V- O

s ^

OSWALD HANFLING

Ayer
 THE

GREAT
PHILOSOPHERS

Consulting Editors
Ray Monk and Frederic Raphael
 Oswald Hanfling

A.J. AVER

Analysing What We Mean

PHCENIX
 A PHOENIX PAPERBACK

First published in Great Britain in 1997 by

Phoenix, a division of the Orion Publishing Group Ltd
Orion House
5 Upper Saint Martin's Lane
London, WC2H 9EA
Copyright © Oswald Hanfling 1997
All rights reserved.
No part of this publication may be reproduced, stored in a retrieval
system, or transmitted iii any form or by any means, electronic,
mechanical, photocopying, recording, or otherwise, without
the prior permission of the copyright owner. Oswald Hanfling
has asserted his moral right to be identified as the author of this work.
A catalogue reference is available from the British Library
ISBN 0 753 80182 5

Typeset by Deltatype Ltd, Birkenhead, Merseyside

Printed in Great Britain by
Clays Ltd, St Ives pic
 A.j. AVER
Analysing What We Mean
'T^he trouble with philosophy, it is sometimes said, is
JL that nothing ever gets settled. Questions posed by the
ancient Greeks are still being asked today and there seems
to be no agreement as to the right answers. Whether, or to
what extent, these complaints are justified will depend on
what the purpose of philosophy is taken to be. But one
may feel that there is something radically wrong with the
whole enterprise and that, if only one could put one's
finger on the flaw, the subject could be transformed, or
replaced by something more fruitful. Such a view was
expressed by Ayer in the opening sentences of Language,
Truth and Logic, together with a confident belief that he
had found the flaw:

The traditional disputes of philosophers are, for the

most part, as unwarranted as they are unfruitful. The
surest way to end them is to establish beyond question
what should be the purpose and method of philosoph-
ical inquiry. And this is by no means so difficult a task as
the history of philosophy would lead one to suppose.
(LTL 45)

What did Ayer mean by 'the traditional disputes of

philosophers'? What is philosophy? According to at least
one influential understanding, it is a quest for tran-
scendent truths, beyond the reach of ordinary or scientific
inquiry. This conception of philosophy was one of the
main targets of Ayer's critique.

We may begin by criticizing the metaphysical thesis

that philosophy affords us knowledge of a reality
transcending the world of science and common sense
... One way of attacking a metaphysician who claimed
to have [such] knowledge ... would be to inquire from
what premises his propositions were deduced. Must he
not begin, as other men do, with the evidence of his
senses? ... [But] surely, from empirical premises noth-
ing ... super-empirical can legitimately be inferred.
(LTL45-6)

'The function of philosophy', he declared, 'is wholly

critical'; 'it is an activity of analysis'; and the idea that
philosophy is 'a search for first principles' was a 'supersti-
tion from which we are freed by the abandonment of
metaphysics' (LTL 37, 62).
With these and similar assertions the young Ayer
embarked on a course of discussion that was designed to
shake the philosophical^ establishment. As we shall see,
and as Ayer would later be thefirstto admit, the book has
many flaws; but few would deny that it was, and remains,
a major contribution to philosophy. One may also be
impressed, and indeed astonished, by its sheer virtuosity
and the extent of its coverage - all the more so in view of
the youthfulness of its author.
 A 'SUCCES DE SCANDALE'

/T began writing Language, Truth and Logicf, Ayer

A reported later, 'in the Christmas vacation of 1933-4
and finished writing it in July 1935, three and a half
months before my twenty-fifth birthday ... The book
enjoyed an immediate succes de scandale.'^ A new edition
appeared in 1946, after which it 'approached the status of
a best seller', with new impressions appearing 'almost
annually for the next twenty-five years'.^ Ayer produced
many other important books, but LTL remains the work
by which he is best known.
In a television interview of 1979, Ayer was asked what
he now saw as the main defects of his youthful work. 'I
suppose', he replied, 'the most important of the defects
was that nearly all of it was false.' This, like some of the
claims in the book itself, was an exaggeration. It is true
that already in the second edition he conceded that 'the
questions with which [the book] deals are not in all
respects so simple as it makes them apear';^ but the
general approach of that book remained with him
throughout his life, as will be illustrated below. The main
change is that the confident, and often over-confident,
style of the early book is replaced by a cautious, painstak-
ing investigation of issues he had earlier disposed of too
quickly and easily. LTL will be the basis of the present
book, but from time to time we shall move forward
to related discussions in later works.
The writing of LTL came about in a rather accidental
way. In 1931, Ayer tells us, he was given two terms leave
of absence from his tutorship at Oxford.

I proposed to spend them in Cambridge, leaming from

Wittgenstein, but [my former tutor] Gilbert Ryle per-
suaded me to go to Vienna instead. He had met Moritz
Schlick, the leader of the group of philosophers and
scientists ... who entitled themselves the Vienna Circle,
at some international congress ... and thought that it
would be a good thing if I could discover what work
was going forward under his auspices. For my part I
had become engaged to be married and thought that
Vienna would be a nice place in which to spend a
honeymoon. I was married to Renee Lees on 25
November 1932 and we went almost immediately to
Vienna.^

He rapidly learned enough German to enable him to

attend the meetings of the Circle.
To anyone acquainted with the philosophy of the
Vienna Circle, its influence on Ayer will be quickly
apparent. This is true both of its ideas, which included
'the elimination of metaphysics', and of the boldness of
its style. But Ayer was also deeply influenced by the British
empiricist tradition of Locke, Berkeley and Hume. (His
appeal to 'empirical premises' and 'the evidence of the
senses' is a sign of this.) There was, however, an important
difference between the new empiricism and that of
previous philosophers: the new empiricism was about
meaning rather than knowledge. The charge against 'the
metaphysician' was not merely that his claims were
unsupported by suitable premises: it was that they were
meaningless, because they failed to satisfy certain condi-
tions that must be satisfied if a statement is to have
meaning.
The new philosophy became known as 'logical empiri-
cism' or, more commonly, 'logical positivism'. The first
has the advantage of indicating the affinity with the
empiricism of Locke, Berkeley and Hume, whose influence
is apparent in Ayer's work; while the qualification 'logical'
indicates the distinctive concern with logical analysis.

THE CRITERION OF VERIFIABILITY

hat not all words or sentences are meaningful is

T obvious if we consider such examples as the nonsense
rhymes of Edward Lear. But a sentence may be nonsens-
ical in less obvious ways. This is so, according to Ayer, in
the case of claims about a super-empirical reality. But
what criterion, if any, is there for distinguishing sense
from nonsense? Ayer's reply to this question was one of
the most prominent features of his book. He gave it as
follows:

The criterion which we use to test the genuineness of

 apparent statements of fact is the criterion of verifiabil-
ity. We say that a sentence is factually significant to any
given person, if, and only if, he knows how to verify the
proposition which it purports to express - that is, if he
knows what observations would lead him, under
certain conditions, to accept the proposition as being
true, or reject it as being false. If, on the other hand,
the putative proposition is of such a character that the
assumption of its truth, or falsehood, is consistent with
any assumption whatsoever concerning the nature of
his future experience, then, as far as he is concerned, it
is ... a mere pseudo-proposition. The sentence express-
ing it may be emotionally significant to him; but it is
not literally significant. And with regard to questions
the procedure is the same. We inquire in every case
what observations would lead us to answer the ques-
tion, one way or the other; and, if none can be
discovered, we must conclude that the sentence under
consideration does not, as far as we are concerned,
express a genuine question, however strongly its
grammatical appearance may suggest that it does.
(LTL48)\

The criterion of vetifiability should be distinguished

from the 'verification principle' of the Vienna Circle: 'The
meaning of a statement is the method of its verification.'
The criterion provides an answer to the question 'When is
a statement meaningful?' or 'What kinds of statement are
meaningful?', while the principle is a claim about what
meaning consists in: it is an answer to the question 'What

