-

..... ..
r.
·.
� •. : t

. �::-......
• • • .¥ ••I

t
.. I· ._ • •
'!-· •••••-.
•• I
•1:..-A• , ...
i· - - .. '

·....
• I

�
._
...
..

.. .
t"'

:}�. �.-r---·.
•

·\
I I
-c:· :.•
l .
... :'c" •
I •

·?
·.::. )7JP? :> • .
.... ... . ...,
-�- l .. :-.
.-- .,.-... « ,
•. .. i=
�. ·. i=
■-•
.·: j(: .. .
:. I •

• '>.."
J. , • :: :·:• •••
:11-:
�- t"'
'

�
•• :·
i,,.,.... � .• .. � . , .........,..: .-
t' �: - _,.,-- ,

��.,, : .,. � ____ ....

. ..... �-<r-•·· .M"'.;·: ...r.-/ "... ..

r . /� �
IN CONVERSATION
W.V. Quine
 IN CONVERSATION
W.V. Quine

Published 1994 by

PHILOSOPHY INTERNATIONAL
Centre for the Philosophy of the Natural and Social Sciences
The London School of Economics and Political Science
Houghton Street
London WC2A 2AE
England

Copyright© 1994 PHILOSOPHY INTERNATIONAL

 CONTENTS

PREFACE

by Rudolf Fara, Executive Director,

Philosophy in Britain, CPNSS, London School
of Economics ill

THE CONVERSATIONS

The Fara Interview

Participants: W.V. Quine, Rudolf Fara 1
The Dennett Panel
Participants: W.V. Quine, Daniel Dennett,
Martin Davies, Paul Horwich, Rudolf Fara chair 3
·The Goldfarb Panel
Participants: W.V. Quine, Warren Goldfarb,
Martin Davies, Paul Horwich, Rudolf Fara chair 9
The Fogelin Panel
Participants: W.V. Quine, Robert Fogelin,
Martin Davies, Paul Horwich, Rudolf Fara chair 14

The Boolos Panel

Participants: W. V. Quine, George Boolos,
Martin Davies, Paul Horwich, Rudolf Fara chair 18
The Dreben Panel
Participants: W. V. Quine, Burton Dreben,
George Boolos, Martin Davies, Rudolf Fara chair 23

The Block Panel

Participants: W.V. Quine, Ned Block,
Martin Davies, Paul Horwich, Rudolf Fara chair 28

ii
 PREFACE

The purpose of this short volume is to provide a hard-copy summary of

the topics covered in the interview and the six panel discussions with
Professor Quine. The verbatim transcripts of the programs numbered
nearly 300 pages of typescript and it was felt that the publication of such
a document was not only redundant with the spoken commentaries, but
also that the needs of the lecturer, researcher or student would be better
served by an explicatory synopsis. This would provide a ready
reference to the content of each program, and could be photocopied for
use as handouts or as a course outline when the series is used as a lecture
supplement.

I cannot acknowledge enough my debt to the participants in the

discussions, nor to those who contributed directly and indirectly to the
project. It is hoped that the aria from I Pagliacci (a favourite of
Professor Quine's) accompanying the acknowledgments on the program
will provide sufficient aural inducement for the viewer to pay a little
more attention than is customary to the credits. With particular regard
to these summaries, Michael Fara wrote them and deserves every credit
for his exegesis and his hard work. To Professor Quine, who not only
checked the summaries and gave invaluable help in their preparation,
but who also inspired the project from inception to completion with his
encouragement, enthusiasm and erudition, I can offer no adequate
thanks. I wish only that the series will be a fitting tribute to his work.

While acknowledging my deep gratitude to all concerned, I take full

responsibility for all aspects of this project. My decision on what would
be presented in the end was final, and the blame for any shortcomings is
entirely mine.

R. V. Fara
Philosophy in Britain
CPNSS
London School of Economics
1994

iii
 THE FARA INTERVIEW

Following a brief biographical introduction, Rudolf Fara opens this

personal, scene-setting interview with questions concerning Professor
Quine's childhood experiences and his relationship with his family.
Quine discusses his early involvement in clubs and societies, familial
influences and his reading interests. Quine's relationship with his
mother was closer than with his father, although he recalls no real
tension in the family. He concludes that he certainly had a happy
childhood.

Fara proceeds to question Quine about his earliest philosophical

thoughts. For Quine these are scepticism of religion and (although he
did not realize it at the time) the beginnings of his doctrine of
extensionality. He discusses how bis fascination with travel relates to his
philosophy, explaining how philosophical reductionism and
foundationalism are similar to "the strange way home" encountered
when travelling. Quine explains that he sees the role of philosophy as
being to ascertain the basis of reality, the nature of the universe. He
disagrees with Russell, who claimed that the essential ideas of the great
philosophers were simple, and he cites Kant as a counterexample. Quine
outlines why he views epistemology ( of ontology on the one hand, and
of mathematical sciences on the other) as being the central philosophical
domain.

Quine discusses how he was faced with a choice of philosophy,

mathematics and linguistics when he went to Oberlin College, and why
he decided to major in mathematics and mathematical logic. Quine chose
Harvard for his graduate studies because Whitehead was there, and he
had great admiration for Principia Mathematica. Quine describes
meeting Russell through Whitehead at Harvard, and he views this as his
most memorable experience as a graduate student there. He states who
influenced him most as a student, and briefly discusses his acquaintance
with B.F. Skinner, who was a contemporary junior fellow at Harvard.

The discussion turns to Quine's experiences in Vienna at the time

of the logical positivists, and he describes the leading figures of that
movement. He outlines the central tenets of logical positivism, noting
that this name was only introduced later, in a paper [1931] by Herbert
Feigl and Albert Blumberg. After Vienna, Quine travelled to Prague,
and he describes his meetings there with Carnap. After Prague came
Warsaw, where Quine met the great Polish logicians-Lesniewski,
Lukasiewicz and Tarski-and he recalls that Tarski impressed him the
most. Quine describes how his approach to logical positivism differed

1
 from that of A. J. Ayer; Quine did not have the sense of destroying a
movement, rather he saw himself as a revisionist.
There follows some comment on Quine's views today. He still
agrees with some of the doctrines of positivism, particularly the
"linguistic tum"-the notion of talking about words rather than ideas­
although he dislikes the positivists' insistence on drawing strict
boundaries within science, between analytic and synthetic truths, and
· between sense and non-sense. He sees extensionalism and naturalism as
being the two doctrines most central to his philosophy, and he briefly
describes what these are.

This program concludes with some questions about Quine as a

person: what food he likes and dislikes, what music he favours, and
what vices he has given into. He states which achievement he is most
proud of, and describes his most embarrassing experience. He names
Aristotle as the philosopher whom he admires most, due to the fact that
he achieved so much on the basis of so little. Quine gives advice to
anyone thinking of studying philosophy for the first time, and ponders
why he enjoys teaching more than being taught. Active versus passive?

