say depends on how well we know how to speak the
language, failing a complete and adequate theory. Native
speakers have a sense of, or a feel for, the logic of the

language and the meanings of its expressions that justifies

2

74

their making claims like Grice's, in the absence of research
and formal studies. Such research would take precedence,
however, since an individual's intuitions tend to reflect
his idiolect and thus may or may not be a true representa-
tion of the language itself.12

On the basis of the detective example Grice adds
that the speaker must not only intend to produce a belief in
his hearer, but further, he must intend that the hearer
recognize the first intention. Grice's analysis requires
intending a double reading of the hearer. He must recognize
the speaker's intention to produce some effect, and he must
recognize what that intended effect is.

What this analysis lacks, according to Grice, is a
further clause which will separate cases of "telling" from
cases of merely "getting someone to think" something. Grice
is trying to separate saying something to someone from
merely arranging matters so that someone comes to believe
something. I believe that Grice is trying to distinguish
between achieving something as a perlocutionary effect and
achieving someone as an illocutionary result. Although Grice
sees the problem, he is unable, as his critics will argue,
to resolve it.

Grice attacks it by considering these three sen-

tences:

(3) Herod presents Salome with the head of St. John the
Baptist on a charger.

75

(4) Feeling faint, a child lets its mother see how pale
it is (hoping that she may draw her own conclusions
and help).

(5) I leave the china my daughter has broken lying

around for my wife to see.

Herod satisfies the two conditions. He intends that Salome
believe St. John the Baptist is dead; further, he intends
that she recognize that he wants her to believe this. But
Herod has not told her that he is dead, he has merely let
her know. If it were not for (1) we might think that Grice
was trying to restrict meaningnn to linguistic communications.
But further examples confirm the breadth of Grice's intended
analysandum.

He offers two cases for comparison, in an effort to

isolate what is needed to complete the definition of mean—

1ngnn.

(6a) I show Mr. I a photograph of Mr. I displaying undue
familiarity to Mrs. I.

(6b) I draw a picture of Mr. I behaving in this manner

and show it to Mr. I.

The photograph is a natural sign to Mr. I. He reads it as
meaning (natural) that his wife has been unduly familiar
with Mr. I. Grice says that neither the picture nor my

showing it meantn anything. Suppose we added an element of

n
conventionality to this under—described example. Suppose
Mr. I suspected that there was something between his wife

and Mr. I. Suppose he hired me, a private detective and

photographer, to check out his suspicions. After some time

76

on the case he and I meet and I hand him the photograph
without saying anything. This photograph is all the proof
he expected, it is all that he hired me for. Could we not
count my showing Mr. I the photograph as meaningfulnn? By
showing him the photo I informed him that he was right
about his wife and Mr. I.

According to Grice there is an important difference
between (6a) and (6b). In (6a) nothing has been meantnn.
Grice claims that this is so because Mr. I is not led to
his conclusion about his wife and Mr. I by reason of my

intentions. But something is meantn in (6b) because there

n
Mr. I is led to his conclusion at least in part because of
his recognition of my intentions.

Grice's analysis is faulty in at least two respects.
First, whatever suggests my intentions to Mr. I in (6b) may
be operative in the case of (6a) as well. Photographs do
not speak for themselves when situations like the above-
described (6a) obtain. Further, Mr. I is led to his conclu—
sion not so much by his recognition of my intention as by
his estimation of what my intention might be. Mr. I merely
thinks that my intention is to inform him of his wife's
behavior, rather than, say, to inform him of my talents as
an artist. I may not have the intention my hearer thinks I
have. He will be led, not by my intentions, but by his
estimation of what they are. Knowing this, I can anticipate

his estimations of my intentions, and rely on his failure to

 

77

estimate correctly or completely to avoid communication and
yet succeed in achieving some intended (perlocutionary)
effect. As we shall see, this is the key to so much of the
criticism of Grice's definition.

If anything is obscure in this article, it is the
exact analysis of meaning that Grice intends to offer.
Grice offers two initial statements of his analysis, both
are given in an unnecessarily restricted form.13 In the
first he claims that three conditions are necessary condi-
tions for "A to mean something by I." These are: (1) "A
must intend to induce by I a belief in an audience," (2) "he
must also intend his utterance to be recognized as so
intended," and (3) "the recognition is intended by A to
play its part in inducing the belief."13 In the second
analysis Grice suggests that 'A meant something by I' is
"roughly equivalent" to 'A uttered I with the intention of
inducing a belief by means of the recognition of this
intention'.l3 Both of these mention A's intention to
induce some belief in his hearer, I. But this can be
generalized to A's intention to induce almost any kind of
effect in I. Grice's final generalized analysis is: "'A
meant something by I' is (roughly) equivalent to 'A intended
the utterance of I to produce some effect in an audience by
means of the recognition of this intention'."l4

Obviously the logical structure of "the Gricean

analysis" is less than perspicuous. It appears that Grice

is not even confident of the analysandum!

78

As to the

analysans, it seems to be the conjunction of three clauses

each of which refers to an intention of the speaker, but it

actually is the existential generalization of the existen-

tial generalization of such a conjunction.

tive terms Grice intends to employ are not obvious.

Even the primi—

In

any case, I offer the following as a reconstructed version

of Grice's original analysis:

(M)

By uttering x A means something if and only if

 

there is some audience, I, and some effect U, such

that:
(i ) A intends
(12) A intends
(i ) A intends
some part

to induce U in I

that I recognize that (i )

that B's recognition of iil) play
I '— 0

in I s productlon of I.

Now that the logical form of the analysis is before us it is

apparent that Grice is committed to the existence of both

an audience and an effect.
ment later.

primitive terms:

We shall return to this commit-

It is also apparent that Grice relies on several

(a) _I_ intends to _I_

(b) _I_ intends that _I_

(c) _I_ induces in _I_

(d) _I__recognizes that _I_
(e) _I_ produces _I__

(f) _I_ plays some part in _I_

I suppose that the first five of these are tolerable, but

the last is surely too vague.

Fortunately, in his notes on

79

the nature of U Grice suggests something that clarifies
what "part" I's recognition of (11) is to play.

Grice puts only two, rather vague, restrictions on
the kind of effect, 2! that A can intend to induce or pro-
duce in I. A must think that it is, in some respect, within
I's control to produce 3. And, A_must believe that I's
recognition of (i1) can function as a reason, in some sense
of the term, for I's production of I. By this Grice wishes
to exclude the case in which I is a direct causal result of
I's recognition.l4

Suppose that A has the three intentions listed in
(M). And, suppose that there is some audience I and some
effect I. Suppose also that I recognizes that A has (11),
thus fulfilling A's (12). I might still not respond with
the production of an instance of U. Yet A would have
meantnn something by uttering I, according to Grice's
analysis. This analysis is not, as Parkinson has suggested,
a causal theory of meaning.15 A's utterance of I is mean—
ingfulnn even if the response (effect) intended in the first
place is not forthcoming.

If Ameantnn something by uttering I, what did he
meannn? He meantnn p.14 This curious reply is what Grice
would have us respond with for "to ask what A meant is to
ask for a specification of the intended effect." Further,

from 'Somebody meant something by I' one can infer 'I meant

something'. This is so, according to Grice, because the two

80

are, roughly speaking, equivalent.l7 Grice suggests that

whenever we are speaking about meaningnn with regards to
utterances, we are making "some sort of reference to some-
body's intentions."l4 We might equate 'Imeansnn that
so—and-so' with "some statement or disjunction of statements
about what 'people' (vague) intend (with qualifications
about 'recognition') to effect by I."14
Anticipating objections Grice indicates that the
UUIy effect associated with the meaningnn of an utterance is
the "immediate effect" that was the speaker's primary inten-
tion. AUy effect, I, that "results entirely" from U is not

18

a part of the meaningnn of I. Grice disclaims "any inten—

tion of peopling all our talking life with armies of com-

18 He claims that no

plicated psychological occurrences."
new problem is created by speaking of linguistic as opposed
to non-linguistic intentions, and this seems plausible.
There is no more difficulty in telling what someone's
intentions are from his non-verbal behavior than there is in
telling them from his speech. In UUIU cases a person can
often make his intentions clear. Failing this the speaker
is "held to convey what is normally conveyed (or normally
intended to be conveyed), and we require good reason for
accepting that a particular use diverges from the general

..19
usage.

81

‘GrOup One

 

This group of problems is comprised of those objec-
tions suggested or advanced by philosophers which simply
are misguided or based on misunderstandings of this theory
or Grice's first analysis. Several candidates for this
group were presented in Chapter I. For example, in Chapter I
we noticed that this theory of meaning does not support
(Thesis 2). Thus, an objection like the following is mis—

guided.

Problem A Because this theory allows that an expression
may mean whatever its author wishes it to mean,
this theory threatens to destroy the "objec-
tivity" of the meanings of our expressions,
which objectivity is the foundation for our com-
munication.

 

Not only does this objection overlook the meaningl/meaning2
distinction, but it also misses the practicality of this

theory. People may mean whatever they wish, but tokens, by

1

virtue of being of a certain type, mean what the type

2

meansz. (Perhaps that is what people in general mean by

1

expressions of that type.) Thus, if a speaker is practical,

Thus, Problem A

 

he will meanl what his expressions mean2.
can be disregarded.
Other objections already discussed result from mis-

classifications of Grice's theory. Grice‘s theory is not a

"causal" theory as Problem B suggests.

 

Problem B This theory is unacceptable because it is a kind
of causal theory of meaning, and thus is subject
to Grice's own critiques of such theories.

 

 

82

Grice has argued that causal theories are unacceptable
because they "ignore" meaning1 and because in the event of
the non-occurrence of the effect, 3, the token uttered will
fail to meannn.ll Surely Grice's own theory does not ignore
meaningl. Nor is it subject to the latter critique. Sup-
pose E is an effect-type. A may intend that gnbe instan-
tiated in I, but I may or may not act in accord with Afs
intentions. Nevertheless A will have meantnn something by
uttering I if he had the three intentions, one of which
being to induce (an instance of) U in I. Further, I will
meannn something because A_meantnn something by uttering it.
Not only does Grice's theory escape the objection of IIUUf

lem B, but Problem B is misguided. Grice's theory is not a

 

"causal" theory, at least, Grice disclaims requiring that U
be instantiated as a causal result of I's recognition of A's

(i ). Thus we can put aside Problem B.

 

1

William Alston misunderstands Grice's theory. He
makes it out to be an ideational theory and thus it becomes
heir to those objections that could be brought against more

traditional ideational theories.

Problem C This theory is a modern version of the ideational
theory of meaning, thus it is unacceptable for it
renders communication impossible.

 

But, as we saw, Alston's objection did not apply. Grice's
theory is not committed to the existence of mental entities
or ideas but to such claims as "people intend" and "people

recognize or understand." Grice has not, as the logical

83

form of his analysis indicates, strictly speaking, even
committed himself to intentions. Nor does he identify "the
meaning of I" as a mental entity. We can disregard IIUU-
IUI_U, then.

Another objection discussed in Chapter I was that
objection we associated with Stampe and, in the footnotes,

Wilson. These philosophers argued:

Problem D This theory of meaning presupposes that to mean
is to perform some kind of mental act. But,
this presupposition is false since to mean is
not to perform any kind of an act.

 

In Chapter I we observed that this objection is without
merit for the presupposition it alleges that this theory
makes is not, in fact, required by this theory. All that is
essential is that to meanl is to intend (or to purpose),
perhaps in some complex way. We need not hold that intend-

ing or purposing or meaningl are acts at all. Thus we can

disregard Problem D as well.

 

There are other misguided objections besides these
four, but they will generate problems that are better clas-
sified under subsequent headings. Let us turn, then, to

the next group.

Group Two

 

This group contains problems of a more sweeping
scope. These problems affect not only this theory of mean-
ing, but several other philosophical enterprises as well.

In a way, these problems reflect some of the philosophical

84

presuppositions of this theory of meaning. For example,
the ontology of this theory of meaning is problematic. We
noticed in Chapter I that both versions of the theory were

committed to propositions. In my reply to Problem B I

 

suggested that we should assume that U ranged over univer—

sals, viz., effect-types. This suggests:

Problem E This theory is objectionable because it presup-
poses the existence of metaphysical entities,
viz. propositions and universals.

 

Problem E is beyond the scope of the present work, so, let

 

us put it aside as was suggested in Chapter I.

Another less pressing but rather interesting diffi-
culty is suggested by the complexity of the analysis of
meaningl. If to meanl is to intend in some complex way,
moreover, if it is to have the three intentions that Grice
lists, how is it that we are not always aware of having
these intentions while we are speaking. This theory of
meaningl makes it difficult to account for the relative ease
with which we communicate. For, not only does communication
involve that A have the three intentions, but it involves

I's correct recognition of A's (i1). We might combine these

two observations to form:

Problem F This theory of meaning is questionable because it
multiplies the speaker's intentions in such a way
as to (l) violate the subjective evidence of
introspection, given that the intentions had when
speaking meaningfullyl do not seem to be as com-
plex or as numerous as this analysis demands;
and to (2) make more difficult the explanation
of human interpersonal communication, given that

 

85

communication demands the meaningful production
of a token as well as the interpretive reading
of that token.

One of Grice's first critics, William Alston, seems

to be suggesting that something like Problem F must be faced.

 

In his 1964 article "Linguistic Acts" Alston first suggests
an example he takes to be problematic for Grice.20
Example 2 (Alston): A man, intending to inform us that
his battery is dead, suffers a slip of the tongue, and
utters the sentence, 'My beagle is dead'.21
The man did not inform us that his battery is dead by utter-
ing those words. Yet, it was his intention to do so. How-
ever Alston's example is not troublesome. All that it does
is indicate the meaningl/meaning2 distinction and urge that
practical speakers will use expressions that mean2 what they
(the speakers) meanl.

Alston goes on to suggest Problem F or something like

 

it by citing Wittgenstein's famous experiment: "Say 'It's
cold here' and mean 'It's warm here'. Can you do it? And
22

what are you doing as you do it?" Again Alston approaches

a vague statement of Problem F when he comments that one's

 

intentions are not completely obvious. One may think that

A was trying to accomplish any one of a number of things by
uttering I, but be unable to guess which of these was Afs
primary perlocutionary or illocutionary intention. Thus he
would not know what Ameansnn by uttering I. Alston remarks,
"it is impossible for one to tell from the context which of

the acts performable by uttering a sentence was" the one that

86

the speaker intended to perform.23 By sayingl 'I'll be
there' someone may be predicting, promising, threatening,
or reassuring. Alston does not notice the obvious replies:
The situation for this theory is no worse than it is for
most others. At least this theory has the virtue of indi-
cating that speakers should be practical in their choice of
tokens. Also, it suggests that our occasional difficulties
in determining what people meanl can be accounted for by
noticing that a speaker's intentions are often not com-
pletely obvious.

Problem F will become more forceful as Grice finds

 

it necessary to expand the number of intentions in his
analysis from three to nine! We shall return, then, to

this problem in the next chapter.

GroupyThree

 

This group of problems will, like earlier groups,
be of only momentary concern, for it is comprised of those
problems which affect issue II rather than issue I. They
suggest difficulties that must be overcome in the effort to
show the relationship between meaning2 and meaningl. Several
of these difficulties were already suggested in Chapter I
when the programs of Leonard and Grice were outlined. Here

are some others.

Problem G This theory wrongly claims that the meaning of an
expression is identical to whatever a person
meant by uttering that expression.

 

87

In Chapter I we concluded that such an objection is mis-
guided because it attributes (Thesis 2) to this theory of
meaning. But, (Thesis 2), while false, is not entailed by
this theory. However, Grice's "rough equivalences" noted

above suggest that Problem G does apply to Grice's formula-

 

tion of the theory as it appears in "Meaning." Later Grice
will distinguish between the "applied timeless meaningnn"
of an utterance—type and the "occasion meaningnn" of an
utterance-type (that is, between meaning2 and meaning3) and

thus avoid this problem. Thus, while Problem G seems to

 

apply to the present formulation, it is easily avoided and
thus can be put aside.

The other two problems in this group are suggested
by the work of Paul Ziff. In "On H. P. Grice's Account of
Meaning," 1967, Ziff examines and attempts to discredit
Grice's theory because "the coin is counterfeit and seems

24 Like Grice, Ziff fails to note

to be gaining currency."
some distinction between meaningl and meaningz. But he
does seem, at times, to realize that such a distinction
might be provided, for he does look at the relationship
between meaning2 and meaningl. Ziff does not object that
the relationship between token—meaning2 and type-meaning2
is poorly executed, but that the meaning2 of sentence-types

(and hence their tokens) is not projectible given Grice's

analysis of meaningl. Ziff's major objection is

88

Problem H This theory is incomplete for it lacks pro-
visions for an essential element in the determi-
nation of sentengg-type meaning: a set of pro-
jective devices.

 

Ziff argues that natural languages are, in their syntactic
and semantic structures, essentially recursive. Thus, for
example, we can ordinarily determine what an utterance
means2 even if it is the first time that we have ever
encountered a token of this particular utterance-type. In

support of his major objection Ziff suggests

Problem I Since the meaning of an utterance-token is what
A meant n by utter1ng it, when we encounter a
token 0? an utterance-type never before encoun-
tered, we cannot figure out what it meansnn. But
this result is contrary to our linguistic experi-
ence. This result is due to the theory's lack

of projective devices for the determination of
meaningnn.

 

For example, consider the sentence
(7) He's a son of a stickleback fish.

The sentence means2 that the male referred to is a son of a
small scaleless fish (family Gasterosteidae) having two or

26 But since

more free spines in front of the dorsal fin.
(7) is rarely if ever uttered, it is impossible to determine
its meaning2 on the basis of what people generally intend by
uttering it, or even what they IUUIU intend by uttering it.
What they "would intend" is a matter for speculation, given
that no projective method is offered by Grice. They might,

for example, intend by (7) to denigrate someone. It seems

clear that this intention is independent of the meaning2 of

89

(7). Ziff's comment indicates that he notices that a
person's perlocutionary intentions may vary independently
of the illocutionary force of what he said. "What people
would intend to effect by uttering such a sentence would
most likely have nothing whatever to do with the meaning of

25

the sentence." That they intend this or that by their

utterance is "wholly irrelevant" to its meaningz.26
A's perlocutionary intentions are irrelevant in the
determination of the meaning2 of I. But if a token means
what someone meantnn by it, and if this is a specification
of his intended perlocutionary effect, then, according to
Grice these intentions do play a part in determining the
token's meaningnn. Ziff's problem, however, is that he is
working without the meaningl/meaning2 distinction. Problem
I is not a genuine problem for Grice's theory, since it
slurs the issues. Perhaps we cannot determine, in the case
of novel utterances, what the speaker meantl by them; but

we can determine what they mean The two matters are dis-

2.

tinct. Problem H, on the other hand, is more serious. In

 

the treatment of issue II Grice will have to provide for
the projective characteristics of natural languages. In
doing this he will provide for our determination of the
meaning2 of novel utterances. The task, as outlined in

Chapter I, is very complex. Problem H is serious, but per—

 

haps not insurmountable. That Grice addresses himself pri-

marily to the definition of meaningl and not to issue II in

90

"Meaning" is a matter of history. Ziff's criticism is only
an indication of the need to approach the second issue to
achieve a complete theory of meaning. It does not under-
mine the work done on issue I, unless that work precludes
the successful resolution of issue II.

There are several other interesting objections in
Ziff's article. These will be mentioned as we move on to
groups four and five. But for now let us turn to a 1969
article by T. E. Patton and D. W. Stampe, "The Rudiments of

Meaning: on Ziff on Grice,‘ to see their defense of Grice's
program against Ziff's major objection.27 Patton and Stampe
tone down Ziff's zealous conclusions by pointing out that
his arguments, serious as some of them are, do not rule out
amendments to Grice's analysis. They feel that no essential
flaw has been discovered. They may have known of Grice's
two articles of the late sixties, and this may account for
their not offering any amendments for Grice. They are con-
tent to argue that a speaker's intentions are not "wholly
irrelevant" to the meaning2 of his utterance.28 If, as
Patton and Stampe say, criticisms of criticisms are subject

to the "law of diminishing fleas,‘ then surely reviews of
them are more so. Thus, I will discuss only their reply to
Ziff's argument.

They point out that Grice's analysis is not restricted

to language. Grice is trying to explicate the concept of

meaningnn not only for sentences like

 

 

 

91

(8) The sentence 'Bovines are ungulate' means cows have
hoofs,

but also for ones like

(9) Those three rings of the bell mean that the bus is
full.

The rings of a bell are not parts of an essentially recur-
sive semantic system. Thus, one need not make appeal to
projective devices to explicate the concept of meaningnn in
such non—linguistic cases. They argue that the sense of
'mean' in (8) and (9) is the same. Hence, there "would
seem to be no reason to suppose" that the kind of meaningnn
that figures in the former has any more need of being
explicated in terms of projective devices than did the
latter. Ziff's mistake was identifying an essential feature
of language with an essential feature of meaningnn. "Our
interest in the meaning of 'mean' is to be distinguished
from our interest in the projective character of language."29
Grice is, they claim, addressing himself to the more primi-
tive philosophical concern, he is asking "What is it for a
mark or movement or sound to mean something?"28
Their last remark is mistaken, but it might be just
a slip. Grice's concern is not, at least as far as his
work in "Meaning" is concerned, with how a mark, movement,
or sound should meanz; but, with what it is for a speaker

to meanl something by uttering a sound or making a mark or

a movement.

92

Turning to the problem of novel utterances, they
admit that Grice's analysis does not help us with the prob-
lem of projection. Rightly, they argue that this shows
that answering the question about what 'mean' means is not
to settle the problem of how speakers are able to determine
the meaning2 of novel utterances. They recast the problem
in Gricean terms, "how are speakers of a language able to
know or determine what is, or would be, intended by novel

30

utterances?" The issue now becomes an aspect of Problem F.

 

However, the matter of how I arrives at his estimation of
A's intentions is, like the matter of Afs wisdom in selecting
the cues he will use to reveal these intentions, a matter

that is independent of providing an analysis of meaningl.

Group Four

 

The problems in group four have to do with certain
logical peculiarities of Grice's analysis of meaning, and
the commitments that the logical structure of Grice's
definition forces upon him. The problems in this group
represent more serious difficulties for this analysis of
meaningl.

Looking at (M) it becomes evident that it is essen-
tial that Grice quantify into the clauses of the conjunction.
But these clauses are referentially opaque. W. V. Quine has
argued persuasively that propositional attitude contexts are

referentially opaque and thus "it is pIima facie meaningless
31

 

to quantify into them." Thus we have Problem J.

 

93

Problem J This theory is objectionable because it involves
quantification into referentially opaque con-
texts.

 

But, even Quine is not prepared to rule out all such quanti—
fication. There are certain "indispensable relational
statements" that seem to involve such quantification.31 For
example, we often find it useful to use a sentence like
"There is something that I intend to do tonight." Thus the
problem is unresolved, and we can side with Grice, at least
for the moment. It seems that an analysis of meaningl must
make use of these opaque contexts and curious quantifica-

tions-—if we are to offer the analysis in terms of inten-

tions at all. Let us hope that in putting aside Problem J

 

we are not deferring a logical problem the eventual resolu-
tion of which will undermine any such analysis of meaningl.
The existential quantification Grice's analysis
requires commits Grice to the existence of effects. We
argued above that I need not respond by producing U, yet A
would still have spoken meaningfully. This reply presup-
posed that there was a way out of the objection that if I
fails to produce 3, then no I exists to satisfy the existen-
tial quantification. The way out was to claim that the
value of 'e' need not be any particular event or situation
that I might bring about. Rather, the value of 'e' was
thought of as a universal, some event-type or effect-type.
This universal may or may not become instantiated in or by

B. As long as the value of 'e' is such a universal we have

94

satisfied the existential quantification. The price is, of
course, a complication of Grice's ontology with the addition
of effect—types, universals, like, for example, believing

that p. We can state Problem K as follows:

 

Problem K Grice's analysis of meaning n entails the exis—
tence of some effect, U, thg nature of which is
unsuitably vague.