8
is meaning?' The criterion is, however, dedudhle from the
principle. If meaning is identified with method of verifica-
tion, then it follows that if there is no method of
verification - if the statement is not verifiable - then it
must be devoid of meaning. There is, however, no
entailment in the other direction, so that one may put
forward the criterion without commitment to the princ-
iple. This was the case with Ayer, though his terminology
is not always consistent (he sometimes referred to his
criterion as 'the principle of verification'). The verification
principle, as advocated by members of the Vienna Circle,
calls for separate discussion, but this will not be our
concern.
The first question that may occur to one on reading the
above passage from Ayer is why, or whether, one should
accept his criterion. That there is a connection between
meaning and verification would hardly be doubted. A
good way of finding out what, if anything, a person
means by a given statement is, indeed, to ask what 'would
lead him ... to accept [it] as being true, or reject it as being
false'. And if the answer is 'Nothing', then we might
perhaps conclude that, as far as he is concerned, the
statement is meaningless. But is this necessarily so? One
of those who refused to accept the criterion was Father
Copleston, who held certain religious beliefs that would
hot satisfy it. In a debate broadcast in 1949, Ayer tried
to persuade Copleston of the error of his position.
He invented the word 'drogulus' for this purpose.

Suppose I say, 'There's a drogulus over there' and you

 say ... 'What's a drogulus?' Well, I say, 'I can't describe
what a drogulus is, because it's not the sort of thing
you can see or touch, it has no physical effects of any
kind, but it's a disembodied being.' And you say, 'Well
how am I to tell if it's there or not?' and I say, 'There's
no way of telling. Everything's just the same if it's there
or it's not there. But the fact is it's there. There's a
drogulus there standing just behind you, spiritually
behind you.' Does that make sense? (ML 41-2)

But Copleston stood his ground against this barrage.

Given that 'drogulus' means a disembodied spirit, he
replied, he would say that the statement offered by Ayer
was 'either true or false, whether one can verify it or not'.
The artificial 'drogulus', taken in isolation, may indeed
strike one as meaningless; but Copleston, in picking up
the phrase 'disembodied spirit', was appealing to a wider
discourse which may be thought to give it meaning. A
similar point may be made about one of Ayer's examples
of metaphysics in LTL: 'The Absolute enters into, but is
itself incapable of, evolution and progress' (LTL 49). 'One
cannot', he commented, 'conceive of an observation
which would enable one to determine' whether this is
true; and until the author of the statement 'makes us
understand how the proposition that he wishes to express
would be verified, he fails to communicate anything to
us'. Now a reader confronted by that example might
indeed regard it as meaningless. But would this be for the
reason given by Ayer? Ayer had taken the example 'at
random' from a work by F.H. Bradley, without any

10
reference to the arguments in which it was embedded; but
perhaps it would not seem meaningless to someone who
had read those arguments - even if it failed to satisfy
Ayer's criterion.
In 1946 there appeared a second edition of Ayer's book,
with a substantial introduction in which he modified
several of its main tenets. In the following extract he
conceded that metaphysics cannot be disposed of as easily
as he had thought; butfirsthe addressed the fundamental
question about the status of his criterion: what reason is
there for accepting it? Is it true to what we actually mean
by 'meaning'?

In putting forward the principle of verificatiorr as a

criterion of meaning, I do not overlook the fact that the
word 'meaning' is commonly used in a variety of
senses, and I do not wish to deny that in some of these
senses a statement may properly be said to be
meaningful even though it is neither analytic nor
empirically verifiable. I should, however, claim that
there was at least one proper use of the word
'meaning' in which it would be incorrect to say that a
statement was meaningful unless it satisfied the prin-
ciple of verification; and I have, perhaps tendentiously,
used the expression 'literal meaning' to distinguish this
use from the others ... Furthermore, I suggest that it is
only if it is literally meaningful, in this sense, that a
statement can properly be said to be either true or
false. Thus, while I wish the principle of verification
itself to be regarded, not as an empirical hypothesis,

11
 but as a definition, it is not supposed to be entirely
arbitrary. It is indeed open to anyone to adopt a
different criterion of meaning and so to produce an
alternative definition which may very well correspond
to one of the ways in which the word 'meaning' is
commonly used. And if a statement satisfied such a
criterion, there is, no doubt, some proper use of the
word 'understanding' in which it would be capable of
being understood. Nevertheless, I think that, unless it
satisfied the principle of verification, it would not be
capable of being understood in the sense in which
either scientific hypotheses or common-sense state-
ments are habitually understood. I confess, however,
that it now seems to me unlikely that any metaphysi-
dan would yield to a claim of this kind; and although I
should still defend the use of the criterion of verifiability
as a methodological principle, I realize that for the
effective elimination of metaphysics it needs to be
supported by detailed analyses of particular meta-
physical arguments. <LTL 20-1)

In describing his criterion as a 'methodological prin-

ciple', Ayer avoids a difficulty I have not yet mentioned. If
the criterion were itself to be understood as a meaningful
Statement, then we would need to consider how it is to be
verified. And if there is no answer, or no clear answer, to
this, then the meaningfulness of the criterion itself will be
put into doubt. Thus the dismissal of certain kinds of
statements as meaningless might itself turn out to be
meaningless! But this difficulty cannot arise if the cri-

12
terion is put forward as a methodological principle, for
such a principle cannot be described (any more than a
suggestion or a proposal) as true or false; and neither,
therefore, can it be subject to the test of verifiability.
But this still leaves the question: should we accept
Ayer's proposal? Is his 'methodological principle' a good
one - one that we ought to endorse? Ayer tells us that he
would still 'defend the use of the criterion', but he does
not say what the defence would be. Perhaps it can be
argued, in favour of the criterion, that it is useful to
distinguish the two kinds of statement on the lines
suggested, even if the description of one of them as
'meaningless' is withdrawn. But how exactly is the
criterion to be formulated and applied? This question is
not as straightforward as it may seem.

THE MEANING OF 'VERIFICATION'

n formulating his criterion, Ayer had to face a difficulty

I about statements whose significance no one, including
himself, would wish to deny.

Consider, for example, the case of general propositions

of law - such ... as 'arsenic is poisonous'; 'all men are
mortal'; 'a body tends to expand when it is heated'. It
is of the very nature of these propositions that their
truth cannot be established with certainty by any finite
series of observations. (LTL 50)

13
This is SO; according to Ayer, because of a discrepancy
between the scope of such propositions and the observa-
tions on which they are based. The number of observa-
tions is, necessarily, finite; but the statements in question
are not restricted to afinitenumber of instances. ('All men
are mortal' does not refer only to people observed so far.)
Whether it follows that such statements 'cannot be
established with certainty' is debatable, but Ayer, like
many others, took it to be so. But then, if to verify is to
'establish with certainty', it will follow that these state-
ments, not being verifiable, must be declared 'not literally
significant', along with those of metaphysics.
To acconmiodate such statements, Ayer introduced a
distinction between 'strong' and 'weak' verification. 'A
proposition is ... verifiable in the strong sense of the term,
if, and only if, its truth could be conclusively established
... But it is verifiable in the weak sense if it is possible for
experience to render it probable' (LTL 50). He would settle
for verification in the weak sense. 'It is only if a negative
answer must be given' to the question of weak verification
'that we conclude that the statement under consideration
is nonsensical' (LTL 52). To make his position clear, he
introduced the term 'experiential proposition'.