Finally, Quine asks the question he would most like answered:

what, in words which we can define and understand now, would be the
briefest intelligible encapsulation of the whole of reality, the nature of
the universe?

2
 THE.DENNETT PANEL

A. Quine's Early Philosophy

1. Quine discusses his opinion of the logical positivists. He agrees

that most of speculative metaphysics should be abandoned, but he
disagrees with the positivists' recognition of a sharp boundary between
metaphysics and natural science. He views philosophy as a part of
science; and the more speculative the question, the more towards the
philosophical end of science it lies. Most of speculative. metaphysics is so
unclear that it makes no scientific sense at all, and so should be rejected.

2. Quine outlines the dogmas that he was attacking in "Two Dogmas

of Empiricism". These are the acceptance of an analytic/synthetic
distinction, a distinction between what is true by virtue of meaning and
what is true by virtue of empirical evidence; and the favouring of
phenomenalistic reductionism, the view that all empirical concepts can
be reduced by definition to sensory input. Quine was arguing for a
replacement of this view with holism. Quine emphasizes that when he
was writing "Two Dogmas..." he was not persuaded that he had
disproved the possibility of a satisfactory definition of analyticity, he
was only maintaining that such a definition had never been given, and
was unlikely ever to be given.

3. Quine has often been considered a pragmatist, and Paul Horwich

inquires into the nature of this pragmatism. Quine responds that he has
never viewed himself as a pragmatist, and he has never even been clear
what is meant by pragmatism. He declares himself to be an empiricist.

B. Scepticism about Meaning

1. Horwich asks Quine how he can reconcile his scepticism about

meaning with his evident fascination with language and the meanings of
words. Quine argues that he was challenging the notion of sameness of
meaning, a rejection of which would ent�il a rejection of the notion of
the meaning of an expression (because, he argues, there is no entity
without identity). He was not questioning our study of the meaning of
language and the meaning of sentences (using "meaning'' as a mass term,
like "sugar"). This is what lexicographers are engaged m, and
lexicography is a worthwhile science.

2. Daniel Dennett outlines Quine' s thought-experiment of radical

translation: imagine a community, completely isolated from ours, which

3
speaks some alien language, and imagine sending two linguists who
independently enter that community and learn the language, returning
with two manuals of translation. Quine's claim is that both of these
manuals will translate the alien language into our language, but that they
will differ in non-trivial ways; and furthermore, there will be no fact of
the matter about which manual is the right one. Dennett goes on to
challenge Quine to offer an example, even an imaginary one, of real
radical translation failure. Quine responds that it is too big a job to
actually reduce a language to a manual of translation, but he defends the
plausibility of the experiment by considering what would constitute a
test of a faithful manual of translation. Quine sees smoothness of
dialogue and success in negotiations when using the language as the only
possible sorts of test, and he asks us to consider two manuals of
translation which equally pass these tests, and furthermore are equally
simple in their translations. Then, he argues, given some monologue in
the alien language, each manual will provide a smooth translation of it.
In order to show that the manuals are different, we cannot simply
compare the two translations, since such a comparison would be merely
intuitive, and would depend on the notion of synonymy. Instead, Quine
proposes a formalized test for their difference: translate the first
sentence of the monologue using one manual, the second sentence using
the other manual, and· continue alternating in this way until the
monologue is translated. Now, if the manuals are non-trivially different,
the resulting translation will fail the test of smoothness, indeed even
coherence.

3. Quine clarifies the nature of his scepticism about meaning,

claiming that the notion of meaning is understandable and useful, but
that it cannot play a role in serious science. It cannot even be used in
theoretical linguistics, although the word is used in practical
lexicography. Martin Davies proposes that, given the fact that Quine
considers philosophy as a part of science, his scepticism about meaning
would make it impossible to ask the characteristically philosophical
question "what do you mean by ... ?". Quine does not view this as a
problem, since he believes it will always be possible to paraphrase such
questions without
. appealing to the notion of meaning-e.g. "in what
way are you usmg the word. . . ?"
. .

4. Quine isolates his linguistic behaviourism as being the main

premise which one would have to deny in order to reject his scepticism
about meaning-we learn language only by observing linguistic
behaviour. Quine accepts Chomsky's proposal of an innate linguistic
readiness, but this does not affect his insistence upon linguistic
behaviourism. He emphasizes that this cominitment does not entail an
adherence to other behaviourist doctrines-e.g. the behaviourism of
Skinner does not concern Quine's philosophy. He sees himself as a

4
behaviourist in the sense that Wittgenstein is a behaviourist, although he
holds that Wittgenstein's doctrine of meaning as use merely presents the
problem, it does not solve it.

5. Dennett quotes Quine as having said "meaning is what essence

becomes when it is divorced from the object of reference and wedded to
the word", and concludes that the challenge to meaning is also a
challenge to the very idea of essence. Quine wholly agrees with this
conclusion, and views "essence" to be interchangeable with "extreme
importance''-if a scientist claims that a trait of an object is essential to
it, he is really claiming that that trait is more important than other traits
of that object.

C. Epistemology

1. Quine gives a characterization of his naturalized epistemology,

placing it at the philosophical end of the spectrum which he calls
science. It is natralistic in the sense that it accepts the scientific finding
that all our data comes to us through our senses, yet it retains a
traditional philosophical approach by being concerned with how it is
possible, on the basis of a.relatively tiny amount of sensory data, that we
manage to arrive at such profound and predictively effective scientific
results. Martin Davies suggests that it might be possible for naturalized
epistemology to coexisf with traditional epistemology. Quine doubts the
demarcation between true belief (other than knowledge) and knowledge
itself, and does not accept knowledge as a legitimate concept. His theory
of knowledge has as one of its findings that there is no such thing as
knowledge, but only true belief warranted in varying degrees.

2. Quine views naturalism as being helpful in dealing with sceptical

questions, and he explains how this is so. With a question such as "how
do I know that I am not dreaming" Quine can only understand the term
"dreaming" relative to the sort of tests that would induce or reinforce
his belief in applying the term. Although it is possible that I am
dreaming, there is a great deal of evidence to suggest that I am not­
systematic memories, clarity of thought, etc. As a consequence of this
evidence, it is very implausible that I am dreaming, and for a naturalist
this is enough to refute the sceptic-nothing is absolutely certain in
science, so one does not require absolute certainty that one is not
dreaming.

3. Quine considers what truth is, and follows Tarski in this regard.
The sentence "snow is white'' is true if and only if snow is white; and if
there is a sentence to which we cannot apply this truth criterion, Quine' s
approach is to blame the sentence, not the concept of truth. He sees one

5
very important trait of truth as being its idealism: our usage is not to
say that truth changes with the progress of science, but rather that our
conception of what is true changes with the progress of science. Davies
criticizes Quine for only considering scientific truth, but Quine replies
that this is the only sort of truth there is-however, his naturalism
allows for a very broad conception of what counts as a science.