 

A partial reply to this problem is to argue that the
analysis does entail the existence of some 3, yet the value
of U is not any particular effect, but a universal effect-
type. We have other clues to restrictions that Grice would
wish to make on the kinds of effect-types A might intend to
instantiate in I. They must be, in A's estimation, within
I's control insofar as I can act to produce an instance of
E in himself by reason of his recognition of Afs intention
that such an instance be produced. Further, we know that

U is, in every case, some perlocutionary effect-type. Thus,

while the nature of U is not completely clear, Problem K is

 

settled sufficiently to allow us not to regard it as one of
the most pressing problems that this analysis must face.
We shall return again to the problem of the nature of U.
Another existential commitment explicit in (M) is
made to audiences. It seems that some I must exist. We
know very little about the nature of such an audience. It
must, we know, be able to recognize certain things, produce
certain effects, and act on the basis of reasons. Rather,

A must believe that it can do these things. But what is

95

this audience like, independent of Afs conception of it?

If I publish an article or address a radio audience, who am
I addressing? Must I, the audience, be a specific person?
What happens if nobody reads the journal article or nobody
listens to the radio program. We should not want to say
that I have failed to meanl something by what I saidl in
such cases. The nature of the audience and its relationship
to A's beliefs concerning it are further obscured by the
opacity of Grice's explicans. Suppose A, a thief, wishes
to get I to believe that there are policemen about so he
saysl "Watch out for the cops." Suppose that A has the
other intentions required but that he does not know that I
is a policeman. Probably, if asked, A would deny that he
intended to get a policeman to believe that there were
police about. Does this entail that, relative to I—as-
conceived-of—by—A, Ameantnn something by uttering I, but,
relative to I (under any other true description), A_meantnn
nothing? The serious problem of the existence and nature
of the audience is noticed by several commentators.32 It
will be settled only toward the end of Chapter III by a
formulation that does not entail the existence of an audi—
ence. In such a formulation the nature of the audience is
determined by A's conception or estimation of the audience

he intends to address. We can formulate the problem of the

audience as follows:

96

 

Problem L Grice's analysis of meaning n entails the exis-
tence of some audience B, t e nature of this
audience is left unsuitably vague. But there
are times at which we should want to say that A
spoke meaningfully even if no audience exists,
or none exists with the characteristics that A
anticipated it would have. '—

As Ziff noticed with regards to this problem, we
cannot think of the speaker as his own audience when no
other audience is available. To do this would be to claim
that he intended to produce an effect in himself by means
of his own recognition of his own intention. An odd situa-

tion indeed.32

Group Five

 

The problems in this group are generated by Grice's
disregard of the i1locutionary/perlocutionary distinction.
It is through this avenue that most of the objections to
Grice's work come, for most, if not all, of the complex
counter-examples Grice deals with turn on the neglect of
this distinction and his failure to provide some substitute
for it. Grice's thesis, according to N. L. Wilson, is that
one can reduce illocutionary intentions to perlocutionary
intentions.5 I believe that this thesis is mistaken; the
reduction will encounter problems that will render it
impossible. Moreover, the failure of this reduction will
render Grice's analysis an inadequate analysis of meaningl.
Here are some of the difficulties that Grice's analysis of
meaningl must face, given its disregard for the illocu-

tionary/perlocutionary distinction. I regard these

97

objections as quite serious, they render Grice's analysis
unacceptable both in its original presentation, and, as I
shall argue at length in Chapter III, in its more recent

formulations.

Problem M Illocutionary intentions cannot be successfully
reduced to perlocutionary intentions because the
illocutionary force of an utterance of I, as
determined by the meaning of I, need not be
identical to the perlocutionary effect (if any)
that A seeks to achieve by uttering x. Afs
perlocutionary intentions can vary ifidependently
of the conventionally determined illocutionary
force of A's utterance.

 

For example, A might deliver the warning "Leave my books
alone" intending to warn I to leave his books alone. Grice's
reduction demands that A have some perlocutionary intention
(perhaps very complex) from which it follows that he has

the intention to warn I. This is not obviously so; but,
assuming it is, it is not obvious what perlocutionary inten—
tion A_must have. It does not seem to follow from the fact
that A has a certain perlocutionary intention that in utter—
ing I_A warned I. One might suggest that it does follow
from an intention like "A intends to get I'to leave A's
books alone" or "A intends to get I to believe that A
intends that I leave A's books alone." But it does not, for
§,EISEE have intentions such as these without having the

intention to warn I.

A problem which is like Problem M, but which is a

 

more serious threat to the reduction is the following:

98

Problem N This theory of meaning is mistaken in that it
requires that whenever one speaks meaningfullyl
one has some perlocutionary intention or other.
But, at least in some cases, to meanl something
by what one says is not to intend to induce any
perlocutionary e fect in one's audience.

 

For example, suppose that I know that I is an exceptionally
incredulous man. Suppose that I promise him that I will
finish my work on time saying "I promise to have the job
done on time." Suppose that I say this for no other motive
than to satisfy my scruples, for I know that I will not
believe me. It would seem that I have meantl something by
what I saidl, but that I have no intention to induce or
produce any perlocutionary effect in I.by what I have said.

Perhaps Problem N is understated, it seems that in normal

 

cases I need not have Ipy perlocutionary intentions to have
meantl something by what I saidl. But, because of the com-
plexity of our acts, because people have purposes upon pur—
poses, it seems difficult to speak of the "normal" case.
Throughout the present work I have assumed that
meaningl is better associated with illocutionary than with
perlocutionary intentions. If Grice's reduction fails, and
if his proposed analysis is better suited to perlocutionary
rather than illocutionary intentions--as I have urged
several times——then Grice's analysis of meaning will be mis-
guided. It will, at least, be incomplete because of its
inability to handle illocutionary intentions. But, if

meaningl is better associated with illocutionary intentions,

99

then his analysis will fail to be an analysis of meaningl.
The arguments in favor of these conclusions are presented

in Chapter III. For now, let us merely state the problem.

Problem 0 This theory of meaningl is inadequate. Since
meaning is best associated with illocutionary
but not perlocutionary intentions, the analysis
offered is an analysis of an incorrect expli-
candum. The analysis offered is an analysis of
'trying to accomplish something by uttering
something' rather than 'uttering something and

 

meaning it'. However, an adequate theory of
meaningl must examine something like the latter
analysandum.

To reinforce the objection that meaning1 is better
associated with illocutionary intentions, and thus to support

the conventions of Problem 0, we can recall that even Grice

 

tried to distinguish between "getting something done by
uttering something" and "telling someone something" (or,

more generally, "uttering something and meaning it").

Several critics will argue that by the addition of clause
(13) Grice did not succeed in capturing the distinction
between "telling I that p" and "getting I to believe that p."
The problem is that (even with all three of the Gricean
intentions) one can accomplish the latter as a perlocutionary
effect of uttering something, but one can still have failed
to do the former in uttering that something. Generalizing
the problem apparent here, we have a problem which becomes
the strategy for several counter-examples to Grice's

analysis.

100

Problem P This theory of meaningl is counter-intuitive
since a number of effects, I, can be achieved
as intended perlocutionary effects of utter-
ances given that A is able to correctly antici-
pate I's, perhaps—mistaken, estimation of A's
intentions. In such cases A will have manipu-
lated I to achieve e, but WE should not want to
say that by uttering-'IAmeantl U.

 

In the counter—examples offered on the pattern of Problem P

 

we will, in general, find that Afs manipulation of I is
unfair, often it is based on deception or on I's ignorance
of some important fact. But the manipulation need not be
unfair. Ifs estimation of Afs intentions can be correct
and can be intended to be correct. Furthermore, §.E§X

meanl I in situations that are based on Problem N. What is

 

crucial is that A can have all the Gricean intentions and
yet we should still find it counter—intuitive to say that

because he had these intentions he meant something

1

(generally that "something" is U) by uttering I, or even

 

that because he had the intentions he meant1 I.

At this point we should take note of certain dis-
tinctions. Let 'a' denote any act, including, for example,
uttering I. We can distinguish "uttering something and
meaningl it" from Grice's ambiguous "meaning something by
uttering something." It is not clear, until Grice tells us,
that the something meant is not I but a specification of
the intended effect, viz., U. We can also distinguish
"uttering something and meaningnn something by it" from

"meaning (to do) something by (doing) I." The latter is

101

synonymous with "intending to do something by doing I." It
does not employ 'meaning' in the sense of either meaningl
or meaning .

nn

Let us employ the strategy suggested by Problem P.

 

Counter~example l (Facione): Suppose that I have an
upset stomach and so I go to the local pharmacist to

ask his advice. I realize, of course, that he cannot
sell me any prescription medicine. I know it is illegal
for him to prescribe, and risky for him to even recom-
mend any kind of medication. To my request for advice
he responds with the statement "Whenever I feel that way
I take some 'Dr. Witch's Panacea'."

The pharmacist, who has the three Gricean meaningnn inten-
tions, must have wanted me to believe that the Panacea
would help me. But this is not what the utterance, as a

token of a certain type, meant Whether we should want to

2.
say, independently of Grice's analysis, that by uttering I
the pharmacist meantl something, is, as far as my intuitions
go, an open question. In part I would venture a negative
answer, claiming that while he did want (intend) me to
believe something, he did not meanl by uttering I what he

intended me to believe. Besides exemplifying Problem P this

 

counter-example attacks the too-easy identification of the
meaningnn of tokens with their author's meaningl.
Counter-examples might be suggested to show that
Grice's definition is, in some respects, too weak, and, in
others, too strong. Some things that we should want to say
are BEE meaningful turn out to be meaningfulnn; other mean—
ingful things turn out to be meaninglessnn. The next

counter-example indicates that the analysis is inadequate in

102

still other ways. For one, the proposed analysandum of

the Gricean analysis seems inadequate to express the way
that meaningl something might be relative to a particular
audience. Also, by fulfilling Grice's analysans the speaker
below has meantnn something (3). I would argue that he did
meanl what he saidl (and so he meantl something), but what
he meantl was identical to the meaning2 of what he saidl.

I would not argue that he meantl I, but that he intended to
bring about U as-a perlocutionary effect.

Counter-example 2 (Facione): I, a spy, come to realize
that enemy agents are watching me. I am to contact my
companion and assure him that our plan can proceed. I
have reason to believe that the enemy knows this, but
they do not know what the plan is nor exactly what our
objective is. I want to abort our mission without let-
ting the enemy know; but I have no prearranged code
worked out with my contact. I make radio contact as all
parties suspect, ordering "Tell headquarters we are
going ahead as planned, all necessary arrangements have
been made." I intend that my contact realize that things
have gone wrong and thus will abort the mission. I
intend that he will realize that it is my intention to
have him think so, and that he will think so because of
that realization. My intention is that my contact employ
the following inference: "I was ordered to contact head-
quarters and to tell them we are going ahead as planned.
But it was never part of our plan to contact headquarters
once the operation had come this far along. Either my
partner has forgotten this, or he has deliberately put
this contradiction into his message. He has never before
been mixed up about our plans. He must, then, have
wanted me to interpret this contradiction somehow. If it
were possible he would have just told me what the situa-
tion was. It must be impossible to do that, which could
be the case only if he is being watched by the enemy and
knows it. But if that is so, then we had better not
move ahead with our plans because if we do we shall all
be captured. This must be what he wants me to understand
by using the phrase 'going ahead as planned'."

According to Grice my utterance is meaningfulnn. It

does have a meaning2 by virtue of which I hope to achieve my

103

purposes. I hope the enemy will take what I said literally
and continue to believe that the plan is moving ahead, 21°
Because they do not perceive my intention that they think
so, I have a chance that my purpose will be achieved. Thus,
relative to the enemy my utterance has no meaningnn, for I

do not intend that they know my intention to achieve e'.

 

Yet it does have meaningnn to my contact, according
to Grice, since I do have the requisite intentions toward
III. However, the token means2 exactly the opposite of
what I meantl by it. This possibility, ruled out by Grice's
definition of token—meaningnn, is what I am counting on to
achieve my purpose of aborting the mission. I hope that my
contact will see that I have ordered him to contact head-

quarters, by knowing what my utterance means I also hope

2.
he recognizes the inconsistency of this request with "moving
ahead as planned" since we did not plan to contact head-
quarters. Seeing this inconsistency he should realize that
something has gone wrong and that we should halt the mission.
He should, at least, realize that I think that something has
gone wrong and that we should abort the mission--ca11 this

realization U. But I have not told him this; although I did

get him to know it. Thus Problem P appears and suggests

 

 

Problem 0.

 

My contact learns my perlocutionary intention by
knowing the conventional meaning2 of I as a token of type I,

and by assuming that I meantl what I saidl. Soon he suspects

104

that I have a perlocutionary intention, and thinks that his
key to learning it is the meaning2 of I and the ostensible

illocutionary force of my uttering I. Thus Problem M

 

appears.
Others have published counter-examples as well.
To support the moot point that Grice's theory lacks pro-
jective devices Ziff offers counter-examples attacking
Grice's definition of meaningnn. Ziff hopes to show that
Grice's analysis "never gets off the ground." But all that
he does show is that, given Grice's theory, meaningless

2

l’ and meaningful2 sounds can fail

Ziff's counter-examples do have virtue

sounds can be meaningful

to have meaningl.33

for they indicate that Grice's analysis of meaningnn is

 

inadequate as an analysis of meaningl. As in Problem 0,
Ziff argues that Grice is explicating an inappropriate
expression.
Counter-example 3 (Ziff): On being inducted into the
army, George is compelled to take a test designed to
establish sanity. George is known to be an irritable
academic. The test he is being given would be appro-
priate for morons. One of the questions asked is:
'What would you say if you were asked to identify your-
self?‘ George replied to the officer asking the question
by uttering (1) 'Ugh ugh blugh blugh ugh blugh blug!‘34
Ziff argues that George meantnn something by (10). He
intended that the officer be offended by (10) because of his
recognition of George's intention to offend him. Thus George
meantnn something by (10), but (10) did not mean2 anything

at all. "On Grice's account good intentions suffice to con-

vert nonsense to sense; the road to Babble is paved with

105

26

such intentions." Ziff rightly argues that George did

not meanl (10) nor mean something by (10). Our pre-

1
analytic intuitions urge us to say that nothing has been
meantl even though the speaker had the three intentions
Grice's analysis requires. We would not expect that George

could paraphrase what he meant by (10). This can be a

l
clue to our pre-analytic intuitions. Ziff claims that
Grice has confused and conflated expressions like "Ameantl
something by uttering II with expressions like "A meant (to
do) something by uttering I." In the above case the latter
is true, but the former is false. Ziff's counter-example
shows that the perlocutionary intentions sufficient for the

truth of "A meant (to do) something by uttering I" are not

sufficient for the truth of either "A meant
35

" "
1 §. or A

meantl something by uttering I." Thus Grice's definition
is inadequate, it must either be reformed or discarded.
Another attack through counter-example is mounted
by P. F. Strawson. Although, Strawson's tone is more
sympathetic and constructive than Ziff's. In his 1964
article "Intention and Convention in Speech Acts" Strawson

attempts to analyze a person's understanding an utterance.36

His discussion concerns How to Do Things With Words, as he

 

attempts to elucidate Austin's concept of "securing uptake"
in order to examine the concepts of an "illocutionary act"

and the "illocutionary force of an utterance."

106

Thinking Grice's notion of meaningnn helpful,

0 a o I 37
Strawson summar1zes and refines the Gr1cean analys1s.

(S) If A meant,1n something by uttering x then

there is some audience, B, and some response, r,

such that: _ —

(i1) A intends to produce r in B by uttering I

(12) intends that B recognize _that (i)

(13) intends that I' s recognition of iil )
function as I' 8 _reason, or part of I1
reason, for response r.

[WWW

I have taken the liberty to make Strawson's redefi-
nition conform to the logical structure I believe he intends
it to have. He claims that his changes are only to make
the nature of Grice's I more obvious, and to remove the
causal connotations of his formulation. However Strawson
has introduced a conceptual change, which he apparently
does not think to be important enough to mention, in the
formulation of (11). Notice that in (M) there is no
apparent connection between A's having uttered I (rather
than having uttered something else, or having done something
else altogether) and the intention to produce U in B.

Strawson has made a connection. Afs uttering x is to be

the means by which A intends to achieve the production of I
in I. Strawson notes that I can be a cognitive response or
attitude as well as an act. He also notes that in every
meaningfulnn speech act A seeks the cognitive response of
I's recognition of A's (i1), beside seeking to induce E

38

in B.

107

Strawson regards Grice's analysis as an effort to
explicate Ifs attempt to communicate, in a fundamental and
straightforward way, with I. This is, I believe, a fair
interpretation. Strawson argues, by counter-example, that
the analysis, (S), is too weak.

Counter- example 4 (Strawson, paraphrased): Suppose
that A intends to induce a belief that p in I by a
certain action, thereby satisfying i _A arranges
convincing-looking evidence that p.where— B will see

it. Suppose that B is watching A work, and that I
knows that B is watching him, and further that A

knows or believes that B does not realize that A is
aware that B is watching. Since I is arranging—
evidence, rather than acting out or pretending to be
doing something, I expects that I will not take the
evidence he saw I arrange as genuine evidence that p.
Rather B will infer from seeing I.at work that I intends
B to believe that p. But this is precisely what I
foresees and intends, thus satisfying (i) Suppose
further that A knows that B generally trasts him, that
is that I believes that I would not want I to think that
2 unless _A thought p were the case. Thus if I has
recognized that A wants him to believe that 3, he will,
at least partly for that reason, believe that E! A
intends that B's recognition of (i ) function just that
way, and so I _satisfies (I 3).

It seems to be a mistake to say that I was trying to com-
municate with I, even though I_has all three intentions.39
I intends that I believe that p, but I tries to get I to
believe that p_without telling him that p.

Grice did not notice, says Strawson, that if I_is
trying to communicate he must have still another intention.
He must intend that I become aware that he, I, is "trying

to let him, I, know" something. We must add another clause

to rule out counter-example 4.

(i4) I intends that I recognize that (i2)40

108

Strawson has made use of Problem P to suggest

 

another clause for the Gricean analysans. He has not, how-
ever, precluded further counter-examples like his own.41
Perhaps some reference to linguistic conventions,

which are independent of Ifs intentions, would insure that

one could not appeal to Problem P to derive counter—

 

examples. As we examine, with Strawson, some of the con-
ventional aspects of illocutionary acts, notice that he
urges that Grice's analysis would not be an adequate
analysis of illocutionary acts. This supports the conten-

tions of Problem 0.

 

Illocutionary acts range in conventionality from
those whose only conventions seem to be ones associated
with their locutionary bases, to those whose conventions
determine their effect (force) independently of their
agent's intentions. For example, in playing auction bridge

a player who utters "Pass" in his turn has passed. On the

 

other hand, if someone says "I'll be there" he could be
warning, promising, predicting, or threatening. Strawson
admits that even his own expanded analysis is inadequate if
applied to illocutionary acts.42 He wishes to claim only
that the four clauses he lists are necessary conditions, not
that they are jointly sufficient. But, whether his analysis
is adequate is somewhat irrelevant, since Grice is not
trying to explicate 'In uttering I_I did . . .'. Yet,
insofar as meaningl is associated with illocutionary inten-

tions, maybe he should have tried to explicate it.

109

Strawson offers two reasons why his revised
analysis is not an adequate analysis of illocutionary EEE§°
Like Alston, Strawson suggests that one cannot easily
determine which of perhaps many illocutionary forces a
speaker might have intended in performing an illocutionary
act. In sayingl 'Don't go' I might have intended to issue
a warning or a command. Yet, perhaps Grice could counter
this objection by altering the content of I. Suppose we
did not think of I, in such a case, as the response that I
not go. Let us, rather, think of it as the response that I
believes that I wishes to ... I not to go (where the lacuna
is filled in by the appropriate illocutionary force, viz.,
the one that I intends). Grice will, as we shall see in
the next chapter, make such alterations in the content of I.
I am anticipating them here only to suggest a way out of
Strawson's objection.

Strawson's second objection is that as the conven-
tions surrounding an illocutionary act become more and more
institutionalized, we come less and less to need all four
of the clauses in the analysis. In some cases we may not
need clause (i1) because no response I is intended. This
suggests that Strawson may be aware of something like our

Problem N. Grice's analysis seems to suit perlocutionary

 

intentions since these require intending to effect some
response in I, but I's illocutionary intentions are a dif-
ferent matter. The only response one need intend is I's

understanding.

110

According to Strawson the success of an illocu-
tionary act requires nothing beyond I's understanding what
is going on, via, perhaps, the conventions associated with
the illocutionary act by virtue of the meaning2 of the
expression uttered in its locutionary base. Thus, perhaps,
the (i3) intention is superfluous. Strawson concludes that
Grice's analysis suits best those illocutions with the
fewest associated conventions; that is, those on the other
end of the scale from the bridge example. But, those like
the bid in the bridge example may require a theory that
claims that the illocutionary force of an utterance is

exhausted by its meaning2.43

However, in reply to Strawson,
Grice may argue that if his analysis suits any illocutionary
acts, it can be made to suit them all. But this remains to
be seen. I am sympathetic to the View that we have already
incorporated more intentions into the analysans than would
be necessary to show the connection between meaningl what
one says1 and one's illocutionary intentions.

These last four problems are very serious attacks
on the adequacy of Grice's analysis of meaningl. We shall,
in the next chapter, pursue his efforts at an analysis of
meaningl to see whether he is able to overcome these prob—
lems, or any of the as yet outstanding problems mentioned
in the present chapter. It is possible that Grice's

analysis is, due to its neglect of the illocutionary/perlo-

cutionary distinction, destined for inadequacy. But suppose

 

 

 

111

that some philosopher proposes an analysis of meaning1 that
lists only intentions and which does not neglect this dis-
tinction. This, it seems, will go far toward fulfilling

the ambitions of this theory of meaning. Let us then search
for an objection that would force us to move beyond the

speaker's intentions to discover all the elements in mean-

ingl.

Group Six

 

This group will contain only one problem in it, but
it will be a very serious difficulty for it turns on the
fact that when one is seeking to communicate he speaks using
signs that he believes that his audience can understand.
This belief concerning how the audience will come to inter-
pret the signs used dictates the practical need to choose
wisely. It also dictates to the theorist the necessity to
move beyond listing merely the speaker's intentions and to
include also his beliefs. Let us look at some of the con-
siderations that suggest this problem as they begin to
emerge in the work of philosophers like John Searle.

Perhaps the most insightful work to issue from
Grice's original theory, due to its effort to link inten-
tions and conventions in a unified theory of meaning, is
John Searle's 1965 article "What is a Speech Act?" and his

1969 Speech Acts: An Essay in the Philosophy of Language.44

 

According to Searle, Grice's analysis is a good beginning,

112

for it marks the connection between meaning and intentions.

It also "captures" an essential feature of communication.
In speaking I attempt to communicate certain things to
my hearer by getting him to recognize my intention to
achieve that effect, and as soon as the hearer recog—
nizes what it is gy intention to achieve, it is in
general ach1eved.

The hearer understands what the speaker is saying as soon

as he recognizes the speaker's (illocutionary) intention

in uttering what he uttered as an intention to sayi just

that.

Searle criticizes Grice's account of meaning for
having confused illocutionary and perlocutionary effects
of utterances. Also Searle restates the criticism that
Strawson's article suggested. Grice has failed, claims
Searle, to recognize the extent to which meaning is a matter
of conventions and rules. Because of the confusion of kinds
of effects just mentioned, Grice has been unable to ade—
quately express the relationship between a speaker's mean-
ingl something and the meaning2 that an utterance has by
virtue of the conventions of a given language. Searle's

criticisms point to several familiar problems. Primarily

he is suggesting Problem M, and his remarks also suggest

 

that Grice's account is incomplete since issue II is not
resolved.
Searle's comments also suggest a new problem. I

believe that we have in Problem Q an objection which will

 

force us to include in our analysis more than A's intentions

113

when he uttered I. We shall have to include certain beliefs

he entertained as well.