Let us call a proposition which records an actual or

possible observation an experiential proposition. Then
we may say that it is the mark of a genuine factual
proposition, not that it should be equivalent to an
experiential proposition, or any finite number of exper-
iential propositions, but simply that some experiential

14
 propositions can be deduced from it in conjunction
with certain other premises without being deducible
from those other premises alone. (LTL 52)

The 'certain other premises' would presumably be about

suitable circumstances. Thus if 'arsenic is poisonous' is
true, it will follow that if one observes ('experiences')
Smith taking arsenic, one will also observe Smith becom-
ing ill or dying. And this would suffice to enable the
general statement to pass the test.
There is another important difference between this
formulation of the test and that with which we began.
What was required there was, as we saw, that the person
making a statement should know what observations
'would lead him ... to accept [it] as being true, or reject it
as being false'. This is not satisfactory, because it does not
distinguish between observations that really support the
statement and those that do not. All sorts of observations
might 'lead' someone to accept or reject a putative
statement, including the 'nonsensical' statements of
metaphysics. In the formulation just quoted, however, the
subjectivity of 'leading someone to accept' a proposition
is replaced by the logical relation of 'being deducible'.
But, we may ask, in what sense would the relevant
'experiential propositions' be deducible? If I am in the
right place at the right time to observe someone taking
arsenic and dying, it does not follow that I will observe
this. It is indeed obvious that we often fail to observe what
is there to be observed. Perhaps the criterion could be
reworded to take account of this point.

15
 There are other difficulties, however. Having introduced
the requirement of 'weak' verifiability, Ayer commented
that it 'seems liberal enough' (LTL 52) to accommodate
laws of nature; but in his new introduction he had to
concede (in view of an objection from Isaiah Berlin) that
'in fact it is far too liberal, since it allows meaning to any
statement whatsoever' (LTL 15). To illustrate this point he
invented a clearly nonsensical statement.

The statements 'the Absolute is lazy' and 'If the

Absolute Is lazy, this is white', jointly entail the
observation-statement 'this is white', and since 'this is
white' does not follow from either of these premises,
taken by itself, both of them satisfy my criterion of
meaning. (LTL 15)

To avoid this embarrassing conclusion, he put forward

an amended, rather complicated version of the criterion,
but this will not be discussed here. It was, in any case,
soon shown that the new version too was untenable.
Further attempts were made by Ayer and others, but
towards the end of his long career Ayer was still comment-
ing on 'the continual failure of attempts to [formulate the
criterion] in such a way as to find a middle ground
between the over-strict requirement ... and the over-
indulgent licensing of gibberish'. He added, however, that
he still did not want to 'discard the concept' of verifiabil-
ity as a criterion of significance.^

16
 VERIFICATION AND ANALYSIS

T "^here is another, simpler way of objecting to the above

formulation of Ayer's criterion, and this will take us
to one of the major themes of the book: analysis.
Consider the simple conjunction The Absolute is lazy and
this is white', and suppose this were put forward as
meaningful on the ground that it entails the observation-
statement This is white'. Would Ayer have to concede
that the statement as a whole is meaningful?
It would seem that the proper way to deal with such
examples is to separate the conjuncts, enabling the first to
be described as nonsense and the second as meaningful.
However, the required separation will not always be so
easy. Take the case of statements about God, one of the
major targets of Ayer's criterion. Suppose someone
explained the existence of order and beauty in the world
by reference to God's creation. Accommodating this to
Ayer's formulation, he might claim that observation-state-
ments to the effect that there is order and beauty in the
world are deducible from the statement that the world was
created by God. Would Ayer have to admit that this
statement passes his test? We can see what the answer
would be from one of his examples. His response was to
separate out the meaningful part of the statement, as he saw
it. Thus 'if the sentence "God exists" entails no more than
that certain types of phenomena occur in certain se-

17
quences, then to assert the existence of a god' will mean
no more than that - though, as he went on to observe, 'no
religious man would admit that this was all he intended
to assert' (LTL 152). Again,

if a man tells me that the occurrence of thunder Is

alone both necessary and sufficient to establish the
truth of the proposition that 'Jehovah is angr/, I may
conclude that, in his usage of words, the sentence
'Jehovah is angry' is equivalent to 'It is thundering'.
(LTL 153-4)

Here we see that the deducibility of observation-state-

ments was not enough: what was needed was to analyse
the statements under investigation, separating out what is
meaningful from what is not. But whereas this seemed
straightforward in the case of 'The Absolute is lazy and
this white', it is not so in the case of statements about
God, for of course the 'religious man', for one, would not
accept the separations proposed by Ayer.
The importance of analysis in Ayer's philosophy is more
radical, however. Let us approach this by considering his
choice of 'This is white' as a typical observation-state-
ment. What did he mean by 'this'? He did not mean such
things as tables and chairs: what he meant was a 'sense-
content' that the speaker was experiencing. The direct
objects of observation, he held, are 'sense-contents' (or
'sense-data': he used various terms at different times) and
not, as one might assume, such things as tables and
chairs. Material things 'are logical constructions out of
sense-contents' (LTL 86); and 'to say anything about [a

18
material thing] is always equivalent to saying something
about [sense-contents] (LTL 185).
What did Ayer mean by 'sense-contents'? In LTL, he
gave little attention to this question. He introduced the
term by reference to the 'ideas' of Locke and Berkeley; but
since their use of this term is notoriously obscure and
controversial, it hardly throws light on the difficulty.
What is clear is that the term in question is to stand for
something with which an observer is more directly related
than he is with ordinary objects, so that his knowledge of
the latter is either inferential (he infers that there is a table
before him from the occurrence of suitable sense-con-
tents) or - as in the passage quoted above - it is a matter of
'logical construction' (to say that there is a table before
him is to say that sense-contents X, Y and Z are occurring
or will occur). But Ayer did not explain what kinds of
occurrences or entities X, Y and Z would be.
In later writings he was more sensitive to the need for
explaining and justifying his talk of sense-contents or
sense-data. In The Problem of Knowledge he devoted a
section to what he called 'the legitimacy of sense-data' (PK
105), in which he examined arguments for and against
introducing such terminology, concluding that 'the argu-
ment from illusion' provides the strongest reason in
favour of it. This argument proceeds from the premise
that things sometimes seem to be otherwise than they
really are (there seems to be an oasis before me, but it is
only a mirage). It is then pointed out that the experience
of seeing (say) a mirage is just like the experience of seeing

19
the real thing; and from this it is concluded that 'when-
ever anyone perceives, or thinks that he perceives, a
physical object, it must then seem to him that he perceives
something or other' (PK 104). The final step is tp 'pass
from "it seems to me that I perceive x" to "I perceive a
seeming-;^", with the implication that there is a seeming-;i^
which I perceive' (PK 105). It turns out, then, that the
direct objects of perception - which in LTL he had called
'sense-contents' - are the 'seeming'-objects to which the
argument from illusion appears to lead.
One difficulty of this line of argument may be men-
tioned here. The claim that 'whenever anyone perceives
... a physical object, it must then seem to him that he
perceives something or other' would not normally be
recognized. On the contrary, 'it seems to him that he
perceives a chair' would be held to imply that he does not
really perceive a chair. Some argument is needed to
defend Ayer's reversal of the logic of these expressions as
normally understood.
Let us turn to another aspect of the programme of LTL.
The task that Ayer set himself was to show that all
genuine statements could be analysed into statements
about sense-contents. Whatever could not be so analysed
must be consigned to some other category than that of
genuine statement. (This, it turned out, was what the
criterion of verifiability really amounted to.) 'Our
remarks', he wrote, 'apply to all empirical propositions
without exception, whether they are singular, or particu-
lar, or universal. Every [such] proposition is a rule for the

20
anticipation of future experience [of sense-contents]' (LTL
134).
But why 'a riile for the anticipation', etc.? This brings us
to phenomenalism - the view that apparently categorical
statements about material things ('Here is a table', 'The
table is white', etc.) are to be analysed in hypothetical
terms ('If anyone looks this way, he will experience sense-
contents X, Y and Z'; or 'If anyone were to look this way,
he would', etc.). It is in this sense that the analysis into
sense-contents must be understood: otherwise, it appears,
it might have the absurd consequence of consigning
ordinary ideas about tables and chairs to the class of
metaphysical 'nonsense'. We normally think of such
objects as enduring at times when they are not imder
observation, but this belief cannot be tested by observa-
tion. It is here that phenomenalism, with its use of
hypothetical statements, must come to the rescue.