D. Ontology

1. Quine is questioned about his doctrine that "to be is to be the

value of a bound variable". Quine completely disagrees with Gilbert
Ryle's notion of different categories, different senses of existence; he
views reality as a scientific term, and the question of what there is as a
part of science, and his doctrine is a way of formalizing this view.

2. Quine explains his thesis of the indeterminacy of reference as

follows. Consider any one-one reinterpretation of everything that there
is; for example, take the "cosmic complement"-i.e. interpret each
object as being the whole universe minus that object. Then reinterpret
all our predicates as being true of precisely the cosmic complements of
everything that they were true of before the reinterpretation. In this
way, it does not matter what objects are being referred to by our
theory, it does not matter what ontology we assume. Quine then raises
the question: what is the point of an ontology at all? His answer is that it
introduces logical structure to the scientific theory, and such structure is
needed in order that the theory imply observation sentences. He adds
that the indeterminacy of reference is trivial and incontestable, and that
the particulars of our ontology, over and above the logical structure that
it imposes on our science, need never have concerned us.

3. Paul Horwich objects that the indeterminacy of reference could be

seen as trivially false as opposed to trivially true, on the grounds that it
is trivially true that "Quine" refers to Quine just as "snow is white" is
true if and only if snow is white. Quine sees this as a misunderstanding
of his position. Given any ontology, the sentence '"Quine' refers to
Quine" is determinately true, but the nature of the Quine that is being
referred to by "Quine" is indeterminate.

E. Contextualization of Quine's Philosophy

1. Martin Davies asks Quine to place his philosophy on the

realist/anti-realist spectrum. Quine sees himself as something of a naive
realist, allowing science to dictate what reality is. However, he has to
reconcile this with the fact that he views the notion of object as

6
spurious. He believes that what saves his position from being pure
idealism is the fact that we form our scientific theory on the basis of
sensory input. Davies questions Quine about the thesis of epistemic
boundedness, or cognitive closure, the view that some important facts
about reality are constitutionally beyond our grasp because of the way
we have evolved. Quine sees himself as hemmed in by his naturalism in
this regard, in the sense that there may be things that are beyond our
grasp, but this would be due to a deficiency in our scientific efforts, not
a deficiency in the makeup of our minds.

2. Quine explains the relationship between his philosophical ideas

and his work in mathematical logic. It was a philosophical drive that led
him to work in mathematical logic, a wish to systematize mathematics.
He also sees a great contribution to philosophy by mathematical logic in
clarifying and simplifying philosophical questions. Quine defends the
study of mathematical logic as being worthwhile by noting that his (and
others') work in the subject has contributed greatly to the development
of computer theory, which in tum has revolutionized technology.

3. Quine outlines his naturalistic account of moral philosophy. He

holds that moral judgements are laid down partly by society and partly
by evolution. Attempts have been made (for example by Kant) to
systematize our moral judgements, but Quine sees such attempts as
hopeless. His naturalism consists in using scientific evidence to persuade
people to act in particular ways-given the evidence, it is likely that
doing A will cause E, and doing B will cause F, and F is the preferable
outcome, so it is best to do B. If two people differ over what they judge
to be the preferable outcome of an action, then there is nothing that
naturalism can do to help, and this why Quine does not view morality as
part of science. He sees a fundamental difference between epistemic
norms and value norms, since there are no experimental checkpoints in
the realm of moral decision making.

4. Given that Quine views philosophy as a subclass of science,

Martin Davies asks for criteria for what would count as a philosophical
scientific question, as distinct from other scientific questions. Quine
gives an example in ontology. The question of what there is, whether
there are or are not unicorns, for example, is a strictly scientific
question. The decision of what ontology we should assume, made on the
grounds of simplicity, is again strictly scientific, until we get to the
point where the ontology is simple enough for scientific purposes. If we
then continue to further simplify our ontology, this must be the work of
philosophy. Quine' s view is that one is a philosopher by virtue of one's
motivation.

7
5. Quine disagrees with Wittgenstein's thesis that all philosophical
questions are pseudo-questions resulting from confusions growing out
of language. He agrees that many questions are pseudo-questions, but
not all are; he gives Russell's theory of descriptions as an example of a
genuine philosophical question-how can we talk about things not
existing when there is nothing there to exist? Quine sees this as a
genuine question because it aids the philosophical project of
clarification. He also disagrees with Wittgenstein's proposed treatment
of pseudo-problems, which is to stay within ordinary language. Quine
prefers the scientific custom of introducing new symbols where (but
only where) they may simplify the problem, and as long as these
symbols are well-defined there is nothing wrong with this approach. He
views the ordinary language philosophers as missing a great deal of
serious and worthwhile philosophy.

F. Philosophy Today

1. Daniel Dennett suggests that American philosophers are generally

sympathetic to Quine's scientific philosophy, whereas their English
counterparts tend to favour the ordinary language approach, and he asks
Quine for his views on this point. Quine sees the opposition in England
as being much less than it was when he was first exposed to it, and notes
that it is not only ordinary language philosophy that he is opposed to. He
is opposed to anyone, including Carnap, who recognizes a sharp
cleavage between philosophy and science.

2.. Quine discusses how his opinion of propositional attitudes has

changed since he decried them in Word and Object. Now he accepts the
propositional attitudes de dicto but not those de re-i.e. he accepts those
of the form "A believes that p", but not those of the form ''there is an x
such that A believes that Fx". Quine still feels that the verb "believes" is
vague-what are the criteria for believing?--but in logical terms the
propositional attitudes de dicto are compatible with the rest of science.

3. Quine appraises the current scene in philosophy, and on the whole

finds it very encouraging, especially in view of the fact that more
philosophers (he gives Daniel Dennett as an example) are
communicating in detail with practical scientists. He is disappointed with
the emergence of deconstructionism, but he believes that the trend will
not survive for long. Quine concludes this program by considering
th-<1se problems he would choose to work on if he were just beginning
his philosophical career.

8
 GOLDFARB PANEL

A. Logical Positivism

1. Quine talks broadly about the status of logic at Harvard before he

went to Vienna in 1932. Although Harvard prided itself on being the
greatest centre for logic in America, he found that he was not learning
any there. The three logicians at Harvard at that time were Whitehead,
Lewis and Sheffer. Whitehead had stopped working in logic, and was
concentrating on metaphysics; Lewis was not teaching.logic; and Quine
viewed Sheffer's courses in logic as altogether too thin. Quine goes on
to talk about the origins of his differences with Carnap over the latter's
program of logical syntax.