 

Problem Q This analysis of meaningl neglects the fact
that in speaking meaningfullyl I_must select a
token which he believes to have some feature
(generally its meaning ) by virtue of which I
will be able to undersgand what A meansl, unless
I is not trying to communicate with I.

Problem Q is not a restatement of the practical con—

 

siderations frequently discussed above. It is a necessity
that I select what he believes to be an understandable
token. According to Grice, Strawson, and Austin, I must
intend that I understand what I_is trying to do in uttering
I. If I does not select a token he believes I can under—
stand, then how can he intend that I shall understand? If
it were only a question of perlocutionary effects, it would
be possible to get I to think whatever I wishes him to
think. But meaningl what we sayl is concerned with illocu-
tionary intentions. I.is telling I what he wishes I to
know, not simply getting I to know it. Grice's analysis is
inadequate for although it provides for the intention that
I understand, it does not provide for Ifs selection of a
token which I believes will make this understanding possible.
In an explication of 'sayingl something and meaningl it'
this condition must be included. Perhaps, however, it need
not be included in Grice's current explicans for he is

analyzing 'By uttering I_I meant something'. Given this

1

explicandum it does not seem necessary to require that I

believe I to be a signal
indicates that Grice has
candum. We shall return

and Problem Q in Chapter

 

troublesome not only for

114

of what he meansl. Yet, this
selected an inappropriate expli-
to the matter of the explicandum

III. These matters could be very

Grice's analysis, but for the

theory of meaning as intention as well.

In an effort to substantiate his criticisms Searle

offers this counter-example. The counter-example, which is

based on Problem P, also

 

erations like those that

indicates some support for consid-

suggest Problem Q.

 

Counter-example 5 (Searle): Suppose that I am an
American soldier in the Second World War and that I am
captured by Italian troops. And suppose also that I
wish to get these troops to believe that I am a German
soldier in order to get them to release me. What I
would like to tell them in German or Italian is that I
am a German soldier. But let us suppose I don't know
enough German or Italian to do that. So I, as it were,
attempt to put on a show of telling them that I am a
German soldier by reciting those few bits of German I
know, trusting that they don't know enough German to
see through my plan. Let us suppose I know only one
line of German which I remember from a poem I had to
memorize in a high school German course. Therefore, I,
a captured American, address my Italian captors with
the following sentence: 'Kennst du das Land we die
Zitronem bluhen?‘ . . . A few imaginative additions to
the example should make the case more plausible, e.g.,
I know that my captors know there are German soldiers
in the area wearing American uniforms. I know that they
have been instructed to be on the lookout for these
Germans and to release them as soon as they identify
themselves. I know they have lied to their commander
by telling him that they can speak German when in fact
they cannot, etc.46

I intend to produce an effect in my captors, namely

that they come to believe I am a German soldier. I intend

that this effect be achieved by means of their recognition

115

of my intention. I intend that they think that I am trying
to tell them that I am a German soldier. According to
Grice, via Searle, my German sentence meansnn "I am a
German soldier." This is not precisely correct for Grice
stipulated that the specification of the intended effect is
the specification of what Imeansnn by uttering I. Since,
according to Grice, I_meansnn whatever Imeantnn by it,
here I_meansnn "Your believing that I am a German soldier,"
or "that you believe that I am a German soldier."

Whether one takes "I am a German soldier" or "Your
believing that I am a German soldier" as a specification of
what the utterance meansnn, the conclusion is unwarranted.

The utterance meant neither of these, it meant "Knowest

2 2

thou the land where the lemon trees bloom?" And neither of

these is what I meantl by the utterance, says Searle. How-

ever, given Problem Q, it is a more correct interpretation

 

of counter-example 5 to say that the American did not mean

 

l
anything by uttering I. He tried to accomplish something,

but he did not meanl what he saidl,

in this situation. In Chapter III we shall see that Grice

nor was anything meantl

considers both interpretations.
Searle seems to realize that some difficulty like

Problem Q is associated with Grice's analysis. Trying to

 

state it he says "What we can meanl is, at least sometimes,

"47

a function of what we are sayingl. If, on the other

hand, any sentence can be spoken with any meaningl

116

whatsoever, given only that the situation makes the appro-
priate intentions possible, then the meaning2 of the sen-
tence becomes "just another circumstance."

Searle attempts to amend Grice's account by making
sure that meaningl is more than just randomly related to
meaningz. If a speaker is using his words literally, he
intends that the hearer's recognition of the speaker's
original intention be achieved in virtue of some conven-

tional or rule governed association that correlates the

 

expressions people employ with the effects sought.

The "extraordinary" fact about human communication
is that a necessary and sufficient condition for success is
that I recognize that I is trying to tell him something,

48

and what I_is trying to tell him. Searle, like Strawson

and Austin, regards intendingI's understanding as a neces-

 

sary condition for meaningl something. However, argues
Searle, it is not a response or a belief that I_is trying
to have I undergo or accept. It is "simply I's under—

standing the utterance" that must be intended. This "effect"

is "the illocutionary effect."48

In the case of illocutionary acts we succeed in doing
what we are trying to do by getting our audience to
recognize what we are trying to do."’ (my italics.)

 

 

Searle slips from talking of "the illocutionary
effect," which seems to be something like the cognitive
response that Strawson tried to capture in clause (i2), to

talking of "3E illocutionary effect." If these latter

117

"effects" are, as it seems, to be drawn from the familiar
examples like warning, advising, requesting, etc., then
"III illocutionary effect" must be something over and above
these. It must, or may be, something which is common and
peculiar to illocutionary acts. In these matters Searle
is not clear. Perhaps "III illocutionary effect" is Austin's
"uptake."

Searle suggests that Grice's reflexive intention be
formulated: "the speaker I intends to produce 32 illocu-
tionary effect II in the hearer I_by means of getting I to

"48 In this formula-

recognize I's intention to produce II.
tion the distinction between one of the more common illocu—
tionary effects and "the illocutionary effect" of "getting

I to recognize I's intention to produce II" is apparent.

Searle offers this analysis of "saying something and mean-

ing it":
I utters sentence I and means it (i.e., means
literally what he says) =df
I utters I, and
(I1) S intends the utterance U of I.to produce in the

hearer I the knowledge, Eecognition, or awareness,
that the states of affairs specified by certain of
the rules of I obtain.

Searle refers to this effect (the knowledge or recognition or
awareness that certain states of affairs obtain) as "the
illocutionary effect, II."

(I2) S intends I to produce II by means of the recog—

Hition of I1.

118

(I3) I intends that (I1) will be recognized in virtue
or by means of I's knowledge of ceigain of the
rules governing the elements of I.

This definition of meaningl introduces a certain
element of conventionality that is absent from Grice's
analysis. But, insofar as Searle has not included a clause
that indicates what I actually believes about I and the

situation of his utterance of I, he has not taken into con-

sideration the objection noted in Problem Q. There is

 

virtue in this analysis. Not only is there an effort to
link intentions with conventions, but the explicandum seems
much more suited to an analysis of meaningl. Searle has
abandoned 'By uttering I I meant something' for 'I utters

I and meansl it'. However, this explicandum is unfortunate
for it ties meaningl too closely to the meaning2 of what is
spoken. O. H. Green noticed moreover that Searle's analysis,
like Grice's, requires the presence of an audience for mean-
ingl to take place.50 I hope to capture some of the virtues
and avoid the difficulties of Searle's analysis in my
analysis of meaningl in Chapter III.

A discussion which is like Searle's in its reliance
on both intentions and conventions, is D. S. Clarke's treat-
ment of communicated signs. Clarke urges that utterance-
tokens express a certain propositional content, their
literal meanings. The utterance of these tokens contains
illocutionary force-determiners like the intonation or the

context of the utterance. No conventional sign can have an

119

unintended illocutionary force, according to Clarke. The
force also depends on Ifs providing I with force-
determiners that aid I in the recognition of Ifs intended
illocutionary force. This second requirement prevents one
from reducing illocutionary force to Ifs intentions in a
Gricean way, since the force—determiners are by and large

51 Clarke's

either conventional or context dependent.
apparent identification of meaning2 and the propositional
content of an utterance—token is perhaps unfortunate, for
some tokens like 'Curses!‘ thus are without meaningz. The
work of Searle and Clarke indicate that although one cannot

neglect intentions in a theory of meaningl, intentions

appear not to be the whole story.

Summary

This chapter has presented Grice's first version of
his definition of meaningnn, as that analysis can be gleaned
from his 1957 article, "Meaning." Roughly, a speaker, I,
meansnn something by uttering I if and only if he intends
to produce some effect, I, in some hearer, I, by means of
the hearer's recognition of that intention.

There followed a critical commentary on this theory
as well as on the entire debate that this article fostered.
As each philosopher addressed Grice's theory his remarks
were noted and, in most cases, evaluated. Only N. L. Wilson's
objections were put off until the next chapter.5t These

encompass Grice's later articles as well as "Meaning."

120

Several objections were expounded. They may be
classified as follows: Group one is comprised of those
problems which are not serious threats to this theory of
meaning. Some misunderstand the theory, others simply do
not apply, others are not well supported themselves. Group
one includes:

Problem A Because this theory allows that an expression
may mean whatever its author wishes it to mean,
this theory threatens to destroy the "objectiv-
ity" of the meanings of our expressions, which

objectivity is the foundation for our communica-
tion.

 

Problem B This theory is unacceptable because it is a kind
of causal theory of meaning, and thus is sub-
ject to Grice's own critiques of such theories.

 

Problem C This theory is a modern version of the idea-
tional theory of meaning, thus it is unacceptable
for it renders communication impossible.

 

Problem D This theory of meaning presupposes that to mean
is to perform some kind of mental act. But,
this presupposition is false since to mean is
not to perform any kind of an act.

 

Some problems were more sweeping in scope or implication,
these were put aside in the hope that their eventual resolu-
tion would not adversely affect the analysis of meaningl.

These problems make up group two and include:

Problem E This theory is objectionable because it presup-
poses the existence of metaphysical entities,
viz. propositions and universals.

 

Problem F This theory of meaning is questionable because
it multiplies the speaker's intentions in such
a way as to (l) violate the subjective evidence
of introspection, given that the intentions had
when speaking meaningfullyl do not seem to be as
complex or as numerous as this analysis demands;

 

 

 

121

and to (2) make more difficult the explanation
of human interpersonal communication, given
that communication demands the meaningful pro-
duction of a token as well as the interpretive
reading of that token.

Another group of problems that can be disregarded, at least
for the moment, are those which affect issue II_but not the
analysis of meaningl. The third problem in this group is,
by far, more serious than the first two. Group three

includes:

Problem G This theory wrongly claims that the meaning of
an expression is identical to whatever a person
meant by uttering that expression.

 

Problem H This theory is incomplete for it lacks provisions
for an essential element in the determination of
sentence-type meaning: a set of projective
devices.

 

Problem I Since the meaning of an utterance-token is
what Imeantnn by uttering it, when we encounter
a token of an utterance-type never before
encountered, we cannot figure out what it means
But this result is contrary to our linguistic
experience. This result is due to the theory's
lack of projective devices for the determination
of meaningnn.

 

nn’

Beyond the problems just mentioned there are some
that are more serious threats to the analysis of meaningl.

Group four contains problems which may be overcome.

Problem J This theory is objectionable because it involves
quantification into referentially opaque con-
texts.

 

 

Problem K Grice's analysis of meaningnn entails the exis—
tence of some effect, I, the nature of which is
unsuitably vague.

Problem L Grice's analysis of meaning n entails the exis—
tence of some audience I, the nature of this

 

 

122

audience is left unsuitably vague. But there
are times at which we should want to say that

I spoke meaningfully even if no audience exists,
or none exists with the characteristics I_antic-
ipates it would have.

Group five is a very serious group of difficulties for

Grice's analysis of meaningl. The problems all are gener-

ated or turn on the illocutionary/perlocutionary distinction

and the result of Grice's failure to provide for it. They

are:

Problem M

 

Problem N

 

Problem 0

 

Problem P

 

Illocutionary intentions cannot be successfully
reduced to perlocutionary intentions because

the illocutionary force of an utterance of I,

as determined by the meaning of I, need not be
identical to the perlocutionary effect (if any)
that I seeks to achieve by uttering I. Ifs
perlocutionary intentions can vary independently
of the conventionally determined illocutionary
force of Ifs utterance.

This theory of meaning is mistaken in that it
requires that whenever one speaks meaningfully
one has some perlocutionary intention or other.
But, at least in some cases, to mean something
by what one says is not to intend to induce
any perlocutionary effect in one's audience.

This theory of meaning is inadequate. Since
meaning is best assoc1ated with illocutionary
but not perlocutionary intentions, the analysis
offered is an analysis of an incorrect expli-
candum. The analysis offered is an analysis of
'trying to accomplish something by uttering
something' rather than 'uttering something and
meaning it'. However, an adequate theory of
meaningl must examine something like the latter
analysandum.

This theory of meaningl is counter-intuitive
since a number of effects, 2! can be achieved

as intended perlocutionary effects of utterances
given that I’is able to correctly anticipate I's,
perhaps mistaken, estimation of Ifs intentions.
In such cases I will have manipulated I to
achieve 3! but we should not want to say that by
uttering I I meantl I.

 

123

The final group of problems directs itself against the
guiding assumption that a theory of meaningl which lists
only intentions of the speaker in its analysans can be an
adequate theory. There is more than just intending involved

in speaking meaningfullyl, or so it seems. Group six is:

Problem Q This analysis of meaningl neglects the fact that
in speaking meaningfullyl I_must select a token
which he believes to have some feature (generally
its meaningz) by virtue of which I will be able
to understand what I meansl, unless I is not
trying to communicate with I.

Let us move to Chapter III to see whether or not

Grice is able to handle these problems.

FOOTNOTES TO CHAPTER II

1H. P. Grice, "Meaning," Philosophical Review, LXVI
(1957), 337—88. This article is also contained in Philo-
sophical Logic, ed. by P. F. Strawson (London: Oxford
University Press, 1967), 39-48; also republished in Problems
in the Philosophy of Language, ed. by Thomas M. Olshewsky
(New York: Holt Rinehart and Winston, Inc., 1969).

2John Holloway, Language and Intelliggnce (London:
Macmillan, 1951).

3H. L. A. Hart, "Signs and Words," Philosophical
QuarterIy, II (1952), 59-62. A review of Language and
Intelligence by John Holloway, Ibid.

4

 

 

 

 

 

 

 

 

H. L. A. Hart, "Signs and Words," gp. cit., p. 61.

5N. L. Wilson, "Grice on Meaning: The Ultimate
Counter-Example," Nous, IV (1970), 296.

6H. L. A. Hart, 9p. cit., p. 62.

7H. P. Grice, "Utterer's Meaning, Sentence-Meaning,
and Word-Meaning," Fougdations of Language, IV (1968),
225-42. H. P. Grice, "Utterer's Meaning and Intentions,"
Philosgphical Review, LXXVIII (1969), 147-77.

8H. P. Grice, "Meaning," op. cit., p. 39. (All
references to this article are made in these notes to corre—
late with the pagination of its reprinting in P. F. Strawson's
Philosophical Logic, IE. III.)

9

 

 

 

IIII., p. 40.
10

italics.)
11

H. P. Grice, "Meaning," gp. cit., p. 43. (My

Ibid., p. 42.

12Ibid., section 1.3, pp. 12-15. See also Benson
Mates, "On the Verification of Statements about Ordinary
Language," Ordinary Language, ed. by V. C. Chappell
(Englewood Cliffs, N.J.: Prentice—Hall Publishers, 1964),
pp. 64-74, for a discussion and critique of this general
position.

 

124

125

l31bid., p. 45.

l41bid., p. 46.

15G. H. R. Parkinson, ed., The Theory of Meaning
(London: Oxford University Press, 1968), p. 184.

16H. P. Grice, "Meaning," 9p. cit., p. 46. Grice
adds: "It may not always be possible to get a straight
answer involving a 'that' clause, for example, 'a belief
that . . .'."

17Ibid., p. 46. Notice the ambiguity here, there
is no reason in the text that suggests that we restrict
the range of 'x' in these expressions to utterance-tokens.

18

 

 

Ibid., p. 47.

19Ibid., p. 48.

20William P. Alston, "Linguistic Acts," American
Philosophical Quarterly, I (1964), 138-46. In a letter
Alston says "I am afraid that I do not have anything over
and above the brief references in publications to which you
refer, except for very scratchy lecture notes." (Letter
dated: July 17, 1970.) Max Black, in a letter, says "I am
sorry to say that my notes on Grice's views are not in a
fit state to be seen yet." (Letter dated: May 4, 1970.)
Both men indicate that they plan to write essays in detailed
examination of Grice's views at some future date.

21William P. Alston, "Linguistic Acts," 0 . gII.,
p. 142. The man could clear things up by saying I meant
to say 'battery' not 'beagle'." Alston says, "When the man
sees his mistake he will say, 'Oh, what I meant to say was
that my battery was dead.‘ Note the indirect discourse
form. He is not just saying that he meant to utter 'My
battery is dead', but that he meant to say that (or tell us
that) his battery was dead." Alston's dist1nctions are
crude versions of our own; but his comment about using
indirect discourse is puzzling. Grice will make much of
this form of expressing meaning in Chapter III.

22Ibid., p. 142. There quoted from Ludwig
Wittgenstein, Philosophical Investigations (New York:
Macmillan, 1953).

23

 

 

Ibid., p. 144.

24Paul Ziff, "On H. P. Grice's Account of Meaning,"
Analysis, XXVIII (1967), 1-8. Ziff's comments have been
utilized by D. S. Clarke, Jr. in his paper "Meaning, Force,

126

and Rhetorical Effect," delivered before the American Philo-
sophical Association at its December 1970 convention. This
paper is abstracted in The Journal of Philosophy, LXVII
(1970), 828-29. Clarke offers a counter-example like

number 3. It proves that he does not notice the meaning /
meaning distinction either. The counter-example appears

on page; 2 and 3 of Clarke's manuscript; it is not generally
available.

25Paul Ziff, "On H. P. Grice's Account of Meaning,"
9p. cit., p. 7.

26

 

Ibid., p. 5.

27T. E. Patton and D. W. Stampe, "The Rudiments of
Meaning: On Ziff on Grice," Foundations of Language, V
(1969), 2-16.

 

281b1d., p. 15.
291bid., p. 3.
30Ibid., p. 4.
31

W. V. Quine, "Quantifiers and Propositional Atti-
tudes," reprinted in his The Wgys of Paradox (New York:
Random House, 1966), p. 189.

32

 

For example see Paul Ziff, 9p. cit., p. 4.

33Ibid., pp. 2 and 4. As usual I have introduced
the appropriate subscripts into the discussion. Ziff's
original phrasing would be that meaningless sounds can be
made to meannn, and meaningful things are not considered
meaningful .

nn

34Ibid., p. 2.

35Ibid., p. 4. In indicating the actual force of
Ziff's counter-example I have taken liberties with his own
views on these topics, especially with regards to the cited
phrases.

36F. F. Strawson, "Intention and Convention in Speech
Acts," Philosophical Review, LXXIII (1964), 439—60.
Reprinted in Problems in the Philosophy of Language, ed. by
Thomas M. Olshewsky, 9p. gII., pp. 259-75. A11 citations
of this article refer to the pagination in Olshewsky's
reprinting.

37

 

 

Ibid., part III, pp. 264-67.

127

38Ibid., p. 264.

391bid., p. 265.

4OIbid., p. 266.

lReviewing Dissertation Abstracts through 1970 we
find only two doctoral dissertations that treat of Grice's
theory of meaning. Both treat it in the context of discus-
sions of other topics and both treat it as a cooperative
effort conducted by Grice and Strawson. Each of the two
examine the Strawsonian revision of Grice's theory that I
have just presented, and each claims to have found diffi-
culties with the revised analysis of meaningn . Stewart
Thau, Linguistic Acts and the Concept of Mean1ng, Doctoral
Dissertation, University of Michigan, Ann Arbor, Michigan.
Abstract on 2085-A, Vol. XXX (1969), Dissertation Abstracts
(Ann Arbor, Michigan: University Microfilms Co.). Alfred
Farnum MacKay, Speech Acts, Doctoral Dissertation, Univer-
sity of North Carolina, Chapel Hill, North Carolina.
Abstract on 4665-A, Vol. XXVIII (1968), Dissertation
Abstracts. MacKay's examination of various recent versions
of theories of meaning that, like Austin's, involve some
concept of a speech act ends with the comment that the
intentions of speakers and the conventions of language make
competing claims with respect to an adequate theory of
meaning. I agree with him on this. Perhaps, too, a compro—
mise is possible.

42

 

 

 

 

 

Ibid., p. 271.

43IIII., see section V, pp. 271—75,

44John R. Searle, "What is a Speech Act?" in Philos—
ophy in America, ed. by Max Black (London: George Allen &
Unwin Ltd., 1965), pp. 221-39. John R. Searle, Ipeech Acts:
An Essay in the Philosophy of Languagg (Cambridge: Cambridge
University Press, 1969).

45John R. Searle, Speech Acts: An Essay in the
PhilOSOphy of Language, gp.‘gII., p. 43.

46

 

 

 

 

 

Ibid., p. 44.

47Ibid., p. 45.

48Ibid., p. 47.

49Ibid., p. 49f.

50O. H. Green, "Intentions and Speech Acts," Analysis,
XXIX (1969), llOff.

128

51D. S. Clarke, "Meaning, Force, and Rhetorical
Effect," 9p, cit., as abstracted in The Journal of Philos-
ophy, 828f.

 

CHAPTER III

THE DEFINITION OF MEANINGl

Introduction

 

Grice published "Utterer's Meaning, Sentence-
Meaning, and Word—Meaning ("UMSMWM") in 1968. It outlines
his theory of meaningnn and elaborates some of the details
of the middle stages of this program.1 In 1969 he published
"Utterer's Meaning and Intentions" ("UMI").2 Here he tries
to provide a redefinition of meaningl in terms of inten-
tions. He hopes to avoid the objections and counter-examples
that several philosophers have presented.

Certain problems prevent formulation of a complete
and definitive presentation of Grice's work and the comments
it fostered. First, many of Grice's critics never published
their counter-examples and objections. Many of the argu—
ments Grice responds to in "UMI" were conveyed to him in
conversation. If this pattern of informal critique is
repeated, we can expect that little will be published con-
cerning Grice's two latest articles. To date only N. L.
Wilson's work is available to philosophers at large.

Grice's reticence to publish is another problem.

No doubt he does deliver public lectures on these and
related topics.4 But few of these lectures find their way

into print. In notes to both recent articles he promises

129

130

that he will revise and restate the views presented in
these articles in a future book, which I presume to be
largely devoted to the topic of meaning.5 The articles,
which are all we have to work from, are thus not considered,
even by Grice, to be his best work.
What I have to say on these topics should be looked
upon as an attempt to provide a sketch of what might,
I hope, prove to be a viable theory, rather than as an
attempt to provide any part of a finally acceptable
theory.6

Since my general aim is to discuss this kind of
theory of meaning, rather than simply Grice's version of it,
the problems which prevent us from knowing Grice's views in
detail do not prevent us from knowing them in outline.

From this knowledge we can infer what kinds of considera-
tions this type of theory seems to necessitate, and what
kinds of problems it seems to encounter.