This explains how it is possible for a material thing to

exist throughout a period when none of its elements
are actually experienced: it is sufficient that they should
be capable of being experienced - that is, that there
should be a hypothetical fact to the effect that, if
certain conditions were fulfilled, certain sense-contents
... would be experienced. (LTL 186)

Statements about material things are to be analysed,

accordingly, 'in terms of the hypothetical occurrence of
sense-contents' (LTL 187).
But^ we may wonder, is this really what we mean when
we speak of tables and chairs? What, apart from avoiding

21
the allegations of 'metaphysics', is the advantage of
representing the meaning of our statements in this way?
According to Ayer, it 'serves to increase our understanding
of the sentences in which we refer to material things' (LTL
91). He conceded that 'there is of course a sense in which
we already understand such sentences ... as "This is a
table", or "Pennies are round"', but, he argued, we

may very well be unaware of the hidden logical

complexity of such statements, which our analysis ...
has just brought to light. And, as a result, [we] may be
led to adopt some metaphysical belief, such as the
belief in the existence of material substances or
invisible substrata. (LTL 91 b)

Here one might object that Ayer's concession - 'there is

... a sense in which we already understand' the statements
in question - does not go far enough. We do not merely
understand those statements: we regard them as having
priority over statements about 'seeming'-objects, etc. Thus
'This is a seeming-table' would have to be explained by
reference to 'This is a table' and not the other way round:
the 'seeming' terminology is parasitic on the other.
In a later work. The Central Questions of Philosophy,
Ayer's views about phenomenalism had changed to a
considerable extent. Here he defended the need for a
vocabulary of 'sense-data', or something of that kind, in
another way, by distinguishing between the 'appearances'
of things and the things themselves - a distinction to
which, he thought, a 'naive realist' could not do justice.

22
His argument in the following extract is not the 'argu-
ment from illusion'; it is an argument about the various
and complex implications of such statements as 'Here is a
table'.

We think of the physical object as preserving its

identity in the various guises in which it appears to us,
but exactly how is this achieved? What is it that
remains constant while its appearance varies? If the
physical object is known to us only through its various
appearances, in what way can we distinguish it from
them? The naive realist ignores these questions ... He
has no convenient vocabulary by means of which he
can refer to the appearances of things, independently
of the things of which we take them to be appearan-
ces. But if we want to discuss the relation of physical
objects to their appearances, we do need such a
vocabulary, and it is just this that the introduction of
terms like 'sensible quality' or 'sense-datum' has been
intended to provide. We may, indeed, not wish to be
committed to all the implications which their use has
been made to carry. Exactly how any such term is to be
construed in order to be acceptable is a question which
we shall have to examine. All that kam now suggesting
is that something of this sort is needed ...
We need only consider the range of the assumptions
which our ordinary judgements of perception carry. To
begin with, there are the assuroptions ... involved in
characterizing anything as a physical object like a table.
It has to be accessible to more than one sense and to

23
more than one observer and it has to be capable of
existing unperceived. In addition, it has to occupy a
position or series of positions in three-dimensional
space and to endure throughout a period of time ...
Neither is it only a question of the validity of these
general assumptions. Our perceptual judgements are
seldom indefinite, in the sense that we claim only to
perceive a physical object of some sort or other. In the
normal way, we identify it as a thing of some specific
kind, and this brings in further assumptions, as, for
example, that the object is solid, or flexible, or that it is
not hollow. These further assumptions may relate to
the purposes which the object serves, as when we
identify something as a pen-knife, or a telephone: they
may relate to its physical constitution, as in the
identification of an object as an orange or an apple,
which denies its being made of wax. They may
presume on the deliverances of other senses, as when
our descriptions of an object which we believe our-
selves to be seeing or touching carries implications
about the way in which it tastes or sounds or smells.
But now can it seriously be maintained that all this
can fall within the content of a single act of percep-
tion? Can my present view of the table, considered
purely in itself as a fleeting visual experience, conceiv-
ably guarantee that I am seeing something that is also
tangible, or visible to other observers? Can it guarantee
even that I am seeing something which exists at any
other time than this, let alone something that is made
of such and such materials, or endowed with such and

24
 such causal properties, or serving such and such a
purpose? I think it evident that it cannot. But if these
conclusions are not logically guaranteed by the con-
tent of my present visual experience, one is surely
entitled to say that they go beyond it, and just this is
what I take to be meant by saying that my judgement
that this is a table embodies an inference. It embodies
an inference, not in the sense that it results from any
conscious process of reasoning, but just in the sense
that it affirms more than can logically be entailed by
any strict account of the experience on which it is
based. What I mean here by a strict account Is one that
is tailored to the experience, in that it describes the
quality of what is sensibly presented/ without carrying
any further implications of any sort. In the normal way,
we do not formulate such propositions because we are
interested not in the data as such, but in the interpreta-
tions which we have learned to put upon them. I
cannot, however, see any logical reason why they
should not be formulable.
If I am right on this point, the naive realists are
wrong in so far as they deny that our ordinary
judgements of perception are susceptible of analysis, or
deny that they embody inferences which can be made
explicit. (CQ 79-81)

It would indeed be wrong to deny that 'our ordinary

judgements of perception ... embody inferences' of the
kind mentioned, which, as Ayer makes clear, are various
and extensive. But what are we to make of his claim that

25
those judgements 'are susceptible of analysis'? Are we to
suppose that 'Here is a table' can be translated into a set of
statements about the occurrence of sense-contents (or
'percepts') of the kind mentioned? If this is what he
thought in LTL, it was no longer so now.

I am not suggesting that physical objects are reducible

to percepts, if this is taken to mean that all the
statements that we make about physical objects, even
at the common-sense level, can be adequately trans-
lated into statements which refer only to percepts. If
. the demand for an adequate translation requires that
the statements referring to percepts set out necessary
and sufficient conditions for the truth of the statements
about a physical object which they are meant to
replace, I think it unlikely that it can be satisfied ...
There is, on the one hand, the difficulty that a visuo-
tactual continuant may be represented by an indefinite
variety of percepts in an equal variety of contexts, so
that if the percepts in question had not occurred, some
others would have done as well, and, on the other
hand, the objection that any description of a particular
set of percepts will be bound to leave open at least the
logical possibility that the observer is undergoing some
illusion. But even if these difficulties could be met,
there is another reason why I do not wish to adopt this
position. The actual percepts that are presented to any
observer, or even to the totality of observers at all
times, are too scanty to answer to our conception of
the physical world ... If the phenomenalist thesis is to

26
 be at all plausible, it has to draw on possible as well as
actual percepts, with the result that most of the
propositions which render its account of the world will
take the form of unfulfilled conditionals. They will state
that if such and such conditions, which are not in fact
realized, were to be so, then such and such percepts
would occur. But apart from the obvious difficulty of
giving a sufficient description of the conditions in
purely sensory terms, I no longer think that such
conditional statements are suitable to play this part.
(CQ 106-7)

With this abandonment of phenomenalism, what is left

of the criterion of verifiability? 'The passage from percepts
to physical objects', he now conceded, is not one 'of
logical construction'. But how, in that case, are statements
about physical objects to be understood? What is their
meaning and how are they to be verified? 'The continued
and distinct existence of physical objects' is, he con-
cluded, 'simply posited'. He would now 'forsake phenom-
enalism for a sophisticated form of realism' (CQ 108). We
may wonder what the younger Ayer would have said to
this. Would he not have regarded the 'positing' of such
objects as a clear example of metaphysics, 'devoid of
factual content'?