2. Quine discusses the positivist doctrine that metaphysics is

meaningless. He agrees with them that much of metaphysics is
unverifiable, but he does not see this as implying that it is meaningless.
For example, most propositions within the social sciences, and even
some propositions in rigorous physics are unverifiable, but they are
useful and helpful to our overall scientific theory, and hence it would be
wrong to regard them as meaningless. Quine is against the positivists'
recognition of a strict demarcation between metaphysics and science.
Martin Davies notes that there are also unverifiable sentences in the
outer reaches of mathematics, and he asks Quine about the difference
between those and other unverifiable sentences, such as "there is a planet
for whose existence we shall never have any evidence". Quine sees the
difference to lie in the structure of connection with other parts of
science: higher mathematical sentences are a clear extrapolation of the
grammar and lexicon of science, whereas other unverifiable sentences
are only vaguely related to verifiable sentences, but the relation is still
strong enough for many of them to be included in our web of belief.

3. · Paul Horwich asks for Quine's reaction to the kind of examples of

meaningless sentences offered by Wittgenstein, "it is five o'clock on the
sun". Quine sees it as characteristic of this kind of example that we can
prove on the basis of accepted scientific principles that there could not
be any criterion by which to decide whether or not we should accept the
sentence within our scientific theory, and this is enough for it to qualify
as meaningless.
4. Quine discusses the impact that Wittgenstein's Tractatus had on
him when it came out, and he concludes that it did not have any positive
effect. He disliked the absoluteness of Wittgenstein's atomism. He also
viewed Wittgenstein's treatment of number as untidy, and clearly

9
inferior to Russell's treatment. He was more sympathetic to
Wittgenstein's later work.

5. Quine considers the relationship between his rejection of an

analytic/synthetic distinction and of phenomenalist reductionism and his
introduction of his holistic "web of belief'. He believes that it might
have been people's failure to reject reductionism, failure to accept
holism, that led them to accept analyticity as uncritically as they do. He
views holism (i.e. anti-reductionism) and scepticism about analyticity as
almost identical. However, he accepts definitions of analyticity which
are adequate to its commonsense extra-philosophical applications, and he
gives such a definition in his The Roots of Reference.

6. Quine discusses A. J. Ayer's relationship with logical positivism,

and he sees Ayer's book Language, Truth and Logic as lying between
the two phases of Carnap which Quine had viewed as abruptly different:
the Carnap of the Logischer Aufbau, where he was taking a
phenomenalist approach; and the later Carnap, where he had
accommodated naturalism within his physicalism. Ayer' s book eases this
transition of Carnap, and successfully reproduces the spirit of the
Vienna Circle in a clear and non-technical manner. Quine views the
book as being important for the ''history and geography" of philosophy.

B. Naturalism

I. Quine' s formulation of his doctrine of naturalized epistemology

was intended as a way of insuring the removal of transcendental
intrusions into philosophy,' either in the way of the possibility of a
priori knowledge, or of divine revelation. It achieves this by
maintaining that the only road to knowledge is through experimental
checkpoints in observation. However, Quine emphasizes that he was not
the creator of naturalism itself-rather he views most traditional
philosophers, and in particular Hume; as being naturalists. Quine says
that he would find it impossible to accept a non-naturalist account of
knowledge, since he would not be able to understand what kind of
evidence would support such an account.

2. Quine feels that he cannot distinguish between externalized

empiricism and naturalism. He introduced the term "epistemology
naturalized" specifically to contrast with Carnap's manufacturing of a
phenomenal language in which science was not expressed, and then
deriving a language about the physical world from that. Quirie's
approach is to view language directly from the standpoint of the
physical world, the already accepted ontology. Quine has written that
empiricism is itself an empirical doctrine, and Paul Horwich suggests

10
that this forces him into considering it conceivable that empiricism
might have to be abandoned if, for example, we found very strong
evidence for clairvoyance or telepathy. Quine modifies his view by
saying that we accept empiricism because observational evidence is our
only road to knowledge, and in the case of telepathy or clairvoyance we
would have to accept that such evidence is not restricted to input
through the five senses that we are presently aware of. He concludes that
he may have been wrong to consider empiricism as itself an empirical
doctrine, unless we construe empiricism as limiting us to the five senses.

3. Quine sees his naturalistic philosophy as continuous with science.

He believes that there are other forms of philosophy, for example moral
philosophy, which, although being worthwhile studies, cannot be part of
naturalistic philosophy. He views philosophy in general as a scattered
aggregate. Martin Davies questions Quine about his opinion of
Strawson's conception of naturalistic philosophy, the aim of which is to
elucidate concepts rather than to detemline what we can know about the
world. Quine views Strawson' s conception as certainly part of
philosophy, but as too narrow for his liking-there is much of
philosophy that is not analysis of concepts. However, he sees Strawson's
project of plotting a conceptual scheme as distinctly philosophical, since
it considers the conceptual structure of the whole way in which we view
the world. Strawson sees the question of which conception of philosophy
to choose as being decided on the grounds of individual temperament.
Quine does not like this notion, since he sees the two conceptions as
being such distinct projects that it does not make sense to choose
between them, any more than it makes sense to choose poetry rather
than science; we want both.
4. Quine stresses that despite his naturalistic insistence upon
observational checkpoints as being the only checks on know ledge, this
does not mean that everything worthwhile in science has to relate
somehow to such checkpoints. There are many scientific hypotheses
which have no possibility, even indirectly, of being submitted to
observational tests, and yet they point the way to further hypotheses
which are testable and are useful to science. He feels that science would
be impoverished without such hypotheses, and he holds this as being one
of his main criticisms of logical positivism. His naturalism consists in
his insistence that there are no checkpoints that are not observational,
not that everything has to be subject to checkpoints. Quine also mentions
that his naturalism does not entail pragmatism-he has never considered
himself a pragmatist, and is not even clear what the term means.

11
C. Extensionalism and Ontological Relativity

1. Quine explains the doctrine of extensionalism. In an extensional

language any two predicates, if they are true of precisely the same
objects, are interchangeable without disturbing the truth values of the
sentences in which they occur. Therefore, in an extensional language,
there is no difference between a property and a class, and so Quine does
not recognize properties. One important consequence of this is that
extensional languages retain the substitutivity. of identity-if two
singular terms are names of the same thing and nothing else then you
can substitute one for the other in any context, without disturbing truth
values. Quine notes that it is hard to make sense of that law failing, but
that if the language is not extensional then it does seen1 continually to be
threatened. He adds that predicates that are true of the same thing may
or may not have the same meaning, so sameness of meaning is a
distinction that goes beyond sameness of extension; this is how Quine' s
extensionalism relates to his scepticism about meaning.

2. Quine does not recognize properties, but Martin Davi�s suggests

that in the fully austere ontology, which consists only of classes of
quadruples of real numbers, neither does he recognize tables and
mountains. Quine agrees with this, but he does not see this as
problematic. From a theoretical point of view it does not matter what
our ontology is (up to isomorphism), so we may as well choose the
simplest one; however, in practice it is more convenient to talk abc;mt
tables and mountains than it is to talk about classes of quadruples of real
numbers, so we choose that ontology instead. However, Quine sees it to
be positively misleading to talk about properties, precisely because of
the problem of individuation-he can see no way of providing a clear
explanation of when to say of property A and property B that they are
the same property. Nevertheless, he accepts the possibility that science
might one day find a need for properties, a need that is more vital than
the need for a criterion of individuation, in which case he would allow
properties.