Grice's two articles are not consistent in every
detail; moreover several notational or minor logical errors
appear in the presentation of the various definitions of
meaningnn. Much of this chapter is a reconstruction of
Grice's efforts to provide a definition of meaningl in terms
of a speaker's intentions. Wherever possible I take the
liberty to make alterations in the notation or logical
structure of Grice's definitions so that these details do
not obscure the main issues. Unlike the usage in the pre-

vious chapters, here 'A' ranges over audiences and 'U' over

speakers, utterers, or authors. This chapter has five

 

131

sections, beside the introduction and conclusion. The
following section develops Grice's four-fold distinction of
the senses of 'meaningnn'. Two sections are devoted to
Grice's efforts to define the primary sense of meaningnn--
"utterer's occasion meaning." The first of these reviews
and comments on Grice's efforts to show that his analysis

of meaningl can be expanded to overcome objections that it
is too weak. This section has six parts; a part dealing
with preliminary matters is followed by five parts devoted
to the successive redefinitions of meaningl.. The second
section deals with Grice's efforts to show that his analysis
is not too strong. This section also is subdivided to
present different redefinitions of meaningl. Then follows

a section offering final assessment of Grice's redefinitions
in the light of the problems developed in Chapter II. This
section also deals with the problems that face any such
analysis of meaningl. Finally an analysis of meaningl is
offered. This analysis is in the spirit of Grice's work,
but is not restricted by the requirement to list only inten—
tions in the analysans. This analysis also, it is hoped,

avoids the problems that plagued Grice's own definitions.

 

Senses of Meanlggnn
The first step to the four-fold distinction is
Grice's apparent recognition of our distinction between mean-
ingl and meaningz. In "UMSMWM" Grice distinguishes between

locutions of these two forms:

 

132

(G7) I meant that . . .

(G8) X (utterance—type) means ". . ."

Each expression is a way of specifying meaningnn. The first
relies on indirect discourse and the second on direct quota-
tion. Grice thinks that this grammatical difference makes
an important semantic difference. He thinks that a different
kind of meaning is being specified when one uses quotation
marks rather than indirect discourse.7

I would argue that the grammatical difference that
Grice refers to is irrelevant. There EEE different senses
of 'meaning' employed in (G7) and (G8). The first seems to
express meaningl, and the second meaningz. But these two

senses of meaning may be expressed using both indirect dis—

course and quotation. Consider:

(1) When Abraham says "Jacob, get yee be gone" he means
"Jacob, go away."

(2) What does 'The cat is on the mat' mean? It means
that the cat is on the mat.

Although expressions like (1) and (2) are found less fre-

quently than ones like:

(3) Abraham means that Jacob should go away when he says
"Jacob, get yee be gone."

(4) The sentence 'The cat is on the mat' means the cat
is on the mat.

or perhaps

(5) The sentence 'The cat is on the mat' means "the cat
is on the mat."

133

(l) and (2) are grammatically acceptable.

(G7) reflects meaningl and is used, says Grice, to
speak of what a particular speaker, I, meant by an utterance—
token on a particular occasion. According to Grice the
paradigmatic expression of standard meaning (meaningz) is
(G8). It indicates what an utterance—type means.

A full specification of the timeless meaning of an

utterance-type, I, is a list of all the meanings it has

2

(or ever had, or will have). The meaning2 of a particular
token, I, of I would not be the same as the timeless meaning

of I. Rather it would be one of the meanings2 from the

list. The "appropriate" one is called the "applied time-
less meaning" of the utterance-type I. Grice's first two

notions of meaningnn are:

Meaningl Timeless meaning of an utterance-type, I (where
I is either complete or incomplete) as specified
by the locution, 'I means ". . .".'

Meaning2 Applied timeless meaning of an utterance—type,
X (where X is either complete or incomplete) as 8

§pecified_by the locution, 'I meant here . . .
The third notion is "utterance—type occasion mean-
ing." It is possible, though not normal, that a person

mean something by a token I other than what I meansz. Con-

1

sider (1) and

(6) Private, when I say "Private, would you like to clean
these weapons for me?" I mean "Private, clean these
weapons!"

Each is a specification of the occasion meaning of the token

used. The first, (1), draws on the meaning1 of the

134

utterance-type and would be identical to a specification
of the meaning2 of that utterance-type. The second, (6),
draws on the meaningl of the author. Grice provides for
the derivation of the meaningnn of an utterance-token from
what its author meansl by it. This is useful when I does
not meanl what I_meansz.
Grice says that although one generally uses the

"conventional implicature" of the utterance-type to deter-

mine what the speaker's intentions are, these intentions

 

determine the occasion meaning of the utterance-type. This

 

 

kind of meaningnn would then be a specification of meaningl,
if Grice's definition proves adequate. Thus, we would be

able to speak of what "I means ," or, oddly, "the meaningl

1

of I." That would be whatever I meant by uttering I of

1
type Ibon that particular occasion. This peculiar notion

of meaningnn is not without its counterparts in idiomatic
English. Suppose I is reading a letter from I and he says
in wonder "What does this mean?" Here I is using an expres-
sion that may be short for 'What does I mean by this'. I

is indicating his readiness to distinguish between what I's
letter means2 and what Imeantl by what he wrote in the

letter.

Grice gives his final two senses of meaningnn as:

Meaning3 Utterance-type occasion meaning, as specified
by the locution, 'I meant ". . ." by I'.

Meaning4 Utterer's occasion meaning, as specified by the
locution, 'I meant by uttering I that . . .'.

135

Again Grice notes that the first three kinds of
meaningnn call for the use of quotation marks, whereas the
latter calls for indirect discourse. I believe Grice is
mistaken in the emphasis that he makes of this syntactical
peculiarity.

Having made the four-fold distinction, Grice hopes
to define meaning4 by listing only I's intentions. On the
basis of this, the other three kinds of meaningnn will be
defined as indicated in Chapter I. Grice recognizes the
artificiality of his distinction commenting:

There is, of course, an element of legislation in the
distinction between the four cited linguistic forms;

these are not quite $8 regimented as I am, for con-
venience, pretending.

Defining Meaning4, I

 

Preliminaries

 

Grice hopes to define
(G9') 'I meant by uttering I that :Ip.'
At times he represents (G9') by
.10

(G9) 'I meant by uttering I that :p.

Often Grice presents this concept in still other forms.

(G10) 'I meant by uttering I that . . .' (as in Meaning4)9
(G11) 'By uttering I I meant that p.'11
(G12) 'By uttering x U meant something.‘ (Used in 1957)11

 

Notice that in each case the past tense, 'meant', is used.

Perhaps this is only an accident of the idiom; perhaps Grice

 

136

has reasons for not using the present tense. It seems,
though, that his definiens is something like "I uttered I
intending thus—and-so." Perhaps then, it would be a trivial
change to use the present tense, 'means', and alter the
definiens to something like "I is uttering I intending
thus—and-so."

The device '*Wp' is peculiar and Grice's explanation
of it is garbled. Moreover his explanation reveals certain
over—simplified and false views about the relationships
between grammar, the purposes to which people put their
language, and propositional attitudes. Grice presupposes
that there is one propositional attitude associated with
each English mood. For example, believing is associated
with the indicative mood, intending (i.e., getting the
audience, I, to intend to do something) with the impera-
tive.12 Further, he presupposes that each mood has only
one primary function.12 For example, the indicative is used
for making assertions, the imperative for issuing commands
(exhibiting that it is the speaker's intention that such—
and—such be accomplished).12

Grice uses 'W' as a variable ranging over proposi-
tional attitudes; for example, believing, doubting, intend-
ing. The device '*' designates the function that maps each
propositional attitude into the mood with which it is "cor-

related." (Here appears the first peculiar presupposition.)

Thus, '* believing' yields the indicative mood; if the value

 

137

of 'W' is doubting, then '*W' yields the interrogative mood;
and if the value of 'W' is intending, then '*W' marks the

imperative mood. Where 2 is a proposition, *Wp is the

 

result of transforming 'p' into the grammatical mood 13.13
For example, let p_be 'it is raining'. I believe Grice
would have us make transformations like these:

Prop. Attitude Sentence
(A) Believing = W *Wp = 'p.‘ It is raining.
(B) Intending = W *Wp = 'p!‘ Make it rain!
(C) Doubting = T *Wp = 'p?‘ Is it raining?

It seems to be a moot point that we can exhibit
a "propositional attitude" by using more than one mood. We
might indicate belief, for example, by using an expression
like "Don't all sane men think that 2?" Thus, the first
presupposition mentioned above is mistaken. The second,
that each mood is used for only one primary function, seems
equally mistaken and restrictive. The indicative mood can
be used to make assertory utterances like predictions,
explanations, descriptions, statements, etc.; but it can
also be used to make promises, to offer advice, to issue
threats and warnings, to bestow names, to offer praise, to
show gratitude, etc.

We might follow Searle's suggestion here. One can
use notational devices if one wishes, but these might better
range over illocutionary forces rather than propositional

attitudes.l4 We might try something like this: Let '*'

 

138

range over those devices by which we indicate a particular
illocutionary force; for example, 'k—' indicates asserting,
'W' indicates warning, 'Pr' indicates promising, '?' indi-
cates asking a direct question, '1' indicates requesting,
etc. Then, we might let '*p' be the result of transforming
E into any grammatical expression appropriate for *-ing
(that) 2.

Perhaps the over—simplified presuppositions led
Grice to slur over the illocutionary/perlocutionary distinc-
tion even in these recent articles. The examples Grice will
use indicate that his fundamental concern is with perlocu-
tionary effects of utterances. Using the indicative mood

the speaker is not so much telling A that p but getting A

 

to think that g believes that p or getting A_to think that
15

 

p by virtue of his realization that g believes that p,

In using the imperative g is trying to get A to do something

 

by virtue of A's realization that g intends that Aydo that
thing. We can speculate that, had Grice been less con-
cerned about propositional attitudes and the "primary pur-
pose" of each English mood, he might be less concerned with
"forcing" E's intentions into a perlocutionary mold. Grice
wishes to say, however, that in every case, Q's utterance

of x is intended to have the (perlocutionary) effect of

 

generating a propositional attitude in A. "The way is

 

opened to a simplified treatment of the M—intended effect, as

being always the generation of some propositional attitude."l6

139

Still another unfortunate aspect of (G9') is the
apparent presupposition that every meaningful utterance
must have some propositional content p. Notice, however,
that this presupposition is not unreasonable given that
Grice is prepared to require that each of gfs meaningfulnn
utterances is intended to foster some propositional atti-
tude in A. But some expressions do not indicate any propo—
sitions. For example, consider 'Hurrah', 'Ouch', and
'Hello'.17 In their normal uses, the speaker is not expres-
sing any attitude, or trying to induce any attitude, toward
any proposition. But, we should not want to say that he
did not meanl anything when he uttered them. Perhaps some

would argue that they do not mean anything. This may be

2
true, but since it goes into the matter of the nature of
meaning2 it is beyond the scope of the present work. It
appears that we should not require that x have some propo-
sitional content when determining whether or not gmeantl
something by uttering it. Perhaps Grice could compromise
by replacing his '*Wp' with something like '*W...'. Here
the lacuna can be filled, in most cases, with a proposition.
But at times one might render "intending to *W..." as

"intending to greet" or "intending to express pain, etc.
Yet this move might be troublesome in that the definition
of meaningnn would be too strong. Except in the case of

greeting g_might not intend any effect in any audience

when he utters one of these kinds of expressions. But in

140

this respect these expressions are troublesome no matter
how Grice alters his position on propositions.

Grice begins his analysis of (G9') by recasting his
1957 explication of (G12). What Grice hopes to show is that
the speaker's intentions (or complex Meaning-intention), are
adequate to be the "supposed link" between the intended
responses of the hearer, A, and what the speaker, g, meant
by uttering x. Adopting Strawson's suggestion, Grice
replaces 'effect' with 'response' and summarizes his 1957

definition of (G12).

(M') U meant something by uttering x if and only if

 

there is some audience, A, and some response, 3,
such that g uttered x intending:

(i1) to induce r in A

(i2) that A recognize that g intends (i )
(i3) that A fulfill (i1) 8? the basis oi his

fulfillment of (i2).

I gather that "Afs fulfillment of (in)" is to be understood
as Afs doing whatever the intention (in) indicates that 2
would have A do. Clause (i3) indicates that Afs thinking
that g intends him to respond with r_be some kind of condi-
tion (or, as Grice would have it, "reason") for A's respond—
ing with r. If £_were merely a causal "result" of (i2) we
would have a counter-example in:

Counter—example 6 (Grice, reconstructed): Suppose that

a performer, g, intends to amuse his audience, A. He

intends, further, that A realize that U wishes to amuse

him, and only that A_be—amused at least in part because
of (as a result of) this realization.ll

141

If, when uttering x (a joke), the performer had these three
intentions, should we say that he meant4 something by utter-
ing x? Notice he has (il) and (i2) and a weaker version of
(i3), viz., that A fulfill (i1) as a result of his fulfill-
ment of (i2). Grice answers "No." He says, "Though A's
thought that g intended him to be amused might be a part-
cause of his being amused, it could not be a part of his

reason for being amused."l8

Thus, according to Grice, (i3)
as stated rules out such cases.

Grice's reply is a slip, one that is easy to make.
His appeal to how A might in fact respond, and why, is
irrelevant. As we noted in Chapter II, the considerations
of importance are not what A does, but what A intends that
A do. However, I believe that Grice is correct to adopt
the stronger (i3). If we are to speak of A's intended
response, it seems appropriate to intend that the response
be within his control. (Of course, I am not happy with
requiring that g intend to achieve some perlocutionary

effect-~the production of r in Af-whenever g speaks mean-

ingfullyl.)

Redefinition I

 

In conversation J. O. Urmson offered this counter-
example to Grice's first analysis:

Counter-example 7 (Urmson): Suppose a prisoner of war
to be thought by his captors to possess some information
which they want him to reveal; he knows that they want
him to give this information. They subject him to tor-
ture by applying thumbscrews. The appropriate analysans

142

for 'They meant something by applying the thumbscrews
(that he should tell them what they wanted to know)‘
are fulfilled: (1) They applied the thumbscrews with
the intention of producing a certain response on the
part of the victim; (2) They intended that he should
recognize (know, think) that they applied the thumb-
screws with the intention of producing this response;
(3) They intended that the prisoner's recognition
(thought) that they had the intention mentioned in (2)
should be at least part of his reason for producing
the response mentioned.
Grice interprets counter-example 7 as a case where
g intends to induce E in A, but fails to intend that the
utterance of A play some part in this process. Thus, Grice
appears to be sympathetic to the conceptual change Strawson
introduced in his formulation of (i1). (See analysis [S],
Chapter II). Strawson requires that g intend to induce r
by means of uttering x. Grice, on the other hand, tries to
provide for the role of the utterance in a reformulation of
clause (i2). According to Grice, counter-examples like 7
can be avoided if, instead of g intending merely that A
recognize in (i2) that g intends (i1), 2 rather intends

that A's recognition derive "at least in part from the

utterance of x." This restricts the intended source of A's

 

information. It is not clear whether Grice wishes the fact
that g utters x to be the source A must intend A_to use, or
whether the x g utters is to be the source.

As a first try we may interpret the restriction to
require that E intend A to recognize (i1) by virtue of some

feature, say the meaning2 of the utterance-token 5° Mean-

ing2 can be broadly construed as deriving from conventions

143

either implicit or explicit. I conjecture that such con-
ventions play at least some rudimentary role in every mean-
ingful speech act (and perhaps every communicatory act
whether it relies on the use of language, gesture, or the
speaker's habitual signal behavior). If this were so,
Grice's work might be undone--meaning2 would be independent
of and more primitive than meaningl. My conjecture may
fail, however, in certain "first time" cases. Suppose 2
produces x, and producing x is a non-linguistic gesture or
action. Suppose that this is the first time that g pro-
duced x to be a sign toward A, and it is not Q's habit to

produce x. Either A_must ask A_what he means by x, or A

1

can guess what 2 means by x. If A guesses he would probably

1
be estimating E's intentions without knowledge of the mean-
ing2 of x. Yet even here Afs actions will be as much a way
of testing his estimation as they are results of that estima-
tion.

The second interpretation focuses our attention not
on the "utterance" as an utterance-token, but as the act
performed by g. Interpreted in this way Grice's restriction
does not require that A read the meaning2 of x. Rather A
is intended to base his estimations of Q's intention (i1) at
least in part on some facet of g's production of x. In

counter—example 7, it is not the "meaning2 of applying some
method of torture that A should focus on, but the fact that

the captors are torturing him. But, if g intends that A

144

recognize their intention to have him reveal some informa-
tion at least in part because they are torturing him, then
Grice's new restriction is satisfied. However, the assump-
tion is, in counter-example 7, that A already knows his
captor's intentions, the torture is simply an inducement to
act.

This interpretation of Grice's restriction puts off,
but only for a time, the difficulty that the first interpre-
tation suggests. One must inquire about what feature of
Q's uttering x is the cue to A that g intends £° It seems
that the meaning2 of the utterance-token g produced is

generally the intended cue. If this is not the case, then

 

g might well have uttered y as well as §° If g intended to
startle A, then the token uttered is of little matter; but

if g intends to startle A this will not be a case of mean-

ingnn, for A would not have all the appropriate intentions.
On the other hand, if g intends to inform A that some com-

plex statement is true (e.g., that the White House is to be
sold at a public auction next Tuesday morning) he would, it
seems, have to use an utterance—token with the proper mean-
ingz. If g intends to communicate, he must choose a token

which he believes A can understand--that is, A must select

a token by which (he believes) A will be able to understand
what he is sayingi. Either interpretation leads to those

considerations that made Problem Q arise.

 

 

145

Which of these two interpretations is Grice's own
seems an unanswered question. His "tobacconist" examples
are explained so vaguely that we cannot tell how to inter-
pret 'utterance'. But, it is a moot point since both

interpretations lead to Problem Q. In any case, since the

 

utterance (ambiguous) was not intended to play some part in
A's recognition of (il) in counter-example 7, Grice offers
the following redefinition with its restriction in clause

(i2):

(RI) U meant something by uttering x if and only if

there is some audience, A, and some response, r,
such that U uttered x intending.

(il) to induce r in A

(12) that A recognize, at least in part from
the utterance of x, that U intends (i1)
that A' s fulfillment of (i2 ) be at leasE
in part A' 8 reason for fulfilling (i1). 0

(13)
Note that Grice has altered (13) as well reflecting his
interest in having U intend that E be a response over which
A has some control, rather than merely an "effect."

Grice could, at any point, deny the authenticity of
a proffered counter-example. This move would preserve his
analysis of meaning4 while shifting the battle ground. The
issue would become how much reliance to place on our intui-
tive or native understanding of meaning. Like other
"abstract" concepts, this one is open-textured, thus our
native understandings will at times differ. There being no
authoritative View to rely on, we shall be asked to accept

or reject the Gricean analysis on the basis of our rejection

146

or acceptance of the purported counter—example. This point

will come, but it has not, I believe, been reached as yet.

Redefinition II, Version A

 

A second kind of counter—example is exemplified by
counter—example 4 (Strawson's) and those suggested to Grice
in conversation by Dennis Stampe and Stephen Schiffer.

Counter-example 8 (Stampe): A man is playing bridge
against his boss. He wants to earn his boss's favor,
and for this reason he wants his boss to win and
furthermore, he wants his boss to know that he wants
him to win. (His boss likes that kind of self-
effacement.) He does not want to do anything too bla-
tant, however, like telling his boss by word of mouth,
or in effect telling him by some action amounting to a
signal, for fear the boss might be offended by his
crudity. So he puts into operation the following plan:
when he gets a good hand, he smiles in a certain way;
the smile is very like, but not quite like, a sponta—
neous smile of pleasure. He intends his boss to detect
the difference and to argue as follows: "That was not
a genuine give-away smile, but simulation of such a
smile. That sort of simulation might be a bluff (on a
weak hand), but this is bridge, not poker, and he would
not want to get the better of me, his boss, by such an
impropriety. So probably he has a good hand, and want-
ing me to win, he hoped I would learn that he has a good
hand by taking his smile as a spontaneous give-away.
That being so, I shall not raise my partner's bid."21

Of all the counter-examples, this is the most
curious. According to Grice we should not want to say in
such cases that the employee meant4 anything by the smile,
not even that he had a good hand or that the boss should
think so. Yet his intending that his boss should think so
satisfies (i1). Since the employee wanted the boss to
recognize, partly because of the sort of smile used, that

he intended him to think so, he satisfies (12). He also

 

 

147

satisfies (13) by intending that the boss's reason for so
thinking be in part based on the success of (i2).

We can note in passing that Grice's analysis of the
(i2) intention sheds little light on the act/object ambi—
guity of 'the utterance of x'. If the act is Grice's focus
then why did he mention the character of U's smile; but if
the object is the focus in (12), why not use the phrase
'the utterance x'? Further we must note the oddity of the
Stampe example. The bridge player is trying to signal to
his boss that he wants him to win, and to know that he wants
him to win. But the player is trying to signal without
appearing to be signaling. The situation is implausible
for if the simulation of a give-away smile is to be taken
as a signal, the simulation is too transparent.22

Strawson's counter—example, which is similar in
strategy, was resolved when the condition, (i4) that U
intend that A recognize U's (i2) intention, was added. In
the case of counter-example 8 we have U's intention (i2),
that the boss know that, (ii), U intends that he think that
U has a good hand. Grice claims that here Q does not have
(14); he does not intend that the boss think that he was
intended to recognize U's (i1). U's plan depends on A's
mistakenly thinking that he is too clever for U. A_is
intended to think that U wanted him to take the smile as a

spontaneous give—away. In fact the boss took U's simulation

148

as a simulation while not realizing that this was how U
intended him to take it.

I find it hard to accept Grice's interpretation of
this counter—example. Nevertheless, he wishes to argue
that it is ruled out by Strawson's analysis. I believe
that Grice would accept the following as a statement of

Strawson's analysis:

(8') U meant something by uttering x if and only if

 

there is some audience, A, and some response, E!
such that U uttered x intending:

(11) to induce r in A

(i2) that A recognize, at least in part from
the utterance of x, that U intends (i1)
3) that A's fulfillment of (i2 ) be at least

in part A's reason for fulfilling (il )
(i4) that A recognize that U intends (i2 ).

(i

Unfortunately another counter-example is available to under-
mine analysis (S') as well.

Counter-example 9 (Schiffer): U is in a room with a
notoriously avaricious man, A, but A is a man with some
pride. U wants to get rid of A. So, U, in full View
of A, tosses a five— —pound note —out of the window. He
intends that A should think as follows: "U wants to
get me to leave the room, thinking that I shall run
after the five-pound note. He also wants me to know
that he wants me to go (so contemptuous was his per-
formance). But I am not going to demean myself by going
after the banknote; I shall go, but I shall go because
he wants me to go. I do not care to be where I am not
wanted."23

Here U has all four intentions. He tossed the money intend-
ing: (i1) that A should leave the room, (12) that A should
think that U wanted him to go, (i4) that A should realize
that U intended A to think that U wanted him to leave

(Strawson's clause), and (13) that A should fulfill (i1) by

 

 

149

reason of his fulfillment of (i2). Either U meant4, by
throwing the money out the window, that A was to leave, or
a further restriction is needed. A's mistake is the key
to the new restriction. A'is intended to think that his
departure was intended to be in pursuit of the money, but
it was not. U intended A to think this, but did not intend
that A realize that this was what U intended him to think.
If A had seen through U's plan, seen his concealed inten-
tion, he would have realized that his departure was not
intended to be in pursuit of the money.

Of course, what A_does think is irrelevant to
whether or not U meant anything by uttering x. Just as

1
whether or not A does 3 is irrelevant. But Grice is sug-
gesting that what is not irrelevant is what U intends A.to
think. U_must intend that A think a number of things, and
it seems also that U should intend that A recognize 211.0f

U's intentions concerning what he wishes A'to think. Thus

Grice presents the "A" version of his second redefinition:

(RII, VA) U meant something by uttering x if and only if

there is some audience, A, and some response, 3,
such that U uttered x intending:
(i1) to induce r in A
(i2) that A recognize, at least in part from
the utterance of x, that U intends (i )

(i3) that A recognize that U intends (i1)

(i4) that A' s fulfillment of (i2) be at1
least_ in part A's reason for fulfilling
(il)

(15) that A think that U intends (i4 ). 24

Note that clause (i4) of (S') has become clause (i3) of (R11,

VA).