27
 STATEMENTS ABOUT THE PAST

/ / ^ ^ u r remarks apply to all empirical propositions

without exception ... Every [such] proposition is a
rule for the anticipation of future experience' of sense-
contents (LTL 134). But what would he say about the
difference between tenses? Are we to suppose that state-
ments about the past are really about the future? This was
indeed his view in LTL. 'Propositions referring to the past',
he stated, 'have the same h3^othetical character as those
which refer to the present and those which refer to the
future' (LTL 134). Thus (to quote from a later work) 'all
that could now be meant by saying that Caesar crossed
the Rubicon was that if we were to look in such and such
history books we should discover that their authors
affirmed it' (CQ 24-5). He was aware that his view might
seem strange, but, he assured the reader,

for my own part I do not see anything paradoxical in

the view that propositions about the past are rules for
the prediction of those 'historical' experiences which
are commonly said to verify them, and I do not see
how else our 'knowledge of the past' is to be analysed.
(LTL 135)

He suspected those who objected to his analysis of

being misled by an assumption that 'the past is somehow
"objectively there" ..., that it is "real" in the meta-

28
physical sense of the terai' (LTL 135). But what makes the
analysis objectionable, and indeed highly paradoxical, is
not that the past is 'somehow objectively there': it is that
it makes no allowance for the distinction between reality
and report: between 'Caesar crossed the Rubicon' and
'Such and such history books say that Caesar crossed the
Rubicon'.
'I do not see', he argued, 'how else our "knowledge of
the past" is to be analysed.' Here, as elsewhere, the
requirement of analysis is more stringent than the original
criterion of verifiability. Asked how I would verify the
statement about Caesar, or what empirical observation
would 'lead me to accept it as being true', I could certainly
answer by reference to a reputable history book, and in
that way establish the credentials of the statement as
meaningful in accordance with the original criterion. But,
as we saw, that criterion had, in the face of certain
difficulties, to give way to the requirement of analysis,
which in its turn produced difficulties including that
about the past.
Here again Ayer showed himself more sensitive in the
second edition. He no longer thought that statements
about the past were really about experiences that would or
might be had in the future, though he still maintained
that such statements could 'be analysed in phenomenal
terms'. The analysis, as now conceived, would be in terms
of sense-contents that the speaker would have experi-
enced if he had been present. A difficulty about this
analysis was, he conceded, that 'these conditions never
can be fulfilled; for they require of the observer that he

29
should occupy a temporal position that ex hypothesi he
does not' (LTL 25). He did not, however, regard this
difficulty as serious, since, he claimed, we could suppose
that the observer occupied a different temporal position
just as we can suppose that he occupies 'a different spatial
position'. But is this true? 1 can suppose that 1 were now in
Cambridge and not in Oxford, but can I suppose that I
was present when Caesar crossed the Rubicon? Could
someone who was present then have been meV
Forty-odd years later Ayer was still trying to analyse
statements about the past on the lines of LTL. Having
reviewed a number of alternatives, he confessed that he
found the position in which he found himself 'counterin-
tuitive'. In the following passage he deals first with
historical statements and then with personal memories
about the recent past.

It is difficult to believe that the statement that Welling-

ton defeated Napoleon at the Battle of Waterloo entails
the report of a [relevant contemporary] conversation,
still less that it is equivalent to that report and a
number of similar pieces of evidence. Should I say that
my current statement that the leaves of the tree under
which I am seated are rustling slightly in the wind will
mean no more to me tomorrow than that I have such
and such a memory, supported perhaps by a meteoro-
logical report?®

The second of these examples involves difficulties of

which Ayer seems not to have been aware. He assumes

30
that he will remember ('have a memory') of the leaves
rustling. But it is quite likely that he will not. We all make
thousands of more or less trivial observations that pass
from our minds quite soon. In that case, must we say that
they did not take place - or that it is meaningless to
suppose that some did?
On the other hand, suppose that Ayer does remember
what happened the day before. In that case 'Ayer remem-
bers that the leaves rustled' will be true. But is this a
statement about a present or future experience? This is
what Ayer must hold, in accordance with the phenomen-
alist analysis. (That is why he speaks, rather oddly, of
'having such and such a memory' rather than of remem-
bering.) In LTL he had claimed that to say of a person that
he remembers 'is to say merely that some of [his] sense-
experiences . . . contain memory images which cor-
respond to the sense-contents which have previously
occurred in the sense-history' of that person (LTL 166).
But one may remember that the leaves rustled without
having such images; and conversely, one may have such
images without remembering that the leaves rustled: one
may be imagining this as opposed to remembering it. (A
similar problem was raised, but not solved, by David
Hume in the eighteenth century.) An alternative view of
memory is that when we say 'Ayer remembers that the
leaves rustled' we are not speaking of the present occur-
rence of images (or anything else); we are saying, rather,
that (a) Ayer knows that the leaves rustled, and (b) he
knows this because he heard them rustling at the time.

31
 MYSELF AND OTHERS

Just as I must define material things ... in terms of their

empirical manifestations, so I must define other people
in terms of their empirical manifestations - that is, in
terms of the behaviour of their bodies, and ultimately
in terms of sense-contents. (LTL 171)

/ T suppose', wrote Ayer, looking back on his career

X towards the end of his life, 'that none of my
philosophical preoccupations has given me so much
trouble as the problem customarily ... described as that of
our knowledge of other minds.'^ The problem is one that,
unlike some of those discussed so far, has troubled many
ordinary people from time to time. I regard other people
as being conscious like myself, but can I really know that
they are? What 1 observe are the physical manifestations
of thoughts and feelings, but, it may be asked, what about
the thoughts and feelings themselves?
For Ayer the problem presented itself in a more acute
form, owing to his reliance on 'sense-contents'. Having
repeated that 'all empirical knowledge resolves itself on
analysis into knowledge of sense-contents', he observed
that a person's sense-contents are 'private to himself. But
from these premises, he pointed out, it seems to follow
that one is

32
 logically obliged to be a solipsist - that is, to hold that
no other people besides himself exist, or at any rate
that there is no good reason to suppose [this]... For it
follows from his prennises, so it will be argued, that the
sense-experiences of another person cannot possibly
form part of his own experience. (LTL 169)

In his response, Ayer insisted that there is no funda-

mental difference between statements about material
objects and statements about other people; and that, in
each case, to suppose that there exists something beyond
sense-contents would be to lapse into metaphysical non-
sense,

just as I must define material things ... in terms of their

empirical manifestations, so I must define other people
in terms of their empirical manifestations - that is, in
terms of the behaviour of their bodies, and ultimately
in terms of sense-contents. The assumption that
'behind' these sense-contents there are entities which
are not even in principle accessible to my observation
can have no more significance for me than the
admittedly metaphysical assumption that such entities
'underlie' the sense-contents which constitute material
things for me ... And thus I find that I have as good a
reason to believe in the existence of other people as I
have to believe in the existence of material things. For
in each case my hypothesis is verified by the occur-
rence in my sense-history of the appropriate series of
sense-contents. (LTL 171)

33
He hastened to asstire the reader that his analysis did not
fail to do justice to the reality of other people.

It must not be thought that this reduction of other

people's experiences to one's own in any way involves
a denial of their reality. Each of us must define the
experiences of the others in terms of what he can at
least in principle observe, but this does not mean that
each of us must regard ail the others as so many
robots. On the contrary, the distinction between a
conscious man and an unconscious machine resolves
itself into a distinction between different types of
perceptible behaviour. The only ground I can have for
asserting that an object which appears to be a
conscious being is not really a conscious being, but
only a dummy or a machine, is that it fails to satisfy one
of the empirical tests by which the presence or absence
of consciousness is determined. If I know that an object
behaves in every way as a conscious being must,
by definition, behave, then I know that it is really
conscious...
It appears, then, that the fact that a man's sense-
experiences are private to himself, inasmuch as each of
them contains an organic sense-content which belongs
to his body and to no other, is perfectly compatible
with his having good reason to believe in the existence
of other men. For, if he is to avoid metaphysics, he
must define the existence of other men in terms of the
actual and hypothetical occurrence of certain sense-
contents, and then the fact that the requisite sense-