3. Davies suggests a possible need for properties arising from a

causal explanation of events. Such an explanation usually entails a claim
that some properties of an event are causally efficacious, while others
are not. Quine denies this entailment, and uses an example of Davidson's
to justify this denial. Imagine a machine which is rotating and heating
up, and a nearby piece of jce melts. We would want to say that the
heating up caused the ice to melt, but Quine' s repudiation of properties
disallows this explanation. To accommodate thls, Quine views the
rotating and heating as a single event, and it is this event that causes the
ice to melt. This does not affect our general law that nearby heating
melts ice, and it does not suggest a law that nearby rotating melts ice-

12
 FOGELIN PANEL

A. Naturalized Epistemology

1. Quine views his epistemology as a chapter of psychology, since it

accounts for the fact that we have such immensely successful scientific
theory based on (relatively) insignificant sensory input.

2. Naturalized epistemology is at the philosophical end of what

Quine sees as the continuum of science be,cause of the crucial role
played in it by the theory of scientific verification, which is traditionally
viewed as normative philosophy.
3. One implication of naturalized epistemology is that philosophers
are very much involved in reportage about other areas of science. Quine
sees no problem with this, and notes that this is exactly the sort of work
carried out by many philosophers; Quine mentions Daniel Dennett as a
notable example.

B. The Web of Belief

l. Quine discusses why holism is central to the web of belief

metaphor. It is a web of logically interlocking scientific theories, with
observation-sentences at the periphery.
2. How the web evolves over time: testing hypotheses by means of
observational tests implied by bodies of theory with critical mass. By
removing hypotheses that fail such tests, we also have to remove any
hypotheses which imply the one we are rejecting, and this will cause
reverberation throughout the web. This procedure of testing and
removing hypotheses is a heuristic norm, not part of the web itself.
3. Why the notion of degree of belief is vital to Quine' s
epistemology. He agrees with Paul Horwich that the web metaphor is
merely compatible with this notion and does not imply it.
4. Quine does not see it as essential to his holism that the web be an
interconnected whole; rather the web is a disjointed set of tenuously
linked and internally dense portions. What is important for Quine is that
all of the web is subject to the same logic.
5. The core of the web comprises mathematics and elementary logic.
Quine explains how this is consistent with his thesis of revisability, the

14
principle that anything in our theory can be changed, although he admits
that he now views this thesis as "at best an uninteresting legalism".
Carnap shared revisability with Quine, but he did not follow through
with the implication that we should question the notion of meaning and
analyticity.

6. Why Quine does not see a need for a bifurcation within the web
between hypotheses and guicling principles (the principle of simplicity,
basic laws of logic, etc.). Each of these principles will have a
representative sentence within the web. For example, the principle of
simplicity will have the sentence: "The chances of getting a hypothesis
that is worth testing are increased by: considerations of simplicity".

C. Scepticism

1. Quine's response to general sceptical challenges: how they can be

resolved by scientific means. An infinite regress of justification is cut
off by the experimental test of a critical mass of theory; this is a
straightforward form of foundationalism.
2. Quine' s response to the "brain in a vat" problem, in terms of
naturalistic plausibility. We trust science which tells us that it is
implausible that we are brains in vats. It is possible that we are brains in
vats, but admitting such a possibility does not threaten our trust in
science precisely because we recognize the fallibility of science.

3. Barry Stroud has argued that Quine's naturalized epistemology is

not sufficient to answer the following sceptical challenge. Given
naturalized epistemology, we are not concerned with knowledge as
classically construed, but rather with degrees of plausibility. If we were
dreaming, and formed a dreamt belief of something true, there would
be nothing plausible about such a belief. Therefore, for anything to be
plausible, we must first rule ·out the dreaming hypothesis, which we
cannot do by means of observational testing, since, with the hypothesis
still active we could not trust the results of the test. Quine' s response is
to allow the dreaming hypothesis into his theory, but simply dismiss it
as being implausible, for lac-k of sufficiently overwhelming
incoherences in our experience.
4. In distinguishing between fanaticism and scepticism, and feigned
doubt and genuine concern, Quine employs psychological judgement:
there would be manifestations which we would expect in one but not in
the other.

15
5. In viewing Davidson's principle 9f charity as a response to
scepticism, Quine pref ers a principle of psychological likelihood of
being believed by the subject in his observed circumstances. This bars
Quine from Davidson's response to scepticism, but he feels that he does
not need one given his overall naturalized epistemology.

6. Underdetermination of theory by data. Quine explains how to

decide between two equally simple competing theories which make the
same predictions. He distinguishes three cases: (i) the theories are
intertranslatable and compatible; (ii) they are incompatible; (iii) they
are compatible but not intertranslatable. In the case of (i) they can be
merged. Davidson has suggested a way of reducing case (ii) to case (iii),
by recasting some shared words of the conflicting sentences as distinct
words. Quine shelves case (iii) by appealing to untranslatability. Truth
applies to sentences, sentences are language-bound, and so truth and
falsity of foreign sentences makes sense only via translation, which is
lacking here.

D. Induction

1. Quine explains why he has written that "the Humean predicament

is the human predicament". A logical justification of induction is
impossible, but an explanation of it is important from a naturalistic
point of view. Unlike Strawson, Quine refuses to give a transcendental
argument. He accounts by natural selection for the good performance of
induction up to now.

2. Quine outlines his belief that an explanation of induction hinges

on the notion of subjective similarity; he explains how this standard can
be elicited using conditioned responses. He concludes that subjective
similarity makes scientific sense, unlike the abstract notion of absolute
or objective similarity.
3. Quine's moral philosophy also uses the notion of subjective
similarity. In spite of this he justifies his claim that the way we acquire
moral values is different from the way we practice science. Science
proceeds by induction and refutation, and its findings then figure in
moral decisions.

E. Epistemic Norms
1. For Quine, the epistemic norms are engineering principles, just
like the norms in building bridges. They are applied cognitive

16
judgements, learned partly by trial and error and partly by reasoning on
the basis of hypotheses which have been checked by experiment.
2. Quine admits that this view forces him to accept that reasoning
which we now regard as fallacious might have to be regarded as valid in
the future, due to a change of consequences of that reasoning, and hence
a forced change of norms.
3. Quine emphasizes an important difference between epistemic and
moral norms. Statements about what we ought to think are derivative
from sufficiently established scientific hypotheses; statements about what
we ought to do have no corresponding cognitive judgements.

4. Quine gives an example of an epistemic norm: 'We ought to think

that the future will resemble the past". This is a cognitive judgement
sustained by induction, and ultimately indeed by natural selection,
although of course it is not a justification of induction.