150

Redefinition II, Version B

 

The frequency with which clause 2 is a stipulation
that A realize that U intended clause 3:; suggests that an
infinite regress may be discovered. If one were to materi-
alize it would be vicious for there would always be a
counter-example to refute each new analysis of meaning4.

Grice believes that no infinite or indefinite
regress will develop; moreover he believes that should one

25 Let us examine the second

appear it would not be vicious.
issue first. Some regresses are not vicious. For example,
moving from the assertion of p to 'it is true that p' to

'it is true that it is true that p' and so on is harmless
because each statement, albeit different from the others,

is logically equivalent to each of the others. Thus all

are true or all are falSe together. Other regressions
manifest distinctions in their first stages but in subse-
quent stages seem to be little more than grammatical manipu-
lations. Such is the case with the difference between 'not
p' and 'not not p'. The grammatical difference apparently
loses importance as one moves to 'not not not p' and 'not
not not not 3' and so on. Each stage beyond the first two
is simply equivalent to one of the first two. Perhaps a
better example would be the following. There seems to be a
difference between 'A knows that p' and 'A knows that he

knows that p' which does not appear as one moves to 'A knows

that he knows that he knows that p' and so on. (But this

151

would not necessarily be the case in every system of
epistemic logic.)

We can formalize this regress by using 'Kap', a
schema that is read as the propositional function "E knows
that p." The values of '3' are persons and the values of

'p' are propositions. Thus we have, for Mr. A:
(8) KAp, and KA(KAp), and KA(KA(KAp)), . . .

Similarly we can formalize the clauses of Grice's analysis.
Let 'Iap' be read "3 intends that p" and 'Tap' be read "a
thinks that p." Now we can indicate the complexity of
Grice's backward facing clauses. Clause (11) of (RII, VA)
requires that U intend to "induce r in A," call this 'fAr'.
Thus we have 'IUfAr'. We can represent (i2), in part, by

18

'IU(TA(IUfAr))'. We can conjoin these two with the third

clause to represent the first three Gricean intentions by:
(9) IU(TA(IU(TA(IUfAr)))) & IU(TA(IUfAr)) & IUfAr

The clause (i4) is that U intend that fAr be based on, or

conditioned by, A's thought that IUfAr. Thus we have for

clause four 'IU(TA(IUfAr)<:>fAr)'. (Where '<:>' is read
"is a condition for.") Adding clause (i5) we have:
(10) IU(TA(IU(TA(IUfAr)<:>fAr))) & IU(TA(IUfAr(:>fAr)

It is by no means clear that the kind of regression
that (9) or (10) suggests is as harmless as that suggested
by (8). It may be possible to disregard more complex

elements like:

 

152

(11) . . . IA(IA(IA(IAp))) or,
(12) . . . TA(TA(TA(TAp))) or,
(13) . . . KA(KA(KA(KAp)))

for these are, at least for practical purposes, semantically
equivalent to 'IAp' and 'TAp' and 'KA(KAp)' respectively.
But if past and future counter-examples suggest anything,
they suggest that one cannot disregard a complex element

like
(14) IU(TA(IU(TA(IU(TA(IU(TA(IUp))))))))

An element like (14) is not equivalent to something like
'IU(TA(IUp))', as counter-examples have indicated. Grice's
analysis threatens to expand conjunction (10) with further

clauses like:

(15) IU(TA(IU(TA(IU(TA(IUfAr)(:)fAr))))) and,

(l6) IU(TA(IU(TA(IU(TA(IU(TA(IUfAr) ©fAr) ) ) ) ) )) , etc.

How many of these clauses will be needed? Is there any
which will be the final clause in the expanded conjunction?
If not, there will always be a counter-example to undermine
Grice's analysis. The regression will be vicious, the com-
plete definiens for 'U meant something by uttering x' will
escape Grice.

The proposed definiens is already objectionable
because it threatens to people "all our talking life with
armies of complicated psychological occurrence." Grice

might have foreseen this problem and postulated a single

 

I. I ‘IIIII‘IIIIIII .‘l'l1
[I‘ll]!!! AI'I

 

153

very complex "meaning-intention" rather than five "regular"
intentions. This move would help to settle the question
'How is it that I do not perceive that I have 2AA of these
many intentions every time I speak, can I have unconscious
intentions?‘ But, it will raise the more difficult prob-
lem "How is it that I do not perceive the complexity of my
intention when I am speaking, can I have a partially uncon-
scious intention?" These questions, in turn, raise prob-
lems. What is "an unperceived intention"? How is the com-
plex "Meaning—intention" to be differentiated from a mere
set of "regular" intentions? Counting intentions is not
easy. U might have to ask himself whether he would agree
that, say, he intended to induce some 5 in some §.bY
uttering E- But whether or not U concedes that he has these
intentions, or others, or only some, or none, Grice's
analysis attributes a complex Meaning-intention to him.

The epistemological problems remain unsettled and
become more disturbing when one notices that, while an
indefinite regress may not in fact materialize, one can,
with little trouble, tack on a number of new mini-intentions
to the Gricean Meaning-intention. Grice argues against the
possibility of producing an infinite or indefinite regress
as follows: A regress cannot get started. For one, the
calculations that would have to be made seem most improbable.
The situations are already complex enough to be nearly mind—

boggling. Consider how much second-guessing, anticipation,

 

154

and manipulation would have to be worked into them to insure
genuine regress. In general, one does not have intentions
to achieve results that one knows he has no chance of
achieving. At best one can only dream of or imagine that
they can be achieved. In human communication the success

of a speaker's intentions or purposes requires and actually
depends on the team work between speaker and hearer. In
response to this argument, one need only recall that we are

not talking about communication but only the intentions to

 

 

induce some response in some audience in a certain way.

 

Taking up the practical argument again, Grice says
that the speaker must suppose the hearer to be capable of
certain very complex inferences already. There seems to be
a limit to what a speaker can reasonably expect, without
having his hearer respond with a look of baffled amazement
as he wonders what the speaker is trying to do. Further,
even if it were possible for A to make the kinds of infer-
ences that it takes philosophers and dramatists days and
hours to work up, it seems that U could not find the cues to
indicate to A that such an abstruse calculation was called
for.26

At some early stage in the attempted regression the
calculations required of §,bY U will be impracticably
difficult; and I suspect the limit was reached (if not
exceeded) in the examples which prompted the addition
of the fourth and fifth conditions.2

This argument is persuasive; it seems that there would be a

practical limit that keeps any infinite or indefinite

 

 

155

regresses from actual realization. But our concern is with
the adequacy othrice's theory, not the practical problems
of enacting a theoretically possible counter—example.

We can test Grice's position by trying to create a
regress. We could try the following pattern for expanding
the Gricean analysis: (a) find the counter-example that
prompted the most recent addition to the current analysis,
(b) alter the situation so that the inference pattern that
U intends that A follow can be expanded by having A recog-
nize U's most complex intention and, (c) let it be U's new,

and concealed intention, that A respond with E in part

 

because of A's newest realization, (d) add, in the typical
3d Egg fashion, a new backward—facing clause to the current
analysis that requires that U intend that his concealed
intention be recognized by A, (e) repeat this process from
step (a). Only practical problems like those Grice noted
prevent the generation of an infinite regress using this
procedure. Each new situation will be counter-exemplary

because each is based on Problem P. Further, not all of

 

U's intentions concerning what he wishes A_to think are
intended to be known to A. It is possible that in such a
situation U could wish to deceive A by virtue of this
intended ignorance.

Let us put this procedure into operation to see how
it works. Counter-example 9 prompted the addition of clause

(i5). We can alter counter-example 9 producing number 9a.

156

Counter-example 9a (Facione): U is in a room with a
notoriously avaricious man, A, Who is a man with some
pride and who is easily insulted. U wants to get rid
of A. So, U, in full View of A, tosses a five-pound
note out of the window. He intends that A should think
as follows: "U wants to get me to leave the room,
thinking that I shall run after the five-pound note.

He also wants me to know that he wants me to go (so
contemptuous was his performance). But I am not going
to demean myself by going after the banknote. I know
that he wants me to go just because I've come to realize
that he wants me gone. How insulting, that he might
imagine that I would not realize this. Well I shall go,
because he wants me gone, and I don't want to be with a
fellow that insults my intelligence."

In counter—example 9a U had all five of the listed inten-
tions. What U wished to conceal from A'is still another
intention: (i6) that A should think that U intended (i5).
Thus we would have (15).
Repeating this procedure we could create another
counter-example.
Counter-example 9b (Facione): Proceed as in counter-
example 9a with the additions that A_is easily hurt
when he realizes that he has been insulted. Suppose
that the inference pattern be altered so that it reads
". . . I know that he wants me to go just because I've
come to realize that he wants me to go. And further
he is trying to insult me so that I realize that this
is exactly what he wants me to think. Well I shall go,
because he wants me gone, and I'm hurt that he tried
to insult me and wanted me to know that he was being
insulting."
Thus we must reveal U's (i7), expressed by (16).
This process could go on first by intending that A
realize U's (i7) because he is perceptive enough to see
that U is trying to hurt him. We might then produce counter-

example 9d by having A realize U's (i8) but not some still

concealed (i9).

157

However reluctant he was to agree that a regress
was either possible or vicious, Grice did take steps to
rule out all counter-examples of the kind we have recently
considered. Counter-examples like 4, and the many 9's call
for U's intending and anticipating some inference pattern

on A's part which contains a mistake that results from A's

 

ignorance of at least one of U's intentions. A may not use
this inference pattern, or A_may realize all of U's inten—
tions, these issues are irrelevant. All that is required
is that U intend that A use this inference pattern and that
A not realize U intends that he use it. Grice puts it
wrongly when he says:
Potential counter-examples of the kind with which we
are at present concerned all involve the construction
of a situation in which U intends A, in the reflection
process by which A'is supposed to reach his response,
both to rely on some "inference-element" (some premise
or some inferential step), E and also to think that U
intends A_not to rely on E.7é
It is a mistake to say that U intends that A think
that U does not want A to rely on E. Such an intention is
not found in any of the counter-examples mentioned, unless
perhaps, it appears in counter-example 8. Grice has appar-
ently confused "U does not intend that A think that U wants
A to rely on E' with "U intends that A think that U does
not want A_to rely on E." If either of these is apparent
in the counter-examples, it is the former and not, as Grice

supposes, the latter. The situation often is that U intends

that A be ignorant of the fact that U actually intends more

158

than A supposes. Grice mistakes intended ignorance for
intended deception. This mistake is reflected in all of
his "Version B" reformulations. These reformulations con—
tain what he calls the "anti-deception" clause. The anti—
deception clause should actually be an "anti-ignorance"
clause. Or, perhaps the Version B reformulations should

contain both kinds of clauses.

(RII, VB) U meant something by uttering x if and only if

there is some A and some E such that:
(a) U uttered E intending:
(11) to induce E in A
(i2) that A recognize, at least in part
from the utterance of E, that U intends
(i )
that A's fulfillment of (12) be at
least in part A's reason for fulfilling
(i ),
(b) it is not the case that for some inference-
element, E, U uttered x intending:
(i4) that—A's determination of the nature of
E should rely on E
that A should think U to intend that
A's dEtermination of the nature of E
should not rely on E.

(13)

(is)
Note that both clause (i3) and (i5) of (RII, VA) are no
longer required in (RII, VB). Perhaps, also, we should
expand the list of curious entities mentioned in Problem E
to include inference-elements.

There may be a way of ruling out both intended
ignorance and intended deception. Thus we would have not
only ruled out counter—examples 4, 8, and the many 9'5, but
also ruled out counter—examples that use the strategy that

calls for A's intended estimation of things to be mistaken

159

in these ways. I suggest that section (b) of (RII, VB) be

replaced by

(b') it is not the case that for some inference-
element, E, and some proposition, q, U
uttered E intending both:

(14) that Afs determination of the nature of
E should rely on E, and, either

(i5) that A should think U’to intend that
A's determination of the nature of E
should not rely on E, or

(16) that A not realize: that U intends
(or believes) that q, and that U
believes that if A knew that U intended
(or believed) that q then A would not
rely on E in his determination of the
nature of E.

I believe that adding (16) in this way will rule out U's
having intentions or beliefs which he wishes to conceal from
A because he thinks that if A knew of these intentions or

beliefs he would decide not to rely on E.

Redefinition III, Version A

 

A feature of most of the counter-examples so far
discussed is that they involve the use of some non-linguistic
utterance-token. Those which incorporate linguistic tokens,
like counter-example 5 (Searle's), present new difficulties
for the analysis of meaning4. Searle's counter—example as
presented in Chapter II is taken from his book. Grice dis-
cussed an earlier version of this counter-example in "Utter-
er's Meaning and Intentions," one which appeared in "What
is a Speech Act?"29 In counter-example 5 the American

soldier hopes to get the Italians to think that he is a

German officer. He utters an arbitrary German sentence

160

hoping they will think, not knowing much German, that it
means2 that he is a German officer.30

Grice is unwilling to take Searle's suggestion and
add an independent element of conventionality to his account
of meaningnn. Grice does not deny that the "conventional
use" of a sentence is normally the key to the recognition of
the intentions of its user. But he wishes to treat the con-
ventional correlation between a sentence and a certain type
of response as just one way of correlating an utterance
with its intended response. He must plead guilty to Searle's
accusation that he treats the meaning2 of an utterance-type
as just another circumstance.31

Grice is not sure that the prototype of counter-
example 5 is genuine. In hard cases like this one the
decision appears to be about the purported counter-example,
but in fact it is about the concept of meaning.32 Grice
has the option of rejecting his analysis of meaning or pro-
posing it as a suggested reformation of the use of the word
'meaning'. In this case, as well as in the case of counter—
example 2, my intuitions do not agree with Grice's. In my
estimation, the general weight of evidence, none of it con-
clusive, favors regarding these as genuine counter—
examples.33

Grice argues that Searle's early version of counter-

example 5 is underdescribed; he tries various ways of amend-

ing it. As presented in Chapter II it is neither

161

underdescribed nor does it lack force; yet Grice's treat-
ment of the prototype is revealing. Grice's first try may
be disregarded because it is based on a number of mistaken
conjectures about how Searle would in fact fill in the
details. His second try approaches Searle's intended inter-
pretation. Grice claims that the American is counting on
the Italians being mistaken about the conventional meaning

of the German sentence. In this case 9y uttering the

 

German sentence the American did mean4 that the Italians
were to believe that he was a German officer. Yet Grice

would not say that the American meant4 by the words that he

 

uttered that the Italians should believe he was a German
officer. The meaning2 (or meaningl) of the sentence is
irrelevant!

Grice argues that U did mean4 something by uttering

x.34 His argument depends upon an analogy. Consider:

(The Port Said Example): The proprietor of a shop full
of knickknacks for tourists is standing in his doorway
in Port Said, he sees a British visitor, and in dulcet
tones and with an alluring smile says to him the Arabic
for "You pig of an Englishman."

Grice says that the proprietor did mean4 that the visitor

was to come into the shop by uttering_E. But, he did not,

 

says Grice, mean4 this byithe words which he uttered.

 

Grice's argument trades on the process/product ambiguity of
utterance, as this ambiguity is preserved in clause (i2) of
(RII, VB). It seems that the proprietor does satisfy the

conditions expressed in (RII, VB), and so, meant4 something.

162

(Note, he would not have satisfied (RII, VB) if (b)'
replaced (b). The proprietor, it is assumed, believes that
q, that his words are insulting by virtue of what they
meanz; moreover, he believes, it is assumed, that if the
British visitor knew that he believed that q the visitor
may not come into the shop. Surely he intends, (i6), that
the visitor not realize these things.)

Grice thinks that the analogy is established. The
proprietor, like the American officer, meant4 something by
his (act of) uttering E, but not by the words that he
uttered. But the analogy is not well founded. For one
thing, there are elements of conventionality in the Port
Said example: the alluring smile, the dulcet tones. To
make Searle's counter-example strictly analogous one would
have to add similar features to the American's utterance of
E, as, for exqmple, its being made in an authoritative tone
of voice.

Other important changes must also be made to make
the analogy successful. The merchant did not intend that
the British visitor understand anything from the Arabic,
the communicatory elements were entirely the tone of voice and
the smile. The British visitor was not even intended to
know or to think that he was expected to know Arabic, to
discern what the merchant wanted him to do. On the other
hand, the American officer intended that the Italians think

that there was some essential feature of the German sentence,

163

viz., its meaningz, which they would be ashamed to admit
that they did not know. Grice's analogical argument that
Searle's counter—example is not genuine, because the
American did mean4 something, fails because the analogy
fails.

Grice's third treatment of Searle's counter—example
is a discussion of my interpretation of it. I suggested
that the American meantl nothing at all by his utterance.31
It was not intended to be communication at all, it was
simply a ruse, a trick. Because the American soldier was
relying on what the Italians might think his German sentence
meantz, rather than, say, on his gestures or tone of voice,
it seems that he was not trying to tell them anything. He

was only playing at trying to tell them something. They

 

might have taken his German to mean anything, from what it

2

does mean2 to what he hoped they would think that it meant

The Italians are to think that they are expected to under—

2.

stand the German, and be unwilling to admit they do not.
Thus U intends to force them into making a guess at its
meaningz. The American means4 nothing by uttering it, how—
ever. He hopes that they will be mistaken about what it

means and that this will be sufficient to get them to think

2

that he means4 something by it, namely, that he is a German

 

officer.

Grice responds, "I do not see the force of this

contention."34 The force of this contention is that if U

164

meant4 nothing by uttering E and still satisfied all the
conditions in (RII, VB), then the analysis is too weak. On
the basis of (RII, VB) U's uttering E is meaningfu14, but
the hypothesis is that it is meaningless4. This objection
is not crucial unless Grice is unable to add further clauses
to (RII, VB) to rule out counter—example 5.
Grice offers another analogy as he tries to show
that my interpretation of counter-example 5 is without
force.
(The case of the Little Girl): I have been listening
to a French lesson being given to the small daughter
of a friend. I notice that she thinks that a certain
sentence in French means "Help yourself to a piece of
cake," though in fact it means something quite different.
When there is some cake in the vicinity, I address to
her this French sentence, and as I intended she helps
herself. I intended her to think (and to think that I
intended her to think) that the sentence uttered by me
meant "Help yourself to some cake." 5

Grice argues, on the basis of this example, that "the fact

that the sentence meant, and was known by me to mean some-

thing quite different is no obstacle to my having meant

36 Grice is correct in this

something by my utterance."
observation, but his argument misses the point. The reason
that someone could mean4 something by such an utterance and
the reason that communication is possible in spite of the
actual meaning2 of the utterance is that the sentence
actually has, in Grice's Little Girl example, another mean-

ingz. Grice was able to communicate with the little girl

because he knew exactly how she would take his utterance.

165

He knew what she thought it meant2 and used it to mean4 just
that.

However, this does not prove that the American
meant4 something for the analogy again breaks down. The
American does not know how the Italian captors will inter-
pret his sentence; that is, the American does not have the
same secure basis that the speaker in the Little Girl
Example has for estimating what A will interpret E to meanz.
The American trusted that the Italians would not know any
of the conventions which would allow them to determine what
the German sentence means . He does not expect before he

2

utters it that they will think that it means something dif—

2
ferent than what he knows it meansz. He expects that they
will be ignorant of what it meansz, that they will not even
have a mistaken idea about its meaningz.

Grice, having found in each of his three tries that
he was unable to understand or formulate Searle's example
in such a way as to make it a counter-example, tries alter-
ing it. By having the American make some gestures, Grice
makes it unnecessary for the Italians to try to guess what
the German sentence means2 on the basis of its being German

and being uttered in these circumstances. Now Grice is

4

prepared to say that the American did not mean that he was

a German officer, but that he only tried to get the

Italians to think "that he meant them to think that he was

"38

a German officer. As revised, this example has become

166

more like the Port Said example. It is not clear why his
conclusions differ in the two cases.

Grice presents (RIII, VA) hoping thereby to rule
out his revision of counter-example 5 and yet retain the
meaningfull utterance in the Little Girl Example. Care-

lessly he omits (i1).

(RIII, VA) U meant something by uttering x if and only if

there is some A, E, E, and 3, such that:
U uttered x intending:

(il ) to induce r in A
that think that
that think that
that think that
that think that
that recognize,
the fulfillment of (i2 ) and (i
that U intends (i1)
that A's fulfillment of (16) be at
least- in part A's reason for fulfilling
(i1)
that A think that U intends (i

has E

intends (12)
correlates E to E
intends (i4)

IEHWHMENW
IGHchx

DJ
('1'
|'-‘
(D
m
(I)
d-
l—l.
:5
'U
9.)
H
(1'
HI
H
O
B

that

pk
V
‘

7).
Note that (il) does not follow from any of the other inten-
tions; it seems odd that Grice would omit it, for it is,
after all, the reason why U uttered E.

The values of 'f' are features of the values of 'x'.
That is, 'f' ranges over characteristics or features of
utterance-tokens. The values of 'c' are "modes of correla-
tion" which associate these features with the response, E.37
(Remember that we argued in Chapter II that the values of
'r' must be universals, "response—types," which U seeks to

have instantiated in A. Thus, a feature of an utterance-

token is correlated by some 9 with some [kind of] response.)

 

167

The most problematic of these new variables is 'c'
for its values are unclear. Grice suggests that a feature
of E may be correlated iconically, associatively, or con—
ventionally with kinds of responses. It seems that Grice
has worked meaning2 into the definiens of meaningl. This
would render his program circular for he intends eventually
to define meaning2 in terms of meaningl! To show how mean-
ing2 has worked its way into the analysis of meaningl con-
sider the following. Suppose that U intends that some A
come to believe that U wishes that the music end. To
achieve this prolocutionary effect U utters "Stop the music!"
His utterance-token has the feature of being an English
sentence. U intends that A think that this is so, and
think that U intends him to think that this is so. U
intends that, and intends A to think that he intends that,
A believe that this feature is correlated by the meaning2
conventions of English with having someone think that some
speaker wants some music stopped. Suppose that U also has
(i6), (i7) and (i8) of (RIII, VA). Thus U has meant4 some—
thing by uttering E. Notice that meaning2 has been used as
a primitive in this analysis of U's meaningl something.

But, perhaps, Grice would not wish to count the
meaning2 conventions of a language as one of the modes of
correlation of features of utterance-tokens with the kinds
of responses that speakers seek to induce in audiences. No

texts support this exclusion, however. The circularity

168

threatened in (RIII, VA) is carried over into all of

Grice's later definitions.

Redefinition III, Version B

 

As formulated by Searle, counter-example 5 does not
satisfy (RIII, VA). I shall argue below that the reason it
fails is because the feature of the token mentioned in (i2)
is not the same as the feature mentioned in (i4) as the
existential quantification with respect to 'f' requires.
This entails that the American did not mean4 something by
uttering E. I prefer this result for I believe that the
American was only playing at trying to tell the Italians
something. He tried to fool them, not to communicate with
them. Thus, I would not claim that he meantl something.

Grice would accept that (RIII, VA) is satisfied
for he thinks that, in the original, the speaker did mean
something. According to Grice (RIII, VA) rules out his own
revision of counter-example 5, yet retains the Little Girl
Example. He supposes the key to this is in requiring that

one and only one feature of the utterance be relied on in

 

(12) and (i4). (If this was Grice's intention, it is not
what (RIII, VA) requires. Again Grice's logic is inadequate
to his plans!). I wish to urge that all Grice needed to
require is that the feature of E mentioned in (12) be the
same as the feature mentioned in (i4); and, moreover, that
the existential quantifier in (RIII, VA) guarantees this

already.