34
 contents do occur In his sense-history gives him a good
reason for believing that there are other conscious
beings besides himself. And thus we see that the
philosophical problem of 'our knowledge of other
people' is not the insoluble, and, indeed, fictitious,
problem of establishing by argument the existence
of entities which are altogether unobservable, but
is simply the problem of indicating the way In which
a certain type of hypothesis is empirically verified.
(LTL 171-2)

In the second edition he drew back from this confident

dismissal of the problem. Though still inclined to the
'behaviouristic' analysis, he confessed: 'I own that it has
an air of paradox which prevents me from being wholly
confident that it is true' (LTL 26). The air of paradox is not
hard to explain. What we are asked to believe is that
when, say, I describe someone as having a toothache, I am
speaking of his behaviour (i.e. of those of my sense-
contents into which his behaviour is to be analysed) and
nothing beyond that.
The air of paradox becomes even more pronounced if
we consider statements in the first person, as when I
describe myself SLS having a toothache. Such statements are
not made on the basis of sense-contents obtained from
observing one's ovm behaviour. But what, in view of this,
is the relation between the meanings of statements in the
first person and statements in the third person? The
difficulty was well described by Ayer in a later discussion.
On the \^ew in question, he wrote.

35
 the statements which I make about my feelings cannot
have the same meaning for any other person as they
have for me. [And therefore] if someone asks me
whether I am in pain and I answer that I am, my reply,
as / understand it, is not an answer to his question. For I
am reporting the occurrence of a certain feeling;
whereas, so far as he was concerned, his question
could only have been a question about my physical
condition. So also, if he says that my reply is false, he is
not strictly contradicting me: for ail that he can be
denying is that I exhibited the proper signs of pain, and
this is not what I asserted. (PK 214-15)

Turning to the problem again in his Central Questions of

1976, he favoured an answer that had been provided by
Hilary Putnam, which he expounded as follows (begin-
ning v^th a quotation from Putnam):

'Our acceptance of the proposition that others have

mental states is both analogous and disanalogous to
the acceptance of ordinary empirical theories on the
basis of explanatory induction.' The main point of
difference is that unlike empirical theories such as those
put forward by scientists the theory that other people
besides oneself have mental states is one that has no
serious rival; in this respect it is like the theory that
there are physical objects. What we establish induc-
tively, on the basis of our knowledge of our mental
states and our observation of other people's behaviour,
is a set of special hypotheses about their mental states:
the alternatives to these hypotheses are hypotheses

36
 which account for the same behaviour in terms of
different mental states rather than hypotheses which
deny to others any mental life at all. (CQ 134)

'None of this'/ he conceded, 'puts the sceptic out of court.'

If he is able to persuade himself that he has been cast

into a world in which only he is conscious, there is no
way in which he can be refuted. He will not even be at
a disadvantage in accounting for the appearances,
since he can believe that everything is and will
continue to be as if other people had minds, though in
reality they do not. (CQ 135)

To this argument he had no answer except to say that

'this is not a theory which I myself find it necessary or
useful to adopt'.
A more radical difficulty, however, is about the meaning
of the supposed 'theory' or 'set of special hypotheses',
when these involve states of affairs that are, by definition,
distinct from one's own sense-contents. According to
Ayer, I can 'establish inductively' that other people have
thoughts and feelings, including toothache. But what,
given the constraints of verificationism, would 'tooth-
ache' mean in this context? The difficulty is similar to that
which arose about our knowledge of physical objects as
distinct from sense-contents. As we saw earlier, Ayer
thought of these as being 'posited' by us, and in the
passage just quoted he compares 'the theory that other
people have mental states' with 'the theory that there are
physical objects'. But in both cases he seems to have gone

37
beyond the bounds of meaningful discourse as he himself
had defined it.

NECESSARY TRUTHS: MATHEMATICS

AND LOGIC

n objection commonly made against empiricism,

A observed Ayer, is that it cannot 'account for our
knowledge of necessary truths' (LTL 96). Here was a
category of statements which, though not empirical,
could not be written off as metaphysical nonsense: they
needed to be treated in a special way. He proposed to
show that such statements are 'devoid of factual content'
(LTL 105-6), since 'they merely call attention to linguistic
usages' (LTL 106). He first made this claim about what he
called 'analytic propositions'.

If I say, 'Nothing can be coloured in different ways at

the same time with respect to the same part of itself, I
am not saying anything about the properties of any
actual thing; but I am not talking nonsense. I am
expressing an analytic proposition, which records our
determination to call a colour expanse which differs in
quality from a neighbouring colour expanse a different
part of a given thing. In other words, I am simply
calling attention to the implications of a certain
linguistic usage. Similarly, in saying that if all Bretons
are Frenchmen, and all frenchmen Europeans, then all

38
 Bretons are Europeans, I am not describing any nnatter
of fact. But I am showing that in the statement that all
Bretons are Frechmen, and all Frenchmen Europeans,
the further statement that all Bretons are Europeans is
implicitly contained. And I am thereby indicating the
convention which governs our usage of the words 'if
and 'all'.
We see, then, that there is a sense in which analytic
propositions do give us new knowledge. They call
attention to linguistic usages, of which we might
otherwise not be conscious, and they reveal unsus-
pected implications in our assertions and beliefs.
But we can see also that there is a sense in which they
may be said to add nothing to our knowledge. For they
tell us only what we may be said to know already.
(LTL106)

A similar treatment could be given, he thought,

to mathematical statements.

Our knowledge that no observation can ever confute

the proposition '7 + 5 = 12' depends simply on the fact
that the symbolic expression '7 + 5' is synonymous
with '12', just as our knowledge that every oculist is an
eye-doctor depends on the fact that the symbol 'eye-
doctor' is synonymous with 'oculist'. And the same
explanation holds good for every other a priori truth.
(LTL113)

Now the statement that an oculist is an eye-doctor may

fairly be described as 'adding nothing to our knowledge'.

39
since the words are, indeed, synonymous. But can the
same be said about mathematical equations? 7 + 5' and
'12' are mathematically equivalent, but they cannot
properly be described as 'synonymous'. Douze is synonym-
ous with 'twelve', '7 + 5' is not. If '7 + 5 = 12' adds nothing
to our knowledge, it is not because the expressions are
synonymous, but because we are most unlikely to be
ignorant of the truth of this statement. But how would
this apply to less obvious mathematical truths? According
to Ayer, his treatment of '7 + 5 = 12' would also hold good
for such statements as '91 x 79 = 7189', even though here
we have to 'resort to calculation ... to assure ourselves that
[these expressions] are s)monymous' (LTL 114).
In the case of such statements as this, however, it
cannot even be claimed that they 'add nothing to our
knowledge' (let alone that the expressions in them 'are
synonymous'). While anyone acquainted with the mean-
ings of 'oculist' and 'eye-doctor' must be aware that an
oculist is an eye-doctor, it is quite possible for someone
acquainted with the meanings of '91', 'x', '79', etc. to be
ignorant of the product of these numbers. He may then
learn what the answer is (thus 'adding to his knowledge'),
either by being informed of it or by working it out for
himself.
Ayer was not unaware of this difficulty. He would need,
he said, to 'explain how a proposition which is empty of
all factual content can be true and useful and surprising'
(LTL 97). He thought he could do so as follows.