F. Knowledge
1. The basis of naturalism is the belief that science is the only road
to knowledge. How then is knowledge itself defined, not to render this a
trivial circularity? Knowledge is warranted true belief, with certain
exceptions, and science does the warranting but cannot guarantee the
truth.

2. Science's warrant itself commonly stops short of experimentation,

Quine holds, and leaves off with the heuristics of analogy, simplicity,
symmetry, and the like. Knowledge should be seen as grading off
correspondingly, but this is contrary to the connotation of certainty that
the word bears. This is perhaps why it seems odd to speak of knowledge
about non-scientific matters.

3. Martin Davies persuades Quine that he should discard his norm of

"don't form beliefs unless they are arrived at by the scientific method",
since we form beliefs which are empirical and have testable
consequences, but are not arrived at by the scientific method.
4. It is part of the concept of naturalism that transcendental
arguments can never be used; naturalism is a strictly empirical doctrine.

17
 BOOLOS PANEL

A. "Two Dogmas of Empiricism"

1. Quine clarifies his scepticism about meaning given in "Two

Dogmas... ". He was trying to abandon the notion of a proposition being
the meaning of a sentence, with an independent abstract existence. It was
in order to achieve this that he introduced the indeterminacy of radical
translation, destroying the notion of synonomy between languages. He
also argues that, even within the same language, there is no objective
behavioural criterion of sameness of meaning, so the notion should be
dispensed with altogether. He does not view the expression "has the
same meaning as" as meaningless, but simply as not based on a
fundamental notion of sameness of meaning.

2. Quine explains exactly what notion of analyticity he was

deploring in "Two Dogmas...". He was inquiring into what a definition
of analyticity coul d be, a definition which would be suitable for
determining what would count as analytic in the domain of theoretical
science. He was not maintaining that an adequate definition was
impossible, just that one had not yet been given, and he viewed it as
unlikely that one would be given at all. He notes that he has now given a
definition, but that it is useless in higher scientific theory. His definition
is given in The Roots of Reference: a sentence is analytic for a native
speaker if he or she learned the truth of the sentence while learning to
use one o f its words. Quine accepts that something is analytic if it is true
by definition, but questions where the definition would be. For Quine,
the acceptance of analyticity verges on the acceptance of essentialism.

3. Martin Davies suggests an alternative working definition of

analyticity: someone does not count as being linguistically competent
with an expression in an analytic sentence unless they judge such a
sentence to be true. Quine concedes that this might be more helpful, but
notes that it is a step away from observable criteria. Martin Davies
pursues this issue by suggesting that we could begin to give an account
of meaning by specifying absolute requirements that had to be met when
using a word. Quine is sympathetic to this suggestion, but does not
believe that it would go far enough to show a dichotomy between the
analytic and the synthetic.

4. Martin Davies suggests that, although definitions of analyticity

seem to be circular, there may be nothing wrong with this provided the
circles are large enough. Quine agrees that circularity is not always
enough for an argument to be rejected, but emphasizes that he was not

18
objecting to the notion of analyticity on the grounds of circularity, but
only stating that none of the five definitions he considered in "Two
Dogmas... " was adequate.

5. Quine responds to the criticism that he dismisses a very important

distinction in language, namely that some words (e.g. "polyglot") have a
very tight connection with certain phrases (e.g. "person who speaks
many languages"), while others (e.g. "butter") do not (e.g. "skimmed
and churned bovine secretion"). Quine denies that he dismisses this
distinction-"polyglot" would test positive to the test of analyticity
given in The Roo�s of Reference whereas ''butter" would not.

B. The Notion of Meaning

1. Quine is asked whether he accepts the notion of "the meaning of a

word". He agrees that we give the meaning of the word "table" by
explaining what it denotes. The term "meaning" is often used for
denotation, and in the case of tables he sees no harm.

2. Quine agrees with Wittgenstein's proposal that the meaning of a

word is its use, but argues that in saying this Wittgenstein is offering the
problem, not the solution. For Quine, meaning is certainly dependent on
use, but it is impossible to equate meaning and use-if they we re
identical then no two expressions would have the same meaning, because
no two expressions can have the same use (they will be used by different
people or at different times). Therefore it is necessary to examine which
characteristics of use are relevant to meaning and which are not. Quine
holds that this is where the problem begins, not where it ends.

3. George Boolos asks Quine what he takes to be a science. Quine

uses the term "science" in the broadest sense, to include history,
cognitive science, linguistics, anthropology, etc. Within this
characterization there are gradations, with rigorous sciences (e.g. logic)
at one end, and less rigorous sciences (e.g. history) at the other. Quine's
attack on the notion of meaning is at the rigorous end of science, where
we are trying to find fundamentals, although there is room for a vague
notion of meaning at the other end. When, for example, lexicographers
study the "meanings" of words, Quine views this as a wholly legitimate
study; but the lexicographers' use of "meaning" does not depend on the
notion of synonymy-they are elucidating language (which is a
worthwhile and scientific venture), but not searching for some objective
notion of meaning.

4. Quine explains the use of his dictum "no entity without identity".
He argues that, in order to make sense of the notion of meaning in

19
general, we first have to make sense of the notion of sameness of
meaning, since if we want to attribute abstract existence to a proposition
(which is the alleged meaning of a sentence) then we must be able to say
that it is the same as (or different from) some other proposition. The
notion of synonomy is sufficient for the notion of meaning, since given
synonomy Quine would define the meaning of an expression to be
precisely the class of all its synonyms.

5. Quine views the segmenting of a text into words as tailored

pragmatically and somewhat arbitrarily to the particular language, and
without interlinguistic significance. Grammar should be viewed in terms
of strings of phonemes, and the work of the grammarian is to formulate
the conditions that qualify a string as part of a language. Quine agrees
fully with Chomsky's suggestion that language is "a dialect with an army
and a navy"-we use the term "language" just for convenience.

6. Paul Horwich questions Quine about Frege's argument that we

need the notion of meaning in order to explain the difference between
"Tom believes that the evening star is the evening star" and "Tom
believes that the evening star is the morning star"; our ordinary notion
of reference is not enough. Quine responds that what we are doing when
we assert these sentences is putting ourselves in Tom's position, quoting
Tom but allowing for some inaccuracy in the word-for-word
reproduction. Then the question of reference becomes the same as the
question of reference in a fable (when we have an ontology of fictitious
imaginary objects), and the way to handle this logically is to think of
such discourse as discourse about expressions.

7. Quine discusses the relationship between his views about meaning

in "Two Dogmas..." and those in Word and Object. In Word and Object
he accepts that we can do a lot with the intuitive notion of meaning,
although it is idle as a technical term in philosophical or linguistic
analysis. The only other difference he recognizes is that in Word and
Object he was no longer concerned with the use of analyticity in
Carnap's epistemology.