169

Let us look at Grice‘s interpretation first. The
"single" feature operative in the Little Girl example is
that the utterance was a particular French sentence.
Although this is not the only feature of the utterance, it
is the only one that U intended to rely on in (i2) and (14).
The revised counter-example 5 finds U relying on several
features of E. It is German, it is spoken with authority,
it is accompanied with certain gesticulations, it is spoken
in a certain peculiar context. Thus it violates Grice's
intended (RIII, VA). The Italians are to think that E has
still another feature on the basis of noticing these several
features, they are to think that it is a particular German
sentence.

Grice's intended (RIII, VA) is to require that one
and only one feature of E come into play. There are
several problems with this requirement. First, the ordering
of the existential quantifiers is crucial, but Grice does
not indicate how they are to be arranged. In the intended
(RIII, VA) the quantification with respect to 'f' is to
employ a numerically distinct quantifier requiring that
there be exactly one 'f'. Such a quantifier might be placed

first, second, third, or fourth in the list of quantifiers.

However, its position is crucial. Consider
(17) (ElX)(EY) FXY
(18) (Ey)(Elx) ny

Note that (17) implies (18), but (18) does not imply (l7).38

170

Second, there is a problem involved with individuating the
features of E, with counting them and combining them.
Should we say that, for example, the American did not rely
on four features of E, but on one complex feature of E?
Third, the requirement does not seem to account for the
fact that we often rely on several features of E to com—
municate. We may rely not only on the words we utter, but
on tone of voice and context. Perhaps also the connotations
of the words employed or the acoustical characteristics of
these words are counted on as well. For these reasons we
can put aside Grice's unicity requirement. Let us assume
that there is in fact no problem here for the existential
quantifier guarantees that whatever E (simple or complex)
that is relied on in (12) is also the one relied on in (i4).
This is sufficient to accomplish the aims that Grice had in
mind. It is also sufficient to rule out the original
counter-example 5 and yet preserve the Little Girl Example.
To go along with (RIII, VA) Grice provides (RIII,
VB). It makes no effort to capture the ill—conceived unicity
requirement with respect to 'f', nor does it include the
important clause (i1). Furthermore, it may be appropriate
to replace (b) by (b') as was discussed in the treatment of

(RII, VB).

(RIII, VB) U meant something by uttering x if and only if

 

there is some A, E, E, and 9, such that:
(a) U uttered E intending:
(i1) to induce E in A

 

 

 

171

(i2 ) that A think that E has f
(13) that A think that c correlates f to E
(14 ) that A recognize, at least in part

from the fulfillment of (i2 ) and (i3 )
that U intends (i1)
that A's fulfillment of (i4 ) be at
least in part A's reason for fulfilling
(i1),
(b) it is not the case that for some inference—
element, E, U uttered E intending:
(16) that A's determination of the nature
of r should rely on E
that A should think U to intend that
A' s determination of the nature of E
should not rely on E.

(i7)

Notice that (RIII, VB) does not require clauses (i3), (i5)
and (18) of (RIII, VA).

In interpreting counter—example 5 my views did not
agree with Grice's. He held that the American meantl some—
thing. I argued that he did not. Apparently Grice thought
that the American satisfied (RIII, VA). I argued that he
did not. Our views concerning counter—example 2, the only
remaining counter—example, would probably differ as well.

I think that both of us would agree that U satisfies (RIII,
VB). Thus, according to Grice, U meant4 something. I would

agree that he meant something, but I would not take his

1
satisfaction of (RIII, VB) as my reason for saying so. I
believe that we would differ in a specification of what U
viz.,

meantl. I would argue that he meant what he said

1 l'
that A should contact headquarters as planned. Grice would
argue, I believe, that U meant something else, viz., that A

should believe that U believes that something has gone

wrong and that the mission should be aborted. Grice is

 

172

committed to saying that this is what U meant because this
is a specification of the response E that U intended to
induce in his contact A.40
The result of this is that meaning4 something is
not the same as meaningl something. (At least, the some-
things meant in each case are not the same.) Thus, Grice's
analysis is an inadequate analysis of meaningl. There seems
to be no potential clause which if added would rule out
cases like counter—example 2. This counter-example does
not depend on deception nor vicious manipulation of A by U.
It is simply a case of trying to achieve some complex perlo-
cutionary effect, E, as a partial result of an illocutionary

act. We should not wish to say that U meant "that you

1

believe that I believe that something has gone wrong and

that the mission should be aborted." Although, we should
wish to say that U tried to get A to believe . . . . We

have the difference between getting A to believe something
and telling A something. Grice noticed this distinction in
1957, but has been unable to provide an analysis of meaning
which is strong enough to capture the latter and rule out
the former.

If Grice should still wish to argue that in the case
of counter-example 2, and cases like it, U meant4 something
I should reply "fine, but he did not meanl what you claim

he meant4.' The views are not inconsistent. What is

173

unfortunate is that the analysis of meaning4 something is
no longer an analysis of meaningl something.

Version A formulations are open to potential
counter-examples. Version B formulations are as well, but
the counter-examples must not rely on intended deception
or intended ignorance. Of course, Grice always has the
option of claiming that proffered counter-examples are
really not genuine. Yet limitations on the concept of
meaning are ill—advised if their purpose is merely to
salvage a theory. That my conception of meaningl differs
from Grice's is clear; that either of us should proclaim
ours to be the "true" one is a mistake. However, I believe
the weight of evidence offered in the above discussions
shows that Grice's conception has serious flaws. Thus, we
should not be surprised to find that his analysis of

meaningl is mistaken.

Defining Meaning4, II

 

The Gricean analysis of meaningnn has also been
charged with being too strong. It seems to require things
which are unnecessary in some cases of speaking meaning-
fullyl and would, thus, exclude these cases.

We can identify the "M—intended response" as what
U meant4 by uttering E.40 The response is always that A
come to have some propositional attitude. Grice has, as
seen above, associated certain attitudes with certain gram—

matical forms. The response to imperatives is "an intention

 

174

on the part of A_to do such-and—such." The response
intended to indicatives is the belief by A that such-and-
such.12 Some have argued that this correlation is over-
simplified. Grice has been offered counter-examples to
support this claim.41 Nevertheless he persists in his
claims. The effect of this is the suggestion that all

meaningful speaking is the endeavoring to generate propo-

l
sitional attitudes in one's audience. Whenever U utters E
meaningfully1 he is trying to get §.t0 W that p, All mean-

ingfull speaking becomes, under Grice's construal, perlo-

cutionary!

Redefinition IV

 

Other counter—examples suggest that A's response
need not follow in every case by reason of A's recognition
that it is U's intention that A produce E.41 Referring to
analysis (RI), Grice is faced with the dilemma: either
drop (i3) because of these new examples, or retain (i3)
because of other examples. Either choice leaves a set of
counter—examples outstanding. Grice avoids the dilemma by
claiming that the problem arises because of underestimating
the subtlety of E in (11). The M-intended effect of utter-
ing E (where E is in the indicative) is not to "get A’to
believe that p" but "to get A.to think that U believes that
p." A more ultimate effect might be, in some cases, to get
A to believe that p himself. Now it is consistent to

. . 43
require (13).

175

Grice dismisses the following counter-example

thinking that his new remarks on the nature of E take care

of it.
Counter—example lO (Grice): "The Countersuggestible
Man." A regards U as being, in certain areas, almost

invariably mistakEn, or as being someone with whom he
cannot bear to be in agreement. U knows this. U says

"My mother thinks very highly of you." U intend; that
A should (on the strength of what U has said) think
that U's mother has a low opinion of him.4
Here, says Grice, there is some inclination to say that,
despite U's intention that A should think U's mother thinks
ill of him, what U meant was that U's mother thinks well of
A.45

(In many respects this counter-example is like counter-
example 2. There the inclination is to say that U £2222
"tell headquarters . . . ." I believe that here, as before,
U meantl what he saidl and that he was trying to accomplish
something by sayingi that, but that he did not mean1 what
he was trying to accomplish.)

Grice dismisses this counter-example claiming that
A is "intended to think that U thinks that p, though not to

think that 13 himself."45

Let 'p' be that U's mother has a
low opinion of A. Thus we have U intends that A think that
U thinks that U's mother has a low opinion of A, and that A
does not think that U's mother has a low opinion of A.
This, however, is contrary to the stated conditions in the
example. U's intention, ultimate if not primary, was to

have A believe that U's mother has a low opinion of A, and

further, to do so thinking that U does not think that p.

176

Let 'p' be that U's mother has a high opinion of A. Doing

 

this demands that we incorporate, independently of what U
might have intended, the conventional meaning of the utter—
ance itself! Now we have: U intends that A think that U
thinks that U's mother has a high opinion of A, and that A
does not think that U's mother has a high opinion of A.
This analysis fits the example better; but we should have
ended up with "A thinks that not-p" and not "A does not
think that p"! Unfortunately for Grice's resolution, one
cannot infer 'TAnot—p' from 'not-TAp'. In counter-example
10 we have a counter—example that Grice has presented him—
self. And, one that he has not been able to resolve.

Grice generalizes his "resolution" to the dilemma
of retaining or dropping the third condition. He distin—
guishes between "purely exhibitive" and "protreptic" utter—

ances. An utterance is purely exhibitive when U intends

 

to impart the belief that he himself has a certain proposi—
tional attitude. If, beyond this intention, U intends to
induce a corresponding propositional attitude in A by
exhibiting his own propositional attitude, then the utter—
ance is protreptic.45

The two versions of redefinition IV are to capture
the exhibitive/protreptic distinction. A charitable rendi—
tion of what Grice seems to regard as (RIV, VA) follows.
Notice that Grice is using (G9') and not (G12). The reasons

why this reformulated (RIV, VA) is "charitable" are that

 

177

Grice again fails to mention U's primary intention, (i1),
and his distinction between exhibitive and protreptic utter-
ances does not emerge since his versions of (i7) and (19)

are substantially alike.46

(RIV, VA) U meant Ey uttering x that *Wp if and only if

 

there is some A, f, E, and g such that:

U uttered x intending:
(i1) to induce in A the belief that U Y's

that p
(i2) that A think that E has E
(i3) that A think that U intends (i2) .
(i4) that A think that c correlates E with w

the response of believing that U W' s
that p

(15) that A think that U intends (i4)

(16) that A recognize, at least in part from

the fulfillment of (i2 ) and (i4 ) that

U intends (i

that A's fulfillment of (i ) be at

least in part A's reason for fulfilling
(i

(i ) thét A think that U intends (i7)

an , for some cases,

(i9) that A, on the basis of the fulfillment
of (i7), himself W that p.

Says Grice, "the nature of the substitution for AXE" shall
determine whether or not (i9) appears. Grice gives us no
further advice about (i9). Probably he intends that it
appear when and only when the utterance of E is protreptic.
The logic of this definiens is peculiar and it seems that
the best explanation of things is that Grice actually has
two kinds of meaning4 on his hands. In the cases when one
means4 something exhibitively the first eight clauses apply.
Every case of meaning4 something is to count as meaning4

something exhibitively. Moreover, in some cases one can

178

mean4 something protreptically. In these cases one must
have all nine of the intentions mentioned.

Unfortunately for Grice, not every utterance is
exhibitive. For example, suppose that you are visiting
Poland with a friend. Suppose that you do not speak
Polish, but that your friend does. You encounter a stranger
who says "Dzien dobre." You turn to your friend and ask
"What does he mean?" Your friend may reply by telling you
that the stranger meant "Hello." We should not want to say
that the stranger exhibited certain propositional attitudes
by what he said, nor that he tried to get you or your friend
to believe that he has some propositional attitude, W,
toward some 2. Since not every utterance is exhibitive,
Grice's analysis is too strong.

The Version B formulation of redefinition IV is:

(RIV, VB) U meant by uttering x that *Yp if and only if

 

there is some A, E, E, and c such that:
(a) U uttered x intending:
(i1) to induce in A the belief that
that p
(i2) that A think that E has f
(13) that A think that c correlates E with
the response of believing that U W's
that p
(i4) that A recognize, at least in part from
the fulfillment of (i2) and (i 3) that U
intends (i )
that A's fulfillment of (i4) be at least
in part A's reason for fulfilling (i1)
and, for some —cases,
(16) that A, on the basis of the fulfillment
of (i ), himself Y that p
(b) it is not he case that for some inference-
element, E, U uttered x intending:
(17) that A's determination that U Y' s that
B should rely on E

H:

W's

179

(i8) that A should think U to intend that
A's determination that U W' s that B
should not rely on E. 47—

The confident Grice proclaims "whether either ver-
sion of redefinition IV is correct as it stands depends
crucially on the View to be taken of an imperatival version
of "The Countersuggestible Man" example."48

Counter—example ll (Grice): "The Countersuggestible
Woman." Mr. B wishes to be relieved of the immediate
presence of Mrs. B, but he regards her as being, so
far as he is concerned, countersuggestible. So he 48
says to her, "Now, dear, keep me company for a little."
If it is correct to say that Mr. B, who clearly did not
intend to have his wife keep him company, "meant by his
remark that she was to keep him company" then redefinition
IV is inadequate. To have meant4 that she was to keep
him company he must have intended that she intend to keep
him company, an intention which he himself never had. We can
correct this small defect by altering clause (i9) of (RIV,
VA) to read:

that A, on the basis of the fulfillment of (i7 ), to
think _U to intend him to W that p.

A similar change is needed in (16) of (RIV, VB).

Grice's analysis is still too strong. As indicated
above, there are cases when a speaker may not be trying to
get his audience to believe that he has, or come to have
themselves, certain propositional attitudes. When U greets
A by saying "Hello" he has not exhibited that he has some
propositional attitude toward some proposition, nor need

he have intended to produce some propositional attitude in A.

180

Thus, U must not have meant4 anything when he said
"Hello."

Whether or not one grants that U meant something

1
by sayingl "Hello" is a matter of little relevance at this
point. What is crucial is whether or not one should decide
this matter by looking at, as Grice's analysis would sug-
gest, U's intention or lack of intention to exhibit or
elicit attitudes toward propositions. It is a mistake to
regard all utterances spoken without the intention to
exhibit or elicit some attitude toward some proposition as

having been uttered without their having been meant But

1‘
this is what accepting the current Gricean analysis entails.
Redefinition V

How does Grice respond to Problem L? Grice argues
that U can still mean4 something by uttering E even if, in
certain cases, there is no A present. There are three
kinds of cases to consider. There are those utterances
which may now or later be addressed to some audience like
the writing in a diary or the posting of signs. I believe
that this group needs some consideration. Then, there are
utterances which the speaker "pretends" to address to an
audience or imagines to be addressing to an audience like
rehearsing a part in a play or in an imagined conversation.
It seems to me that Grice might just as well say that in
these cases U did not mean anything. In these cases the

speaker knows that no actual audience is present, or at

181

least he believes that none is present for the moment.
Grice's final classification includes those "internal"
utterances for which the speaker neither thinks that there
will possibly be an actual audience, nor imagines himself
to be addressing an actual audience. I believe that Grice
has neglected still another group of utterances, viz.,
those which do not involve intending to communicate to any
audience, but nevertheless involve meaningl what one saysl.
This class would include making a pledge or bestowing a

name.

Grice approaches Problem L by requiring a less pre-
cise notion on U's part of his prospective audience. The
speaker need only intend that his utterance would induce a
particular kind of response in a particular kind of audi—
ence, should the audience be present.49 U's conception of
this audience might be fairly indefinite, but he should,
according to Grice, be able to mention some characteristic
by which to identify his intended audience.

Grice presents his final redefinition. (Wisely he
disregards the Version A form.) By (RV, VB) Grice hopes to
solve Problem L by accounting for "the examples that need
to be accounted for."50 Let '¢' range over properties of
persons. The properties are to be selected from a more
restricted range in accord with examples like "being a
passer—by" or "being identical with Jones" or "being a native

speaker of English." Now we can present (RV, VB). But we

182

must allow ourselves more than the usual amount of liberty
with the text. Not only does Grice neglect to list U's
intention to induce the M—intended effect, but he also
includes improbable requirements. For example, in Grice's
original (RV, VB) U must intend that E have the character—

istic of being an utterance—token such that everyone who

has the property ¢ would believe certain things about it.51
Moreover, he neglects what he used counter—example 11
to establish about the nature of clause (i6). Grice's impre-

cise (RV, VB) seems to avoid requiring the existence of an
audience. I believe that my reconstruction of it does
avoid this, but I notice that it does not settle anything
with regards to the nature of A. Further, notice that the
logical structure of definiens of (RV, VB) is that of a
universally quantified conditional which has been existen-
tially generalized. The conditional has a complex conjunc-
tion as its consequence. Notice also that (RV, VB) should
probably be divided into two definitions, one for meaning4
something purely exhibitively, and one for meaning4 it pro—

treptically as well.52

(RV, VB) U meant by uttering x that *Wp, if and only if

there is some ¢, some f, some c, and no E, such
that U uttered_x intending that: for every A, if
A has ¢, then _
(i ) A believe that U W' s that p
and (i ) A think that E has f
and (i3) A think that c correlates f with the
response of believing that U Y' s that p
) A think that there is some ¢' such that:
U intends that for all E, if E has ¢'

and (i4

183

then, E think, via thinking that E has
E and that g correlates E with the
response of believing that U Y's that E,
that U Y's that p

and (15) A thifik, at least in part by reason of
his fulfillment of (i4), that U T's that

E
[and, for some cases, (i ) A think, on the basis
of the fulfillment of (15), that U intends

that he W that p]
and (i7) A rely on E in coming to think that U

Y's that pt or that U intends that A W
that p —

and (i8) A think that there is some ¢' such that:

U intends that for all B, if B has ¢'

then, E think that U intends that E not
rely on E in coming to think that U W's
that p [or that U intends that E T that *

2.]
Grice intends that the bracketed clauses in (i7) and (i8)
replace (without the word 'or') the phrase 'that U W's that
2' whenever clause (16) is operative. Clause (16) is
operative if and only if the utterance is protreptic. I
believe that it would improve Grice's definition if we
allowed that the bracketed clauses in (i7) and (i8) operate
in disjunction with 'that U W's that p' when (i6) is opera-
tive. This prevents U from intending to deceive A with
regards to either the exhibitive or the protreptic aspects
of his utterance.

(RV, VB) avoids part of Problem L for from the truth
of the definiendum one cannot infer that some audience
exists. One might infer that U conceived of his intended
audience as all those who had ¢. If ¢ is a property like

"being an English speaking person living in East Lansing in

184

the year 2171" the audience need not even be conceived of

as currently existing.

Problems

The weight of evidence is, I believe, against Grice.
The discussions above indicate that Grice was unable to
defend his analysis from the charge that it was too strong.
To some extent he did treat of Problem K and Problem L,
but he produced counter-example 10 for himself. The unre—
solved counter—example 2 indicated that his analysis is 1.
also too weak. His work suffers in several other respects
as well. On the whole it is confusing, at times careless.
His use of formal devices gives a false sense of precision.
His style lures the reader into several small concessions.
His analysis of examples is often odd, if not counter—
intuitive, and his method of amending definitions seems, at
times, gg Egg. If the Meaning-intention seemed improbable
on the basis of the subjective evidence of introspection
when we considered the (1957) analysis, consider how much
more improbable (RV, VB) is.

One philosopher who tries to show that Grice's work
is not only faulty in detail but also in principle chooses
to base some of his arguments on Problem F, this problem of
introspective evidence. N. L. Wilson hopes to show, by
presenting his "ultimate" counter-example, that Grice's
analysis is too strong. Wilson's counter—example is not

like earlier ones for it does not attack a particular clause

185

or aspect of Grice's definition. This counter-example
attacks the enterprise of defining meaning in terms of
complex intentions. Here is an adaptation of Wilson's
counter—example:

Counter-example 12 (Wilson): Suppose I am conversing
with Grice. I say 'Snow is white.‘ By uttering 'Snow
is white' I mean that snow is white. According to
Grice it follows that I intended a number of things,
namely all the things mentioned in his definiens. Now
I do intend to say that snow is white, but the only
other intention I have is to avoid having any of the
intentions Grice attributes to me in this or any sub—
sequent analysis. I shall, with every revision he 2?
offers, avoid trying to intend what his analysis claims

I intend.

 

Wilson's counter-example trades on the fact that a
person can believe that p and fail to believe that g even

though 2 entails g. Wilson is arguing that U can meanl

something and not intend certain things, even though mean-
ingl something entails, 323E Grice, intending these things.
The standard method for discovering what U in fact intends
is simply to ask him whether or not he intends such-and-such.
For example, Grice could ask whether U intended to have him
rely on some E which U believes is false. Wilson completes
the scenario of his discussion with Grice anticipating this
kind of question:

Grice: So. By uttering 'Snow is white' you meant
something. Did you intend to produce this
response in me?

Wilson: (correctly—-that's essential): No sir!

Grice: That response?

Wilson: No. There is no response such that I intended
to elicit that response from you. But you
can't claim that I wasn't really talking to
you. Of course I was. If you hadn't heard me
I would have repeated myself in a louder voice.

 

 

 

186

As a matter of fact I didn't even intend that
you should abandon your views. If I had, I
might, of course, have played into your hands.
In any case--and this is more to the point--
you can only intend to produce an effect if
you deem it possible to do so (as you have
pointed out) and everybody knows it is not
possible to persuade a philosopher to give up
a cherished doctrine. To repeat, I wasn't
trying to produce any particular response in
you at all. My only intention was to behave
in a counter-exemplary way.

 

Wilson's counter—example stands or falls with his

"dogmatic" claim that intuitively we can see no inconsis-

 

tency in the final act of his scenario.53 However, all

that Wilson has shown is that there seems to be 22 intuitive

 

inconsistency in claiming that U_meantl something and deny-
ing that he had all the intentions that Grice attributes to
him. Wilson has not shown that it lg inconsistent to claim

that U meant something and to deny that he had all the

l
intentions that Grice attributes to him. As far as that
goes, Wilson has shown nothing.

In alluding to Problem F we are not, however, posing

 

a serious threat to those theories which seek to analyze
meaningl in terms of complex intentions. One can argue that
in these matters our intuitions are in need of reform. How-
ever, there may be some point to comparing our intuitions
when it comes to evaluating the adequacy of a particular
analysis. As was the case when counter—example 5 and
counter—example 2 were discussed, our intuitions may tell

us that a situation has been poorly understood, or a concept

has been inadequately expressed. Wilson is right to try to

 

 

 

187

point out that the final analysis Grice offers is not one
which captures the intuitive sense of meaningl. Part of
the reason for this dissimilarity between Grice's meaning4
and our intuitive notion of meaningl is suggested in EEppf
_1_e_m_<_>_.

Problem 0, like the others in group five, is

 

generated, at least in part, because of the illocutionary/
perlocutionary distinction. Unless Grice can achieve the
reduction of illocutionary intentions to perlocutionary

intentions, Problem 0 will stand as a serious objection to

 

his analysis of meaningl. His analysis seems to be better
suited to the definiendum 'By uttering E U tried to accom-
plish something' than to either 'U uttered E and meant what
he saidl' or 'U uttered E and meant something'. His
analysis seems to be an analysis of trying to candidly
accomplish some perlocutionary act by means of uttering E,
and not an analysis of meaning1 something.

Problem M urges that the reduction cannot be

 

achieved. One reason why the reduction will fail is that,
given the different EEEEE of responses or effects U may

intend, it appears that illocutionary intentions are simply
not perlocutionary intentions. In general there seem to be
three kinds of responses or effects U may intend. U may

intend (a) to secure uptake, that is to have A understand
what U_is doing in uttering E. U may intend (b) to secure

some perlocutionary effect. He can "try to get A to" react

188

in a certain way. The intended reaction may be that A
think that U intends that p. U does not have to be candid
with A about this: U may intend that A not realize what
U's intended perlocutionary effect is. Finally, U may
intend (c) that his utterance of E have some illocutionary
force. U may wish to tell A that p, for example. Grice
tries to assimilate all three kinds of intended results to
the type (b) response. Note that in the case of the type
(c) result there is an associated type (a) response in
cases when U is trying to communicate with A, for in such
illocutions uptake must also be intended.

Judicious paraphrase might indicate how to capture
type (a) responses in type (b) terminology. U can intend
to "get A to think that U intends to" do such-and-such.