The power of logic and mathematics to surprise us

40
 depends, like their usefulness, on the limitations of our
reason. A being whose intellect was infinitely powerful
would take no interest in logic and mathematics. For
he would be able to see at a glance everything that his
definitions implied, and, accordingly, could never learn
anything from logical inference which he was not fully
conscious of already. But our intellects are not of this
order. It is only a minute proportion of the conse-
quences of our definitions that we are able to detect
at a glance. Even so simple a tautology as '91 x 79 =
7189' is beyond the scope of our immediate apprehen-
sion. (LTL 114)

Now it is true that our intellects are limited in the way

described by Ayer. But this seems to support the claim that
he was concerned to deny: namely, that mathematical
statements can add to our knowledge, and do not merely
'tell us only what we may be said to know already'.
The treatment of logic and mathematics in LTL, like
that of other topics I have discussed, continued to trouble
Ayer in later years. His vacillation on the subject was
recollected by Professor F.M. Quesada, who reported a
conversation between them as follows:

'Seventeen years have elapsed since in 1951, Freddie,

you attended the Lima Congress. Do you still think that
logic and mathematics are what you thought at that
time?' Ayer was silent for a few seconds ... Then, with
slow and hesitating words, he replied: 'Well,... f guess
I don't think now the same way

41
This conversation, held in 1968, was recalled by Quesada
when writing his contribution to The Philosophy of A.J.
Ayer in the late 1980s. Responding to that contribution,
Ayer confessed: '1 have not yet wholly freed myself from
[my] perplexity' on this subject." Nevertheless, having
added some fresh arguments, he concluded: 'I adhere to
the line that I took in Language, Truth and Logic all those
years ago.'^^ 'A fairly competent mathematician', he
observed, 'can tell at a glance that 37 + 37 = 74. He has to
use pen and paper to discover that 37 x 37 = 1369.'^^ Yet,
he argued, 'since multiplication is definable in terms of
addition, it is absurd to put these propositions into
separate categories'.
But from the point of view of the programme of LTL,
there is good reason for putting them into separate
categories. The essential question there, as we saw, is
whether mathematical statements can be informative -
whether they have 'factual content' or merely 'tell us what
we already know'. According to the empiricism of Ayer
and others, only empirical statements can be informative.
But the informative capacity of '37 x 37 = 1369' is not put
into doubt by the claim that multiplication is definable in
terms of addition. (It may be worth pointing out that
addition itself can be informative: this is so where the
sums to be added are large or many.)
A relevant distinction can be made by reference to
knowledge of meanings. Let us take '2 + 2 = 4' as an
example. Anyone who understands the meaning of this
sentence must recognize that what it says is true. But this
is not so when we turn to '37 x 37 = 1369': in thi§ case

42
knowledge of meaning does not entail knowledge that
the statement is true, and that is why the statement
can be informative. It is true that there is a way of arriving
at this information which is different from that of
arriving at empirical information, but this difference is not
captured by Ayer's account. (I have preferred '2 + 2 = 4' to
Ayer's '37 + 37 = 74', because recognition of the latter
may depend, as he puts it, on being 'a fairly competent
mathematician'.)

THE ANALYSIS OF 'STATEMENTS OF VALUE'

/ ' ^ h e r e is still', wrote Ayer in Chapter 6 of LTL, 'one

A objection to be met before we can claim to have
justified our view' - namely, that all genuine statements
of fact are analysable into statements about sense-con-
tents. 'It will be said that "statements of value" are
genuine [statements], but that they cannot with any show
of justice be represented' in this way (LTL 136).
In response to this challenge, he admitted that 'ethical
concepts' cannot be analysed in the approved way, but
immediately explained this in a way which would leave
intact his thesis about 'genuine statements of fact'. The
concepts in question, he explained, 'are mere pseudo-
concepts'.

The presence of an ethical symbol in a proposition

adds nothing to its factual content. Thus if I say to

43
 someone, 'You acted wrongly in stealing that money/1
am not stating anything more than if I had simply said,
'You stole that money/ In adding that this action is
wrong I am not making any further statement about it.
I am simply evincing my moral disapproval of it It is as
if I had said, 'You stole that money/ in a peculiar tone
of horror, or written it with the addition of some
special exclamation marks. The tone, or the exclama-
tion marks, adds nothing to the literal meaning of the
sentence. It merely serves to show that the expression
of it is attended by certain feelings in the speaker.
If now I generalize my previous statement and say,
'Stealing money is wrong,' I produce a sentence which
has no factual meaning - that is, expresses no
proposition which can be either true or false. It is as if I
had written 'Stealing money!!' - where the shape and
thickness of the exclamation marks show, by a suitable
convention, that a special sort of moral disapproval is
the feeling which is being expressed. It is clear that
there is nothing said here which can be true or false.
(LTL142)

He considered a difficulty that had been raised by G.E.

Moore. If ethical statements are neither true nor false,
how is it possible that people disagree about them? He
illustrated the difficulty as follows.

If a man said that thrift was a virtue, and another

replied that it was a vice, they would not ... be
disputing with one another. One would be saying that
he approved of thrift, and the other that he didn't; and

44
 there is no reason why both these statements should
not be true. (LTL 146)

Ayer's response was to admit, and indeed insist, that 'it is

impossible to dispute about questions of value'. What
moral arguments are really about are always 'the facts of
the case', and not the value that is to be put on them.

When someone disagrees with us about the moral

value of a certain action or type of action, we do
admittedly resort to argument in order to win him over
to our way of thinking. But we do not attempt to show
by our arguments that he has the 'wrong' ethical
feeling towards a situation whose nature he has
correctly apprehended. What we attempt to show is
that he is mistaken about the facts of the case. We
argue that he has misconceived the agent's motive: or
that he has misjudged the effects of the action, or its
probable effects in view of the agent's knowledge; or
that he has failed to take into account the special
circumstances in which the agent was placed ... But if
our opponent happens to have undergone a different
process of moral 'conditioning' from ourselves, so that,
even when he acknowledges all the facts> he still
disagrees with us about the moral value of the actions
under discussion, then we abandon the attempt to
convince him by argument. (LTL 146-7)

What would Ayer say about the endeavours of moral

philosophers? Should their work be consigned to the
waste-bin along with that of metaphysicians? No: there

45
was something to be done in moral philosophy, but it was
not much.

We find that ethical philosophy consists simply in

saying that ethical concepts are pseudo-concepts and
therefore unanalysable. The further task of describing
the different feelings that the different ethical terms are
used to express, and the different reactions that they
customarily provoke, is a task for the psychologist. (LTL
148)

Ayer-s account of moral statements (his talk of 'special

exclamation marks' and 'a peculiar tone of horror', etc.)
may strike readers as crude and perhaps even ridiculous;
his position became known as 'the boo-hooray theory'.
However, the view that moral statements are neither true
nor false, that they are expressions of feeling and not
statements of fact, has been widely held both inside and
outside philosophy, both before and since Ayer. (That's a
value-judgement' is sometimes thought sufficient to con-
sign a statement to the realm of personal feeling as
opposed to objective fact.)
Now it is clear that moral statements are different from
empirical ones; and if 'genuine factual statement' is
defined to mean 'empirical', then it will follow that moral
statements are not genuinely factual. But should this
definition be accepted? Is it a suitable way of marking the
difference between moral and empirical? In ordinary
language we commonly speak of truth and knowledge with
regard to moral statements. Thus a person who regrets
what he did may say 'I knew it was wrong' or 'It's true that

46
I ought to have done something else'. And in a court of
law it may be important to establish whether the defend-
ant knew he was doing wrong.
In this matter, as in others, Ayer was more sensitive in
later writings to the dangers of departing from the
ordinary uses of words. 'Certainly', he admitted in an
essay of 1949,

when someone characterizes an action by the use of an

ethical predicate, it is quite good usage to say that he is
thereby describing it; when someone wishes to assent
to an ethical verdict, it is perfectly legitimate for him to
say that it is true, or that it is a fact, just as, if he wished
to dissent from it, it would be perfectly legitimate for
him to say that it was false. We should know what he
meant and we should not consider that he was using
words in an unconventional way. What is unconven-
tional, rather, is the usage of the philosopher who tells
us that ethical statements are not really statements at
all but something else, ejaculations perhaps or com-
mands, and that they cannot be either true or false.
(Ml 231-2)

In view of these concessions, one might wonder why, as

stated at the start of his essay, he 'still wished to hold' the
view in question. In the following passage he gave his
reasons for doing so. The issue, as he recognizes, is of
wider interest and is, indeed, of fundamental importance
for philosophical inquiry.