C. Indeterminacy of Translation
1. Before introducing radical translation in Word and Object, Quine
claims that a speaker's language can be mapped onto itself in such a way
that a) the speaker's dispositions to verbal behaviour remains invariant,
yet b) the mapping is no mere correlation of sentences with equivalent
sentences. Paul Horwich asks Quine whether such a mapping has ever
been done, and Quine responds that it has not. However, he gives an
example of how such a mapping could in principle be carried out. Given

20
two languages and two manuals of translation between the languages,
translate the first language into the second by means of one manual, and
then translate it back again by means of the other manual. This will give
the required permutation of the language.

2. Martin Davies presents Quine with Gareth Evans' criticism of the

indeterminacy of translation, given in "Identity and Predication". Evans'
criticism is as follows. Given a _word such as "Gavagai", we may
translate it either as "rabbit" or as "rabbit-part", and it is indeterminate
which translation would be correct. However, if we allow adjectives in
the alien language, it seems impossible to decide how we should
translate those adjectives given a particular choice of translation of
"Gavagai". Quine responds that it is a question of fact whether such a
decision is possible, and he believes that it is possible. He also adds that
"Gavagai" is an observation-sentence, and so is not used as an example
of indeterminacy of translation- it is used to illustrate the
indeterminacy of reference.

3. Quine is pressed on what he means by "indeterminate". He states

that a term is indeterminate if the sentence predicating it has no truth
value. Paul Horwich and George Boolos attack this definition, asking
Quine to consider the truth value of the sentence "either p or not-p",
where p has no truth value, and Quine concludes that such sentences
must be viewed as meaningless. Boolos follows this by considering cases
of vaguen�ss: "either this•is blue or it is not blue"; Quine's view is that
there is imperfection which one allows for in such cases-we proceed
on the fiction that each term has its fixed denotation, but in fact vague
terms could have any of an indefinitely wide band of alternatives. Quine
does not assimilate indeterminacy of translation with vagueness of
predicates; the only cases in which he conjectures indeterminacy of
translation is when they are drastic, and cannot be handled in a
borderline sense.

D. Meaning and Linguistic Behaviourism

1. Quine states that the single pivotal premise in his argument for
meaning and scepticism is linguistic behaviourism-not to be confused
with specific theories of operant behaviour, conditioned reflex, etc. The
important consideration is complete dependence on observable
behaviour; that is how, Quine maintains, people learn language. Quine
feels that the notion of innate knowledge does not threaten this position,
because even if all our linguistic capabilities were innate, we would still
only be able to interpret someone else by observing his behaviour.

21
2. Quine considers the differences between him and Chomsky. He
feels that they differ in the question of how much innate endowment we
have before learning �anguage, but he sees this merely as a question of
fact to be settled by experiment. Quine also does not agree with the
work of Chomsky's followers done in semantics, although this
disagreement would probably have no bearing on Chomsky's work.
Quine states that Chomsky, in his paper "Quine' s Empirical
Assumptions", completely misunderstood him. He therefore feels that
the differences that are sensed between them are mfferences in what one
of them thought the other was thinking, not what he was really thinking.

3. Martin Davies questions Quine about the intuition that many

people have that it is possible to imagine computers talking and
providing behavioural responses which seem no different from our
own, yet they would still not be said to be thinking beings-there must
be something more to language than just behaviour. Quine responds that
this is an intuition which he does not have-there is no limit in principle
to the extent to which computers can replicate our behaviour, and
whether they were rational or not would only be a matter of degree.
4. Quine outlines his doctrine of the inscrutability of reference. It is
the doctrine that one can reinterpret all the objects that are referred to
in a theory, substituting them for other objects, provided merely that
the class of new objects is in one-one correspondence with the class of
original objects. The thesis is that all the reinterpreted sentences will
retain their truth values, and any possible empirical evidence about the
truth of sentences will remain undisturbed by the reinterpretation. This
makes reference, and even existence, much less absolute and less
significant for science. What matters when we quantify over objects is
not what the objects are, but merely how they relate logically to one
another. The discussion ends with Professor Quine indicating that the
best formulation of his views on meaning is to be found in the second
edition of Pursuit of Truth and he briefly mscusses the origins of these
views.

22
 DREBEN PANEL

A. Modality

1. George Boolos probes Quine' s views on modality and analyticity

by asking the provocative question: "Could one million have been less
than seven?". Quine treats such questions in two stages. Firstly he deals
with the modal construction by what he calls "semantic ascent", dealing
with just the symbols; so in effect the question becomes "Is one million
less than seven T'. Having done this, Quine asserts that the sentence "One
million is not less than seven" is mathematically true. He goes on to
explain what constitutes the firmness of mathematical truths: when there
is a bundle of hypotheses (some empirical, some mathematical) which
has "critical mass"-that is, it implies the possibility of an observational
test-fails an empirical test, then we must reject at least one hypothesis
from that bundle. When choosing which hypotheses to reject, we will
always choose empirical ones before mathematical ones, since a
rejection of a mathematical hypothesis would entail a devastating upset
of our whole theory. It is. this primacy of mathematical truths that
constitutes their firmness.

2. Quine explains what "quantifying in" is, a:i;id why it is such a bad
thing. Quantifying in is when there is a modal operator in the scope of
which lies a variable bound by a quantifier which lies outside that scope.
Quine asserts that accepting this has at least two undesirable
consequences. Firstly it leads to essentialism, since if we allow the
sentence '�for some x necessarily Fx" then we allow that there is
something which is essential to any object fulfilling the predicate F;
secondly, it disallows the substitutivity of identity.

3. Quine then takes this notion a step further. Not only does he not
allow quantifying in-it makes no sense, in his view, to say that some
numbers are necessarily odd-but also he finds no place for a concept
of necessity within science. However, he makes it clear that there is
nothing logically wrong with the notion of necessity, but rather science
has no use for it; it would take an empirical discovery of profound
import for Quine to accept the concept of necessity.

B. Quine's .Characterization of Logic

1. Quine reconciles the fact that in Mathematical Logic (written in

1940) he set out a system of set theory, whereas later he wrote that logic
is to be identified with quantification theory. He asserts that his views

23
did not change, but it was simply that irt 1940 everyone saw the term
"logic" as incorporating set theory, a tradition derived from Frege and
Russell. He made the later point because of his firm belief in a
fundamental distinction between the end of quantification theory and the
beginning of set theory. This he had recognized already in 1940,
although submitting still to the established terminology.

2. Quine begins to give some descriptive characteristics of logic.

One is that, as Tarski argued, logic is essential to all forms of discourse,
while enjoying topic-neutrality as Ryle observed. Another is that all of
logic is "potentially obvious"-everything valid is either obviously so,
or can be made to be obviously so by a series of obviously valid steps.
In other words, logic has a complete proof procedure.
3. Quine does not see it as disturbing that there is no complete
procedure to show invalidity. He gives two reasons why it is intuitively
preferable that logic should concentrate on validity rather than
invalidity. One is that, given an empirical discovery, we broaden our
knowledge by drawing the logical consequences of that discovery; the
other is that if one shows that the deduction of something from true
premises is invalid, one still has not shown that the conclusion is false.