But such a paraphrase is inadequate. If one intends that

A understand, then one cannot intend to deceive A; however,
if one merely intends to get A to think something, then one
may also intend to deceive A. Thus type (a) responses are
not type (b) responses. But perhaps we would grant that a
judicious paraphrase used in conjunction with some kind of
anti-deception clause would allow the reduction of type (a)

to type (b). Can type (c) results be assimilated?

In uttering E, U may have intended to tell A that p,

or perhaps simply to greet A. In this case U need not
restrict himself to intending to achieve some entirely cog-

nitive response in A. Perhaps the only cognitive response

 

 

 

189

U intends is uptake. All type (b) responses are, in the
Gricean analysis, cognitive. They follow the pattern: U
intends to "get A to think that . . . ." But this would
make the reduction impossible for intending to perform an
illocutionary act is not in every case to intend to induce
an entirely cognitive effect in A. Moreover, and apart
for Grice's analysis, intending to perform an illocutionary
act need not entail intending to achieve any response in A
beyond uptake. However intending to accomplish a perlocu—
tionary act entails intending to achieve some effect in, or
response by, A, but it does not entail intending uptake.
The reduction, then, cannot be accomplished.

Counter—examples generated in accord with Problem P
would not apply to an analysis of meaningl that recognizes
the perlocutionary/illocutionary distinction. Cases of U's
intending to produce perlocutionary effects in A by uttering
E can be ruled out; only intentions to achieve the illocu—
tionary results of uttering E would be considered in an
analysis of meaningl.

Other distinctions would help with Problem P also.
We can distinguish between three general kinds of speech
acts: (A) the acts of simply uttering words or sayingl
something. (B) the acts by which U intends to accomplish
something as the perlocutionary effect of uttering something.
Perlocutionary acts may or may not be candid. That is, U

can intend to have A be ignorant of U's actual intentions,

 

190

intending to appear to A_to be revealing other intentions
by which U hopes to mislead A_or deceive A, (C) the acts
of doing something in sayingl something, the illocutionary

acts. In the case of illocutionary acts the notion of a

 

non-candid act is inappropriate. To perform an illocutionary
act U must, if communication is involved, intend that A
understand what U is doing. Thus, U intends that A know
what U's actual intentions are. By virtue of U's intending
uptake on A's part illocutionary acts that involve communi-
cation are, of necessity, candid. Thus, anti-deception
clauses, and anti—ignorance clauses, are superfluous in
such cases.

Grice's analysis of meaning4 requires that U intend

some candid perlocutionary effect. In Problem N this was

 

noted. Grice's analysis is not an adequate analysis of
meaning1 for one need not intend an effect of kind (b) to
perform an act of kind (C).

One might suggest further support to the illocu-
tionary/perlocutionary distinction by noting that these two
kinds of intentions can vary independently in a given
situation. N. L. Wilson argues that Grice's theory is sub-
ject to many counter-examples because his position entails
the View that these two kinds of intentions do covary.53
Although Grice has never said this, it follows from his

reductionist thesis that for each illocutionary intention U

might have there is some set of perlocutionary intentions

191

which we can attribute to U such that the illocutionary
intention is nothing more than the set of these perlocu—
tionary intentions.55 Here we might expect U‘s illocu-
tionary intention and the set of perlocutionary intentions
to vary directly.
Wilson offers the following example to indicate
that a reduction cannot take place because of the indepen-
dence of the two kinds of intentions.56
Counter—example 13 (Wilson): If I wish my guests to
leave, there are any number of different things I might
say (and mean by what I say) in order to shoo them out.
On the other hand, if I say, "It's getting pretty late,"
meaning that it's getting pretty late, there might be
any one of a number of different things I expect of my
audience.
"There just is not" says Wilson, begging the question, "the
kind of covarience necessary to make Grice's theory go."57
What Wilson can argue is that the covarience seems counter—
intuitive on the basis of examples like 13.
Still another avenue of attack is open to support
the claim of Problem M and urge that the reduction Grice's
thesis requires must fail. Strawson has argued that some
illocutionary acts are so circumscribed by conventions that
one cannot perform them, if one is in the appropriate circum—

58

stances, even without the intention to perform them. The

 

example of making a bid in a game of bridge has already been
cited. All the player need do is saya or utter one of the
appropriate words or phrases at the appropriate time. One

need not have intended anything, much less M-intended to

 

 

192

make a certain bid or to bid at all, for the bid to have
been made officially. Likewise, one can think of some
illocutionary acts which cannot be performed, no matter

what one intends to do, unless one is in the appropriate
circumstances. A bid, for example, cannot be made unless
the bridge game is in the auction and not in the play
stages, and unless it is U's turn to bid, and unless U
selects the word or phrase that he will sayl from among a
small group of appropriate expressions. That U should utter
'one heart' intending to bid one heart at any other time is
not a sufficient condition for his having performed the
illocutionary act of bidding one heart; no matter how firm
his intentions in the matter were. In such a case we might
what he said

say that U meant but that his illocutionary

l 1’
act misfired. There is something in these highly institu-
tionalized illocutions that approaches ritual. Their suc—
cessful performance is removed from the control of solely
the "meaning intentions" of the agents involved. Thus, not
only is Grice's analysis of meaningl inadequate, but by
Virtue of restricting itself to listing only the intentions
of the speaker it is irreparably inadequate.

Furthermore, Grice's theory of meaning is irrepa-
rably incomplete. A theory of meaningl must account for U's
illocutionary intentions. If in uttering E U told A that p,
or promised A that p, then U meant what he said . If in

l 1

uttering E U warned or requested or advised A, then he

 

193

meantl what he saidl. A sentence of the form "In uttering

E, U ...—ed A" entails a sentence of the form "U meantl E.
A sentence of the form "U meantl E" entails a disjunction
of sentences of the form "In uttering E U ...-ed A." In

each case the lacuna is to be filled by the specification
of an illocutionary act. Grice's analysis of meaning4 is
best suited to expressing U's candid perlocutionary inten-
tions. Since illocutionary intentions cannot be reduced
to perlocutionary intentions, Grice's analysis is not an
adequate expression of illocutionary intentions. Nor can
it become so. Thus Grice's analysis is not an adequate
analysis of meaningl.

The problems in group five (Problems E, U, U, and E)
primarily attack Grice's theory of meaning, in that it over—
looks the illocutionary/perlocutionary distinction. They
do not show conclusively that any theory of meaningl which
restricts itself to listing only U's intentions in the
definiens is going to be inadequate. However, the discus—
sion of highly ritualistic illocutions suggests that this
is the case. In the case of some illocutionary acts, U's
intentions seem to be neither necessary nor sufficient
condition for the successful performance of the act. Often
the acts depend on using certain expressions, often on
using these expressions at certain times or in certain con—
texts, or with certain prerequisite conditions having been

satisfied. But, if U believes that he has satisfied all

 

194

the prerequisites, that he is speaking in the right context
using the right expression, and if he intends to perform an

illocutionary act, then he will have meant what he saidl.

1
His illocutionary act may misfire because his beliefs might

have been mistaken, yet he will have meant something in

l
uttering E. An adequate theory of meaningl will capture
the role played by U's beliefs concerning the choice of his
utterance token and the circumstances of his utterance of
that token. A theory of meaningl that lists only U's

intentions cannot capture these elements.

Problem Q first suggested the necessity of U's

 

selection of a token which he believes to have some feature

by virtue of which A will be able to understand what U_is

 

doing in uttering E, whenever U is trying to communicate
with A. This problem, like the analogous ones just men-
tioned, suggests that the narrow restriction to listing only
U's intentions is unfortunate and should be disregarded in
developing a theory of meaningl.

U's choice of a token will, in most cases, be deter-
mined by the meaning2 of the expressions in U's language.
Meaning2 is determined by the conventions of the language.
These conventions may not exist independently of any inten-
tions ever had by men at any time, but they are independent
of a particular author acting at a particular time. To

accomplish one's linguistic purposes one does not create

195

these conventions at the moment of speaking, but one speaks,
at that moment, in accord with them. Searle says,

One can in certain special circumstances 'request' some-

one to leave the room without employing any conventions,

but unless one has a language, one cannot request of

someone that he, e.g., undertake a research project on

the problem of diagnosing and treating mononucleosis

in undergraduates in American universities.59

Are these conventions reducible to intentions at
any point? In the above quotation Searle is not referring
to those conventions that regulate behavior that exists
antecedently and independently of those conventions, like:
the rules of etiquette, or the ethical code of a profes-
sion, or the rules of thumb that a cook or gardener might
follow. Searle is talking about those rules that consti-
tute the very activity in question, like the rules that
taken together define playing American football.60 Regula-
tive rules may be codified custom, or the result of delib—
erate and intentional formulation. Constitutive rules seem
to be intentionally formulated in every case; as for example
when one sets about creating a new game or a new language.
However, the intentions that one has to constitute a language
are not the intentions that one has if one is trying to speak
in accord with a language or to use a language.
Breaking a regulative rule is to do something wrong;

but when one breaks a constitutive rule one has failed to do
what one intended to do. To violate the rules of football

or the constitutive conventions of a language is to void

one's act; the act has, to use Austin's expression, misfired.

196

To violate these rules is, if done deliberately, to consti-
tute a new game or new language, or modification of the
existing situation.61 One can, if one chooses, set out to
play a football-like game, just as one can, if one chooses,
speak in ways that violate existing linguistic conventions.
But one cannot do this in a practical way unless one
announces these intentions beforehand. And one cannot be
said to be playing football or speaking the original

language but playing something like football or speaking

 

something like the original language.

 

It is possible to violate rules simply out of
ignorance. A child learning to speak does not abide by all
the rules, nor does it always succeed in communicating what
it intends to communicate. The process of learning what
the conventions are and how to follow them suggests that
the question of the relation of conventions to intentions
may be resolved if we can determine how language originally
came to be instituted and constituted among men. But this
historical question is irrelevant. However a language came
to be constituted, it is no longer within the range of the
individual's power to drastically alter a given language by
his intentions or purposes on specific occasions in the
normal course of his linguistic activity. He may wish to
alter the language, or create a new language, but these
intentions are not the more parochial purposes and inten-

tions that Grice and Leonard spoke about and which are part

197

of one's everyday linguistic activity. Thus it seems that
an adequate theory of meaningl does not have to deal with
the reduction of conventions to intentions; rather it must
provide for U's intention to speak in accord with what he
believes to be the conventions of his linguistic community.
It does this by noting that whenever U speaks meaningfullyl
he does believe that his utterance has some feature by
virtue of which it will be understood by the person with
whom he is trying to communicate. In most cases this

feature is the conventional meaning of the utterance.

 

A Definition of Meaningl

If we follow the suggestions made in the section
above, we may be able to generate an analysis of meaningl.
We must not neglect the illocutionary/perlocutionary dis—
tinction. We must drop the restriction that we list only
U's intentions in the definiens. We must remember to pro-
vide for his belief that he is speaking in accord with the
conventions of his linguistic community (construed broadly),
i.e., in a way that he believes will be understood.
Recalling Problem 0 and several other discussions we must
remember to alter the definiendum. We should, however, avoid
making a person's meaningl something too directly connected

with the meaning2 of what he says This is a fault that

l'
Searle's definiendum, roughly: "U utters E meaning literally
what he says," possesses. For, it is possible that people

meanl something other than what their utterance means We

2.

 

 

198

must also take care to avoid Problem L, the problem of
requiring an audience. Moreover a sophisticated analysis
of meaningl something would be able to handle utterances in
which no propositional content was found, as well as those
in which a propositional content is utilized. Likewise the
analysis should provide for the possibility that U's
intended audience be thought of as reading the utterance

at a time which is either the same as or later than the
time at which U produced the utterance.

We might stop, having seen that an analysis of
meaning1 that restricts itself to listing intentions is
never adequate. But, it seems that we would be unwise to
neglect making use of the hints that the above discussions
have brought to light. I cannot claim that the following
analysis of meaningl is adequate, but I am confident that
it is a move in the right direction. It tries to incorpo-
rate the many virtues and avoid the many vices of several
of the earlier analyses.

Let me offer two examples to help with presenting
this new analysis of meaningl. The analysis will be adequate

to handle the meaningful utterances in

1
("The Greeting Example") At about 10:30 in the morning
John walks outside his house. He sees his neighbor and
greets him saying "Hello."

as well as in
("The Request Example") Later that same morning John

thinks of inviting his out of town in—laws for dinner
on a coming holiday. He sits down and writes in a

 

 

199
letter "Why don't you come for dinner on the holiday?--
We would love to have you visit."

The analysis of meaningl shall rely on the follow-
ing notational devices. Seven variables will be used. The
variable 'x' ranges over utterance tokens. In the examples
these are 'Hello' and 'Why don't you come for dinner on the
holiday?—-We would love to have you visit'. U ranges over
speakers; the person who produces the utterance token in
each example is John. The variables 't' and 't" range
over moments of time. In the greeting example the moment
of the production of E is the same as the moment when U
anticipates that his neighbor will hear what was said. In
the case of the request John anticipates that his in-laws
will read the letter at some time later than the moment he
is writing it. The device '*' ranges over illocutionary
forces. When it appears it is to be replaced by some
specification of the illocutionary force of U's utterance
of E. In the first case John greets A and in the second
John requests ... of A. The '...' and 'A' are schematic
devices (not variables). 'A' is to be replaced by a speci-
fication of U's conception of the audience he intends to
communicate with. In treating these examples we would
replace 'A' with 'his neighbor' and by 'his out of town in—
laws'. The schematic device '...' is to be replaced by a
specification of the propositional content of E where appro-
priate. In the first example there is none, in the second

the lacuna is replaced by "that you should come to dinner on

 

200

the holiday." The variable 'f' ranges over features of

the utterance token E. The variable 'c' ranges over the
ways in which these features are correlated with illocu-
tionary acts. Generally the utterance tokens are correlated
with illocutionary acts by virtue of their meaningz.

In selecting the appropriate definiendum it is
important to choose something which is the middle ground
between the convention-bound idiom "U utters E and means
it" and the under—developed (G12). We might choose Grice's
definiendum, but this does not seem suited to anything but
perlocutionary intentions. Moreover, we wish to use a
definiendum which allows us to mark the fact that in com-
munication the speaker has some concept (perhaps quite
vague) of who his audience is. Attempted communication
involves, at least, the belief that there is some possible
audience for the author to speak to. I trust that some-
thing like the definiendum I have selected will prove to be
adequate for the expression of meaningl. It will be obvious
that the phrasing I have selected yields a sentence of the
form "U uttered Emeaningl to do such-and-such." Grice
urged that the sense of 'meaning' in 321 such sentence is
not the relevant sense of 'mean', that is, not meaningnn.
However, it seems that if there is any truth at all to
Grice's generalization it is that if one replaces the 'such-
and~such' by a name of a perlocutionary effect, then the

relevant sense of 'mean' is absent. However, it seems that

201

if one replaces that phrase by a word naming an illocu-
tionary force it is possible that the appropriate sense of
'mean' is present. Moreover, we cannot be sure how much
faith Grice himself had in his generalization for it seems
that whenever one gives the Gricean specification of what
U meansnn it turns out to be that U means to do such-and—

such. That is, U means to induce in A the belief that etc.

(Def. Ml) U utters x at time t meaning to * A ...

 

if and only if there is some E, g, and El (later

or the same as t) such that:

U utters E at ttme E

(l) believing that A would think, at El, that E
has E_which g correlates to *-ing ...

(2) intending to * §.--- in uttering E

(3) intending that A think, at El, by virtue of
believing that E has E which E correlates to
*-ing ..., that U intended to * A ... in
uttering E at E _

Clause (1) of (Def. Ml) takes care of Problem Q.

 

In the Little Girl Example we saw that U could mean some-

1
thing even if he did not believe that the utterance-token
he used meant2 what he meantl. It is sufficient that U
believe that his intended audience will take the utterance-
token to have a certain meaningz. Clause (1) is not too
strong, although it may seem so. One might consider the
case of Mr. Jones who has an English-speaking foreign friend.
One day the friend became so excited about something that he
began to tell Jones about it in his native language, but

Jones did not speak that language. We should not want to

say that the friend did not mean anything, nor that he would

   

202

claim that he thought that Jones did know the foreign
language. However I believe that we can say that implicit
in the friend's behavior is the belief on his part that
Jones does understand. As soon as Jones indicates that he
does not the friend will remember this and speak English.
Clause (2) presents U's illocutionary intention, clause (3)
his intention to secure uptake. As was argued above, with
illocutions no anti-deception nor anti-ignorance legislation

is required. We avoid Problem L by the use of schematic

 

letters and by altering the definiendum make it apply to
cases of attempted communication.62 Some illocutionary
acts, however, do not fit this pattern. For example, if
one is naming a child or performing some kind of pledge in
which a certain formula might have to be spoken. Also,
someone may wish a concept of meaningl which better suits

the notion of meaningl literally what one says We might

l’
adopt a definiendum like that used by Searle in the second

chapter and offer the following auxiliary definition:

(Def. Ml)+ U utters x at time t meaningl what he saysl
if and only if there is some E, and U such that:
U utters E at time
(1) believing that
to *—ing ...
(2) intending to * ... in uttering E

t
E has E which c correlates

I might suggest one final definition. We might, in the
spirit of the theory of meaning as intention, define "E

means2 * ..." as "generally the use of E is to * someone ..."

203

or, "we conventionally utter a token of E when we wish to
* someone ... ."

I believe that those problems of more immediate
importance, namely those in groups four, five, and six in
the summary of Chapter II, can be resolved favorably for
these definitions of meaningl. Moreover, these definitions
have a simplicity to them that goes far toward settling

Problem F.

 

Conclusion

 

Having surveyed the efforts of Grice (and some of
his supporters) to provide an adequate analysis of meaningl
by listing only the intentions of the speaker we concluded
that such an analysis must fail. In Grice's case it fails
because of his neglect of the illocutionary/perlocutionary
distinction and the several problems this neglect fosters.
In the case of a theory restricted to listing only inten-
tions we found that fatal neglect of the conventional
elements in communication.

Throughout the discussion of Grice's efforts to
analyze meaningl fruitful insights were provided as to the
requirements of an adequate analysis. (Def. M1) is a first
step toward such an analysis. It is in the spirit of the
theory of meaningl as intention, but it does not neglect the
illocutionary/perlocutionary distinction, nor does it
restrict its analysans to listing only intentions. I

believe that the analysis avoids the immediate problems

 

 

204

that plagued Grice's work and the theory of meaning as

intention in general.

 

 

 

 

 

FOOTNOTES TO CHAPTER III

1H. P. Grice, "Utterer's Meaning, Sentence-Meaning,
and Word-Meaning," Foundations of Language, IV (1968),
225-42, referred to in these notes as 1TUMSMWM."

2H. P. Grice, "Utterer's Meaning and Intentions,"
Philosophical Review, LXXVIII (1969), 147-77, hereafter
ll UM]: . ll

 

 

3N. L. Wilson, "Grice on Meaning: The Ultimate

Counter—example," Nous, IV (1970), 295-302.

4Mr. Dennis Bird informed me that he has heard radio
lectures Grice has given on the topic of speech acts at the
University of California. "UMI" was originally a public
lecture delivered in 1968 at Oberlin College. Professor
Frye of the University of Pittsburgh indicated that these
tOpics often occurred in conversations and seminars Grice
conducted.

5"UMI," pp. plE., p. l74n. "UMSMWM," pp, plE,,
p. 225n. (Rumor has it, Via Professor Frye, that the manu-
script was accidentally destroyed.)

6"UMSMWM," pp, cit., p. 225.

7Ibid., p. 225f.

8The discussion of the four-fold distinction is
drawn from "UMI," pp. cit., pp. 147-50.

91bid., p. 149.

lo"UMSMWM," pp, cit., p. 226.
ll"UMI," pp. cit., p. 151.
12

"UMSMWM," pp. 913., pp. 226, 230f. "UMI," pp. 913.,
pp. 166, 171. The assumption that each grammatical device

is correlated with one and only one mood is false, just as

the assumption that each mood is to be correlated with one
primary function. For example, the '1' is generally asso-
ciated with the imperative mood for it often is used to
indicate commands or entreaties. But it can be used in

other contexts as well. One finds these examples on page 1150

205

206

of Webster's New Collegiate Dictionary (Springfield, Mass.:
Merriam Co. Publishers, 1961 edition): (a) Oh that those
lips had speechl, (b) Is the writer to become the slave to
the publisher!, and (c) Egad! I had no such intention.

To exemplify Grice's association of intending with the
imperative we might claim that when I say "Open the door!"
I intend to have g intend to open the door.

l3"UMI," pp, cit., p. 171. "UMSMWM," 93, Cit-r P- 226-

14John R. Searle, S eech Acts: An Essay in the
Philosophy of Language (Cambridge: Cambridge University
Press, 1969?, P. 31.

15

 

 

 

"UMSMWM," pp. cit., p. 231.

lGIbid., p. 230.

17John R. Searle, pp. cit., p. 30.

18In this definition, and in those that follow, I
have taken liberties with the Grice text to try to insure
uniformity of format and consistency of presentation. For
example, at times Grice uses 'produce' and at other times
'induce' in the first clause. Since 'induce' seems better
suited to the kinds of effects intended, I use 'induce' in
all the definitions. At other times Grice omits clauses
that seem to be required, or he submerges clauses inside of
other clauses. I try to make all the necessary clauses
explicit. Further, I have simplified constructions like
'g intended that p and g intended that g' to 'g intended
that p and g'. I trust that nothing important is lost in
this economy.

l9"UMI," pp. cit., p. 152.

201bid., p. 153.

211bid., p. 154.

22N. L. Wilson, pp. cit., p. 299.

23"UMI," pp. cit., p. 155.

24Ibid., p. 156.

251bid., p. 157.

261bid., p. 158f.

27Ibid., p. 158.

207

281bid., p. 159.

29John R. Searle, "What is a Speech Act?" Philosophy
in America (1965), 229f.

30Note the use of indirect discourse, rather than
quotation marks to specify the purported meaning2 of the
sentence.

31

 

 

Chapter II, supra, in "Group Six."

32Cf. Morris Weitz, "The Role of Theory in Aesthet-
ics," Journal of Aesthetics and Art Criticism, XV (1956),
27-35, as reprinted in Perspectives in Philosophy, ed. by
Robert N. Beck (New York: Holt Rinehart and Winston, Inc.,
second edition 1969), pp. 386-93. There is a striking simi-
larity between what the theoretician of art is doing and
what I believe Grice may be doing if his theory is pushed
to its limit. The theoretician reaches a stage at which he
is no longer discerning a characteristic of objects but
making prOposals as to how the concept of art (meaning)
should hereafter be circumscribed.

33Cf. John Wisdom, "Gods," published in Essays on
Logic and Language, ed. by Anthony Flew (Oxford: Basil
Blackwell and Mott Ltd., 1951), pp. 187-206. Note section
6.5 where Wisdom discusses the problem that one faces when
both sides of a dispute are using reasons that are less than
totally convincing. Wisdom likens the situation to a case
in a court where a decision must be made on the basis of
the evidence at hand.

34

 

 

 

 

"UMI," pp. cit., p. 162.

3SIbid., p. l62f.

361bid., p. 164.

37Ibid., p. 163.

38(17) expands to '(Ex)(z)((Ey)FzyEx=z)' and (18)
to '(Ey)(Ex)(z)(FzyEx=z)'. From the former one can deduce
the latter, but one cannot deduce the former from the latter.

39"UMI," pp. cit., p. 165.

4OIbid., "The identification of what 2 meant by E
would turn on the identification of the M—intended response
or effect." p. 165. In the cases of indicative and impera-
tive utterances "the M—intended response will be a proposi-
tional attitude." p. 166. But, "The way is opened to a

 

 

 

208

simplified treatment of the M-intended effect as being
always the generation of some propositional attitude."
"UMSMWM," pp. cit., p. 230.