When a philosopher asserts that something 'really' is

47
not what it really is, or 'really' is what it really is not,
that we do not, for example, 'really' see chairs and
tables, whereas there is a perfectly good and familiar
sense in which we really do ... it should not always be
assumed that he is merely making a mistake. Very often
what he is doing, although he may not know it, is to
recommend a new way of speaking, not just for
amusement, but because he thinks that the old, the
socially correct, way of speaking is logically misleading,
or that his own proposal brings out certain points more
clearly. Thus, in the present instance, it is no doubt
correct to say that the moralist does make statements,
and, what is more, statements of fact, statements of
ethical fact ... But when one considers how these
ethical statements are actually used, it may be found
that they function so very differently from other, types
of statement that it is advisable to put them into a
separate category altogether; either to say that they
are not to be counted as statements at all, or, if this
proves inconvenient, at least to say that they do not
express propositions, and consequently that there are
no ethical facts ... It is merely a matter of laying down
a usage of the words 'propositions' and 'fact', accord-
ing to which only propositions express facts and ethical
statements fall outside the class of propositions. This
may seem to be an arbitrary procedure, but I hope to
show that there are good reasons for adopting it. And
once these reasons are admitted the purely verbal
point is not of any great importance. If someone still
wishes to say that ethical statements are statements of

48
 fact, only it is a queer sort of fact, he is welcome to do
so. So long as he accepts our grounds for saying that
they are not statements of fact, it is simply a question
of how widely or loosely we want to use the word
'fact'. My own view is that it is preferable so to use it as
to exclude ethical judgements, but it must not be
inferred from this that I am treating them with
disrespect. The only relevant consideration is that of
clarity. (MJ 232-3)

The qualification 'although he may not know it', in the

second sentence, should be noted. What Ayer is saying is
that philosophers have sometimes failed to recognize the
true nature of their claims. The philosopher teUs us, say,
that we do not really see chairs and tables. Now this claim
is not empirical: it is not as if further observation might
reveal that we do, after all, see chairs and tables. The
philosopher's arguments, whatever they are, are meant to
rule out the logical possibility of our seeing chairs and
tables. But this, according to Ayer, amounts to 'laying
down a new usage of words'; so that statements that are
possible, and indeed quite common, in the ordinary usage
are not permissible in the new one. The philosopher's
claim may look like a statement of fact ('we do not really
see ...'), but what he is really doing is to 'recommend a
new way of speaking'.
It is in this sense that Ayer wishes his conclusion about
moral statements to be understood. What we must
consider, therefore, is not whether that conclusion is true,
but whether the new way of speaking, under which 'true'

49
and 'know' are improper, would be beneficial. (The
question is similar to that which arose about accepting the
criterion of verifiability as a 'methodological principle'.)
The implication at the end of the passage is that the
new way of speaking would have the advantage of clarity;
but is this really so? What is impprtant is to bring
out the differences between moral and empirical
statements (which Ayer proceeded to do by means of
a detailed example). Thus one might draw attention to
the difference between 'How do you know?' in the cases
of 'He took the money' and 'He acted wrongly'. But it
is not clear what could be achieved by introducing a
new way of talking which, in any case, could not displace
the existing one.

50
 CONCLUSION

A yer's work, as we have seen, is open to various

XXobjections, but this does not detract from its import-
ance. Ayer himself would not have been daunted by such
objections and perhaps he could have answered them in
ways 1 have not anticipated. But, in any case, he was
himself, as we have seen, constantly questioning and
revising his earlier ideas. Just as one must admire the
bravado of his early book, so one must be impressed,
when reading his later work, by his cautious and painstak-
ing treatment of the questions at issue, and his constant
striving to do justice to alternative views before arriving at
his own conclusion.
Ayer produced a vast amount of writing in his career,
only a fraction of which - though an important fraction -
could be presented here. A bibliography of his writings,
running to seventeen pages, includes books, articles and
letters to newspapers on many topics, both inside and
outside philosophy. His output, starting with a book
review in 1930, continued steadily until his death, and in
a sense even beyond that: the article 'What I saw when I
was dead' appeared in the Sunday Telegraph in August
1988! This was written after a 'near death' experience, but
his real death occurred some ten months later.

51
 ABBREVIATIONS
used to refer to works by Ayer

CQ The Central Question of Philosophy (Penguin,

Harmondsworth, 1976).

LTL Language, Truth and Logic (Penguin, Harmondsworth,

1971).
MJ 'On the analysis of moral judgements', in Philosophical
Essays (Macmillan, London, 1954).

ML The Meaning of Life and Other Essays (Weidenfeld and

Nicolson, London, 1990).

PK The Problem of Knowledge (Penguin, Harmondsworth,

1956).

52
 NOTES

I am grateful to Peter Hacker for comments on an earlier draft.

1. B. Gower (ed.), Logical Positivism (Groom, London, 1987),

p. 23.

2. L.E. Hahn (ed.), The Philosophy of A.}. Ayer (Open Court,

La Salle, Illinois, 1992), p. 18.

3. Ibid., p. 7.

4. Ibid., pp. 15-16.

5. The words 'sentence', 'proposition' and 'statement',

appearing in various combinations in Ayer's writings,
may be puzzling to readers not familiar with the literature.
There are philosophical reasons for distinguishing between
them, but I shall not go into these. On the whole I agree
with Ayer's view, in a later book, that 'statement' will do
for most purposes in expounding his views, and will
follow this usage where appropriate.

6. Hahn (ed.). The Philosophy of A J. Ayer, p. 302.

7. Hahn (ed.). The Philosophy ofAJ. Ayer, pp. 601-2.

8. Ibid., p. 598.

9. Ibid., p. 467.

10. Ibid., p. 478.

11. Ibid., p. 483.

12. Ibid., p. 481.

53
C:oN.sui;nN(; KUITORS:

R A Y M O N K A N D FRK.DERK: R A P I I A E I

AJ. Ayer 1910 - 1989

Ayer is best remembered for Lmignage, Truth mid Logic (1936),

which introduced British and American readers to the logical
positivism of the Vienna circle. Hanfling shows in this
introduction to Ayer's work how he turned this philosophy
into a form of British empiricism in the tradition of H u m e .

According to Ayer, philosophy is an activity oi malysis.

Metaphysical truths can be neither established nor refuted by
philosophical enquiry: they are meaningless. In support of this
claim, he deployed his 'principle of verifiability'. But he found
it difficult to refine the principle 'in such a way as to find a
middle ground between [an] over-strict requirement' which
would disqualify perfectly ordinary statements as meaningless,
and 'the over-indulgent licensing of gibberish' - including
that of metaphysics.

Oswald Hantling recently retired f r o m his philosophy

professorship at T h e O p e n University. His publications
include Logical Positivism, Essential Readings in Logical
Positivis7ii, Wittgenstein's Later Philosophy and The Quest
for Meaning.

Cover painting":
Swinging by VVassiiy
Kandin.sky, courtesy of the
T i t e Gallery