41"UMI," pp, cit., p. 166.

421bid., p. 169f.

43Ibid., p. 171.

44Ibid., p. 167.

45

Ibid., 172.

46

47
(RIV, VB).
48

P
Ibid., p. 172f.
P

Ibid., . 173. Here Grice gives directions for

Ibid., p. 173.

49Ibid., p. 174.

50According to Grice one need not account for some
cases of verbal thinking in which the speaker is not
actively framing the thought but is more of a passive
listener to his own mental speaking. Ibid., p. 175.

511bid., p. 175.

52Ibid. Grice says "I could present a more formal
version which would gain in precision at the cost of ease
of comprehension." (I think he is mistaken on both counts.)
He offers the following as (RV, VB):

Redefinipion V
"U meant by uttering E that *Wp" is true iff

 

(E¢)(Ef)(EC)=

uttered x intending x to be such that anyone who has
would think that

 

9
¢

(1) g has E

(2) f is correlated in way p with w-ing that p

(3) TE¢'): g intends p to be such that anyone who has
¢' would think, via thinking (1) and (2), that g
W's that p

(4) in View of (3), g W's that p;

209

and
II. (operative only for certain substituends for "*Wp")
p uttered x intending that, should there actuallypbe

anyone who—has ¢, he would via thinking (4), himself
W that p;

 

and
III. It is not the case that, for some inference-element E,
g intends égto be such that anyone who has ¢ will both
(1') rely on B in coming to 9+ that
and (2') think that (E¢'): U intends x to be such that
anyone who has ¢' will come to 9+ that p without
relying on E.

Notes: (1) "W+" is to be read as "W" if Clause II is opera-
tive, and as "think that U W's" if Clause II is
non-operative. —

(2) We need to use both "¢" and "¢'", since we do
not wish to require that g should intend his
possible audience to think of E's possible
audience under the same description as U does
himself. —

53N. L. Wilson, pp. cit., p. 296.

54Ibid., p. 296f.

55Grice did suppose a covariance between the kinds
of results g intends and what E_might have meant by uttering
§.40 This generalizes his thoughts on the supposed covari-
ance of grammatical moods and kinds of M-intended effects.
Grice's supposition first appeared in "Meaning." Wilson
takes Grice's supposition out of context, and he does not
present Grice's own revised interpretation of his remarks
as this appeared in "UMI" on page 166. Further Wilson mis-
construes the force of Grice's supposition claiming that it
asserts a covariance between illocutionary and perlocution-

ary intentions. Ibid., p. 296. Wilson's construal is
unfair, but interesting.
56

Ibid., p. 297f.

57Ibid., p. 298.

5 8 ' ll ' I!
Chapter II, supra, in Group Five.

59

John R. Searle, Speech Acts: 'An'Essay'in'the
Philosophy of Language, op. cit., p. 38f. Searle argues fur—
ther that “some system of rule governed elements is necessary
for there to be certain types of speech acts, such as prom—
ising or asserting." p. 38.

 

 

 

210

60£§£Q-, pp. 33-42. Here Searle develops the dis—
tinction between constitutive and regulative rules.

61Although this is off the point, it seems that
much of the modern confusion in matters of morals and ethics
is perhaps due to the tendency to treat ethical codes as
constitutive rather than regulative. Novel behavior becomes
a "new morality" rather than a violation of morality. This
raises a question: should ethical codes be thought of as
constitutive of "ethical" behavior or regulative of human
behavior?

62To make the rendering of '* A ...' and '* ..."
into English more idiomatic it would be necessary to provide
for each verb that replaces '*' a note on its grammar. One
might, for example, note that it is proper to use a 'that'-
clause with the verb 'state' and that it is proper to write
"state pp_A that such-and-so" rather than "state A that
such-and-so."

 

 

BIBLIOGRAPHY

 

BIBLIOGRAPHY

Books

Alston, William P. Philosophy of Language. Englewood
Cliffs, N.J.: Prentice—Hall Publishers, 1964.

 

Austin, J. L. How to Do Things with Words. New York:
Oxford University Press, 1965.

 

Ayer, A. J., ed. Logical Positivism. New York: The Free
Press, 1959.

 

Beck, Robert N., ed. Perspectives in Philosophy. New York:
Holt, Rinehart, and Winston, Inc., second edition
1969.

 

Black, Max, ed. Philosophy in America. London: George
Allen & Unwin Ltd., 1965.

 

Chappell, V. C., ed. Ordinary Language. Englewood Cliffs,
N.J.: Prentice—Hall Publishers, 1964.

 

Flew, Anthony, ed. Essays on Logic and Language. Oxford:
Basil Blackwell and Mott Ltd., 1951.

 

Holloway, John. Language and Intelligence. London:
Macmillan, 1951.

 

Kerner, George C. The Revolution in Ethical Theory. New
York: Oxford University Press, 1966.

 

Lambert, Karel, ed. The Logical Way of Doing Things. New
Haven, Conn.: Yale Press, 1969. (Essays in honor
of Henry S. Leonard.)

 

Leonard, Henry S. Principles of Reasoning, New York:
Dover Publications, 1967.

 

. Principles of Right Reason. New York: Henry
Holt and Company, 1957.

 

 

Olshewsky, Thomas M., ed. Problems in the Philosophy of
Language. New York: Holt, Rinehart, and Winston,
Inc., 1969.

 

211

 

 

212

Parkinson, G. H. R., ed. The Theory of Meaning. London:
Oxford University Press, 1968.

 

Passmore, John. A Hundred Years of Philosophy, New York:
Basic Books, Inc., 1957, revised 1968.

 

Quine, Willard Van Orman. The Ways of Paradox. New York:
Random House, 1966.

 

. Word and Object. Cambridge: The M. I. T. Press,
1960.

 

Searle, John R. Speech Acts: An Essay in the Philosophy
of Language. Cambridge: Cambridge University
Press, 1969.

 

 

, ed. The Philosophy of Language. London: Oxford
University Press, 1971.

 

Strawson, P. F., ed. Philosophical Logic. London: Oxford
University Press, 1967.

 

Wittgenstein, Ludwig. Philosophical Investigations. New
York: Macmillan, 1953.

 

Articles

Alston, William P. "Linguistic Acts." American Philosoph—
ical Quarterly, I (1964), 138-46.

 

 

"Meaning." Encyclopaedia of Philosophy, V, 223-41.

 

Austin, J. L. "A Plea for Excuses." Proceedings of the
Aristotelian Society, XLII (1956-1957), the presi-
dential address, 1-30. Reprinted in Ordinary
Language, 41-63.

 

 

Belnap, Nuel D., Jr. "Questions: Their Presuppositions,
and How They Can Fail to Arise." The Logical Way
of Doing Things, 23-37.

 

 

Callaghan, William J. "Henry Leonard at Michigan State
University." The Logical Way of Doing Things,
295—312.

 

Cavell, Stanley. "Must We Mean What We Say?" Ordinary
Language, 75—112.

Cohen, L. Jonathan. "Do Illocutionary Forces Exist?"
Philosophical Quarterly, XIV (1964), 118—37.

 

 

 

 

213

Green, 0. H. "Intentions and Speech Acts." Analysis, XXIX

Grice, H. P. "Meaning." Philosophical Review, LXVI (1957),
377—88. Reprinted in Philosophical LOgic, 39—48,
and Problems in the Philosophy of Language, 251-58.

 

 

"Utterer's Meaning and Intentions." Philosqph-
ical Review, LXXVIII (1969), 147-77.

 

. "Utterer's Meaning, Sentence-Meaning, and Word-
Meaning." Foundations of Language, IV (1968),
225—41. Reprinted in The Philosophy of Language,
54-70.

 

 

 

Hamblin, C. L. "Questions." Encyclopaedia of Philosophy,
VII , 49-53.

 

 

. "Questions Aren't Statements." Philospphy of
Science, XXX (1963), 63-63.

 

Hart, H. L. A. "Signs and Words." Philos0phical Quarterly,
II (1952), 59-62.

 

Leonard, Henry S. "Authorship and Purpose." Philosophy of
Science, XXVI (1959), 277-94.

 

. "Authorship of Signs." Papers of the Michigan
Academy of Science, Arts, and Letters, XLV (1960),
329—40.

 

"Interrogatives, Imperatives, Truth, Falsity,
and Lies." Philosophy of Science, XXVI (1959),
172-86.

 

"Reply to Professor Wheatley." Philosophy of
Science, XXVIII (1961), 55-64.

 

Mates, Benson. "On the Verification of Statements about
Ordinary Language." Ordinary Language, 64—74.

 

Patton, T. E., and Stampe, D. W. "The Rudiments of Meaning:
On Ziff on Grice." Foundations of Languagp, V
(1969)! 2'16.

 

Quine, W. V. "Quantifiers and Propositional Attitudes."
The Wpys of Paradox, 183—94.

 

Searle, John R. "What is a Speech Act?" Philosophy in
America, 221-39.

 

214

Stabl, Gerald. "A Review of the Leonard-Wheatley-Leonard-
Hamblin Debate over the Attribution of Truth Values
to Interrogatives." Journal of Symbolic LOgic,
XXXI (1966), 666—68.

 

Stampe, Dennis W. "Toward a Grammar of Meaning." Philo-
sophical Review, LXXVII (1968), 137-74.

 

Strawson, P. F. "Intention and Convention in Speech Acts."
Philosophical Review, LXXIII (1964), 439-60.

Weitz, Morris. "The Role of Theory in Aesthetics." Journal
of Aesthetics and Art Criticism, XV (1956), 27-35.

Wheatley, J. M. 0. "Note on Professor Leonard's Analysis
of Interrogatives, Etc." Philosophy of Science,
XXVIII (1961), 52-55.

 

Wilson, N. L. "Grice on Meaning: The Ultimate Counter-
example." Nous, IV (1970), 295-303.

 

Wisdom, John. "Gods." Essays on Logic and Language,
187—206.
Ziff, Paul. "On H. P. Grice's Account of Meaning."

Analysis, XXVIII (1967), 1-8.

Other Sources

 

Alston, William P. letter dated July 17, 1970.
Black, Max. letter dated May 4, 1970.

Clarke, D. S. "Meaning, Force, and Rhetorical Effect."
Unpublished manuscript read before the American
Philosophical Association, December, 1970. Abstracted
in The Journal of Philosophy, LXVII (1970), 828-29.

Dissertation Abstracts. Ann Arbor, Michigan: University
Microfilms.

 

Encyclopaedia of Philosophy, Ed. by Paul Edwards. New York:
Macmillan and The Free Press. Eight volumes, 1967.

MacKay, Alfred Farnum. Speech Acts. Doctoral Dissertation,
University of North Carolina at Chapel Hill, 1967.
Abstract no. 4665—A, Dissertation Abstracts, XXVIII
(1968) 1 pt. 5"A

 

 

 

 

215

The Oxford English Dictionary, Ed. by James A. H. Murray,
Henry Bradley, W. A. Craigie, and C. T. Onions.
New York: The Oxford University Press, 1961.
Twelve volumes and supplement.

 

Thau, Stewart. Linguistic Acts and the Concept of Meaning.
Doctoral Dissertation, University of Michigan, Ann
Arbor, 1969. Abstract no. 2085-A, Dissertation
Abstracts, XXX (1969), pt. 2-A.

 

 

Websters New Collegiate Dictionary. Springfield, Mass.:
Merriam Co. Publishers, 1961 edition.

 

 

 

APPENDIX

APPENDIX

Two men are primarily responsible for presenting
the theory of meaning as intention or purpose. Each
developed his own version of the theory, and moreover, they

published their work nearly simultaneously and, from all

 

appearances, independently. The two versions are distinct
in formulation, although they seem to have originated from
the same basic spark of insight, that meaning can be
explained in terms of an author's purposes or intentions.
These two philosophers and originators are Henry S. Leonard
and H. Paul Grice.

Herbert Paul Grice was born in 1913 in Birmingham,
England. He studied at Oxford, where he became a Fellow
and Tutor in St. John's College at the start of the second
world war. He acquired the respect of his many students and
colleagues during his nearly thirty years at Oxford. Grice,
who was P. F. Strawson's tutor, was influenced by the work
of the Ordinary Language philosophers, but his work remains
original in style and argument. He was thought of as a
lively and critical teacher, but he has published relatively
little. He wrote three articles directly concerned with
meaning. These were mentioned in Chapter I. His articles

"Personal Identity" (1941), "Metaphysics" (1957), "The

216

217

Causal Theory of Perception" (1961), and "Some Remarks about
the Senses" (1962) reveal an interest in metaphysical and
epistemological questions. Yet he is most famous for "In
Defense of a Dogma," a reply to critics of the analytic/
synthetic distinction, which he published in collaboration
with Strawson in 1956. Grice taught at Cornell and then at
Brandeis. In 1966 he finally left Oxford permanently to
become a professor of Philosophy at the Berkeley campus of
the University of California, after being a William James
Lecturer at Harvard for one year.1

The following are H. P. Grice's publications:

"Personal Identity," Mind, L (1941), 330-50.

"In Defense of a Dogma," Philosophical Review, LXV (1956),
141-58.

"Meaning," Philosophical Review, LXVI (1957), 377—88.

"Metaphysics" with D. F. Pears and P. F. Strawson, Tpp
Nature of Metaphysics, ed. by D. F. Pears (London:
Macmillan Co. Ltd., 1957), 1-22.

"The Causal Theory of Perception," Proceedings of the
Aristotelian Society, Supp. Vol. XXV (1961), 121—52.

"Some Remarks about the Senses," Analytic Philosophy, ed. by
R. J. Butler (New York: Barnes & Noble, Inc.,
1962), 133-53.

"Utterer's Meaning, Sentence-Meaning, and Word-Meaning,"
Foundations of Language, IV (1968), 225—42.

"Utterer's Meaning and Intentions," Philosophical Review,
LXXVIII (1969), 147-77.

 

 

 

 

 

 

 

 

 

Henry S. Leonard was born in 1905 in West Newton,
Massachusetts. In 1931 he received his Ph.D. from Harvard
University where he taught until 1937. Leonard also taught
at Rochester University for a year; he took at post at Duke
University in 1937. Leonard's earliest articles indicate

the convergence of four forceful philosophical perspectives:

218

Logical Positivism, the speculative philosophy of Whitehead,
the pragmatic tone of Peirce, and the concern for logic,
meaning, and coherent conceptualization of C. I. Lewis. In
1936 Leonard published "Logical Positivism and Speculative
Philosophy"; later he published "The Pragmatic and Scien—
tific Metaphysics of Charles S. Peirce" (1937), and
"Gestalt Psychology and Physicalism" (1939). But Leonard

was not to be a chronicler but a creater of philosophy.

 

Soon he became involved in the philosophy of science and
logic. In 1940, collaborating with Nelson Goodman, he
published the classic "The Calculus of Individuals and

Its Uses." Some years later, in 1947, the Duke University

Press published Leonard's first book, Logic, Language, and

 

the Methods of the Sciences. Two years later Leonard was

 

invited to chair the Department of Philosophy at Michigan
State University. At about this time he renounced associa—
tion with the Logical Positivist movement in the interesting
"Ethical Predicates." This was followed by another article
on logic, "Two-Valued Truth Tables for Modal Functions"
(1951). His interests were, however, turning toward the
philosophy of language and the philosophy of logic, as is
evident in the philosophical and original logic text

Principles of Right Reason (1957). This book was revised

 

and reprinted ten years later as Principles of Reasoning.

 

During the interim Leonard's work on the philosophy of logic

and theory of meaning continued with the articles mentioned

219

in Chapter I. He also wrote "The Logic of Existence"
(1956) and "Essences, Attributes, and Predicates" (1963).
These two articles were instrumental in originating what
later came to be known as "Free Logic." After two years
at the Princeton Institute for Advanced Study, Professor
Leonard was appointed, in 1961, to the coveted position of
University Professor at Michigan State University. He
returned to the philosophy of science to publish another

substantial work, The Use and Abuse of Measurement as a

 

Facet of Scientific Research (1962). He also made an

 

excursion into metaphysics with "The Mental and the
Physical" (1964). However, his major devotion was the
development of an expanded formal language, "Language W."
Unfortunately this work was never finished, although he
did publish "Synonymy and Systematic Definitions" in 1967
and he had written a substantial set of unpublished notes,
"Notes on Language W" which are now held by Michigan State
University. Professor Leonard had begun preparing for pub—
lication a collection of his major essays; he had prepared
introductions for some of these. Richard Rudner is now
about to publish this collection of the late Professor
Leonard's work.2

The following are Henry Leonard's publications:

Books and Monographs:

Logic, Language and the Methods of the Sciences (Durham,
N.C.: Duke University Press, 1947).

Principles of Right Reason (New York: Henry Holt and Co.,
1957).

 

 

220

Principles of Reasoning (New York: Henry Holt and Co.,
1967).

The Use and Abuse of Measurement as a Facet of Scientific
Research (Seattle, Wash.: University of Washington
Press, 1962).

 

Articles:

"Singular Terms," Harvard University, Graduate School of
Arts and Sciences, Summaries . . . (of Ph.D.
Dissertations). 1932, 298-300.

"Logical Positivism and Speculative Philosophy," in Philo-
sophical Essays for Alfred North Whitehead, ed. by
Otis H. Lee (New York: Longmans, Green and Co.,
1936), 125-52.

"Gestalt Psychology and Physicalism," Journal of Unified
Science (Erkenntnis).

"The Calculus of Individuals and Its Uses," with Nelson 1
Goodman, Journal of Symbolic Logic, V (1940),

45-55.

"Ethical Predicates," Journal of Philosophy, XLVI (1949),
601-7.

"Two-Valued Truth Tables for Modal Functions," in Structure,
Method, and Meaning, ed. by Langer, Henle and
Kallen (New York: Liberal Arts Press, 1951), 42-67.

"The Logic of Existence," Philosophical Studies, XII (1956),
49-64.

"Interrogatives, Imperatives, Truth, Falsity, and Lies,"
Philosophy of Science, XXVI (1959), 172-86.

"Authorship and Purpose," Philosophy of Science, XXVI
(1959), 277—94.

"Authorship of Signs," Papers of the Michigan Academy of
Science, Arts, and Letters, XLV (1960), delivered
at the 1959 meeting, 329-340.

"Reply to Professor Wheatley," Philosophy of Science,

XXVIII (1961), 55-64.

"Essences, Attributes and Predicates," Proceedings and
Addresses of the American Philosophical Association,
XXXVII (1964), the presidential address, 25—51.

"The Mental and the Physical," Centennial Review, VIII
(1964), 337-52.

"Synonymy and Systematic Definitions," The Monist, LI
(1967), 35-68.

"Notes on Language W," in the possession of Professor
Herbert Hendry, Department of Philosophy, Michigan
State University, East Lansing, Michigan.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Reviews:

"A System of Logistic by W. V. O. Quine," Isis, XXIV (1935),
168-72.

"Testibility and Meaning by Rudolf Carnap," Journal of
Symbolic Logic, II (1937), 49—50.

 

 

 

221

"The Pragmatism and Scientific Metaphysics of C. S. Peirce:
A Review of Volumes V and VI of the Collected
Papers of Charles Sanders Peirce," Philosophy of
Science, IV (1937), 109—21.

"The Logical Syntax of Language by Rudolf Carnap," Isis,
XXIX (1938), 163-67.

"The Thought of C. S. Peirce by T. A. Goudge," University
of Toronto Quarterly, XXI (1951), 93-97.

1Thomas M. Olshewsky, Problems in the Philosophy of
Language (New York: Holt, Rinehart, and Winston, Inc.,
1969), p. 747. John Passmore, A Hundred Years of Philosoppy
(New York: Basic Books Inc., 1957, revised 1968), p. 470n.
John R. Searle, The Philospphy of Language (London: Oxford
University Press, 1971), p. 145.

 

 

2William Callaghan, "Henry Leonard at Michigan
State University," The Logical Way of Doing Things, ed. by
Karel Lambert (New Haven, Conn.: Yale University Press,
1969), essays in honor of Henry S. Leonard, 295-312. Con—
versations with students, friends, and colleagues of Henry
Leonard: Herbert Hendry, Richard Rudner, and William
Callaghan.

 

 

 

 

.;\'_.444\.~.‘.\‘
. ‘

.4 .."""
'44 4.. 4»-
4-....‘C.....

 

.444... 4
H 2443.. «414-44, ~.

4. u...
"‘.‘.\" ... ..
n

444...«.-. «ix-:4 -
.4~. 44

4 4. .'.. . . . 2..
. . . . . 4. .4
"4‘ x 4 - 4 4 . ‘
..-- ..
It: 4 .‘.? . -4444- ... . .
44.4.1441..." -
4 .4444 44 4. $4.41"- .4
.. “42..."... .531: 9-...-.
44.-.31.-..“ , 3-
_4 4:. _4 4 4‘
, .4.“
.. .4 44%....“
. ......433 4

‘4 v ‘
"‘.. . . 541;. v
. --.\-.-4\\-‘.--44\ - ‘4
\3 ‘ M44 .... 4 ..:;::..4_..
4 .. . - '
444- “444-. .‘.,4.-\-....... -.~. 44 4A 4'4-1- CF ‘ ~
.:4:~.4:4 44} ‘44»4-" 441-7“ 4 4 44. >4
\_ 44 w

. .. .
\ ‘x ‘4‘“ "V‘c\‘h - ‘
.... 4...... .... -.
44.433344. ."3 3.1;“: 4
'1‘: h “ ‘ “‘ . I
14- . . »
N“ ‘ "‘
~144:w:'14 444 4.44 4
.44... ...-...
W... “Axum... .41.. Y‘ ‘7 ‘
-¢-~~'>44 44\ 3744:044'4-44
'- 'r 4433

14 4 :51; .'4 :‘4 ..‘4
a“ ‘::A‘\‘:‘ I‘\
“3“ ‘W
44
W“ “‘4. £11.74 4??»444444‘4a? hunk“
V4444“- 444444. ""31?“ «mun-\- 4- m1?“ 4.. .44
34v»: '34,: 21:4}:44-3‘ ”a ‘53:: 4'4“ W4
-.-.4
~44 4: 3:?“ fl '?‘:‘4‘ " M
L
44444.. '3": :
4Y1‘44'17V‘VL‘.“ «~44 <‘44 41‘“. 4 44 ~ 5".“ 144:7“?
. 'I-v .
4 “44‘444144454-444'944 >444 .44\ ‘4‘? 444444‘1154‘2'”
...‘-4.: 44:3. «‘W‘ «(4% -v-44~4~v 44
.. v‘fi.‘“‘"“" “v .

4‘ ..
. -
_.
n:- \v‘qm' v4 4 ‘4
3.2-,“ V 44 .'11r‘
. - ‘47: .11. 154%.." 2:.
'14 > ‘W‘-44w 1“

... ‘...'.UH
" . .. . . - ,
. »x . ‘ v \ ‘ ‘
, . . w
‘4' ‘ . 3.. >4 ,
v
44:.44 1,44 4
4- 144414 - l .u . 3""""
q \ I4I
. ....
”,5.” a}. .. '
4. . an.
4445-4243944Lm44444 714:5";
4 , 4. .4
44m 4:414-«3‘4: y,‘ ‘ 4 R14:
4:44' -
“
..

~41

L‘sé‘é'?‘ ‘1‘:
1,3355?“ 13.“- . mi.“

44"

$4444
4445. 349‘. ‘
i

\ ‘W‘.”
._ . .:.,E..{:~.~.

V“ '
3‘4‘:\: \I.~ I in
4‘61\'-~lt- ..:‘:‘\‘ ‘;;.“ '
Ear-2:333:11. '{~:’:"" .
I- 1"
. 4 - > ‘ 'I
“in": .72.} .. “.-.? ...: . 7'
42451394 s‘éS. 4 4 u»

 

I .u ‘ "":,“'..."
W..u.s...‘.&m
H... ('m'”
~\\"\- >4
.' .I

I 4
r >I‘4-I‘
.“.,...l.. ... --
nu-