4. Martin Davies encourages Quine to clarify his criterion of

potential obviousness. He argues that it is obvious that flesh decays, but
that this is not the sort of obviousness that would apply in logic. He
questions what further criterion is needed to separate the right sort of
obviousness from the wrong sort. Quine responds by pointing out that
potential obviousness is merely a trait· of logic, not a sufficient condition
for something to be logical. Quine believes, with Tarski, that it is
impossible to provide necessary and sufficient conditions for what
constitutes logic; we can only specify logic for our purposes,
enumerating the logical vocabulary. Burton Dreben notes that this
distinguishes Quine from Carnap in this regard.

5. George Boolos notes that _there are principles of set theory which
are also potentially obvious, such as the axiom of extensionality. Quine
agrees with this, but notes that this does not make such principles
logical. All of logic is potentially obvious; this is a reason for
distinguishing logic from set theory. Quine gives two further
distinguishing characteristics: logic is complete whereas set theory is
not; and logic is neutral in its ontology, whereas the ontology of set
theory demands sets.

24
C. Ontology
1. Quine argues that the only interest in questions of ontology is a
stylistic interest-an interest in getting a more economical, elegant
formulation of the theory. For example, from a structural point of
view, it is less confusing to talk about sets of quadruples of real
numbers than it is to talk about tables, mountains, etc.

2. Quine defends his view that ontological questions cannot even

arise with regard to a decidable theory, and in fact they can only arise
within a theory with an infinite (or at least indefinitely open) domain of
discourse. Within a decidable theory one can set up a recursive
definition so that every true sentence in the theory reduces to zero
equals zero, and every false sentence to zero equals one. Clearly there is
no call for an ontology here. Then, in theories with finite and
exhaustively labelled domains of discourse, there is no need for
quantification (simply use truth-functions instead), and so ontology is
again unnecessary. Even in theories with infinite domains of discourse,
choice of ontology is irrelevant because of the indeterminacy of
reference.
3. George Boolos presses this argument by considering arithmetic.
Arithmetic without multiplication is decidable, but arithmetic with
multiplication is undecidable. It seems strange that the introduction of
multiplication brings with it a genuine question of ontology. Quine gives
two responses to this. Firstly, in practice it is simpler to assume an
ontology than not, so that we do not have suddenly to adopt one when
we move into an infinite domain of discourse. Secondly, the role of
ontology is to reinforce logical connections within a system, and it just
happens that arithmetic without multiplication is so weak that such
reinforcement is not needed.
4. Quine considers the epistemic status of mathematical sentences
which can never play a role in applicable science (for example the
axiom of choice and the continuum hypothesis). These sentences warrant
investigation because they consist in the same mathematical vocabulary
and syntax as other sentences which are applicable; and it is more
convenient to include such sentences in our theory than not, because to
dismiss them as meaningless would require an extremely complicated
mutilation of our grammar. Quine then discusses the problem of how to
determine the truth-value of these sentences. We cannot use the criterion
of demonstrability, because Godel has shown that there is no complete
proof procedure even for applicable mathematics, let alone inapplicable
mathematics. Quine settles for considerations of simplicity and
economy, and in particular he favours Godel's criterion of
constructability, and adopts the axiom V = L. Quine sees no distinction,

25
at this extreme, between saying that the criterion is a guide to truth and
that it constitutes the truth; just as, at the other extreme, he saw no
distinction between some changes in logical theory and changes in
terminology.

D. History of Logic

1. Quine defends his claim that "logic is an ancient subject, and since
1879 it has been a great one". Although this seems to slight Boole's
work, Quine argues that Frege's leap in logic was so great that, if a
boundary had to be drawn in the history of logic, it would have to be
with Frege. It is noted that Peirce independently achieved almost the
same results as Frege, as little· as two years later; however, although
Peirce had the quantifier, he did not have a full system-there are no
precise rules.

2. Quine discusses the impact of the Godel incompleteness theorems

on philosophy in the United States when they came out. He says there
was little impact on the Anglo-Saxon world for two or three years.

3. Quine discusses Andrew Wiles' discovery of a proof of Fermat's

last theorem, and views it as very exciting. Quine has an amusing
combinatorial equivalent of the th�orem, published in Quiddities, and he
tells the story of its discovery.

4. Quine outlines his experiences in logic when he was in Europe in

the 1930s. He went to Carnap in Prague, and then to Lukasiewicz in
Warsaw; it was through Lukasiewicz that he met Tarski. Quine views
Tarski's truth definition as the first term in a sequence of denotation: a
one-place predicate denotes each of the things it is true of, a two-place
predicate denotes each pair of things it is true of, etc. A zero-place
predicate is simply a sentence, and truth is its analogue of denotation.
,
· Truth, like denotation, can be explained by "disquotation, : just as
"rabbit" denotes rabbits, so "Snow is white" is true if snow is white.
Disquotation can be parlayed into a formal inductive definition of
denotation for formalized languages with finite vocabularies of
primitive predicates; and a formal definition of truth then emerges as
the zero case. This was Tarski's work.

5. Quine sees this recursion as doing full justice to the notion of

truth, except for one point. When something accepted as true is
disproved, we do not say that it ceased to be true; we say that it turned
out to have been false. The scientist is seen not as ·making truth, but as
fallibly seeking it.

26
6. Quine gives a brief outline of how much mathematical logic he
feels it is appropriate for a philosopher to -be aware of: a minimum of a
knowtedge of truth-function theory and quantification theory, together
with a complete proof procedure of those plus identity. Beyond that,
Quine thinks it desirable for the philosopher to be acquainted with the
logical paradoxes, and the set theoretical axioms in several alternative
set theories, and the completeness proof for logic. Also Godel' s
theorem, with some acquaintance of its proof, although not necessarily
in exhaustive detail.

27
 BLOCK PANEL

A. Philosophy of Mind

1. Quine justifies his view that inverted spectrum situations are

possible but irrelevant to the linguistic assessment of communication. He
sees introspection as indispensable heuristically, but it is not
explanatory. He goes on to say that, since no objective test of the
situation is possible, the inverted spectrum problem is paradoxically
meaningless; paradoxical because of the vividness of its apparent
meamng.

2. Quine explains why he believes that it is possible for two different

beings to exhibit the same behavioural characteristics while at the same
time be in different neural states. The full explanation of linguistics is
not to be found in behaviour, but in the nervous system; however, this
further explanation is irrelevant to the systematization of linguistic
behaviour.

3. Quine holds that the ground for maintaining an identity theory of

mind is that all observational evidence for our overall theory would
remain undisturbed by identifying the mental and the physical. He goes
on to say. that the difference between eliminativism and physicalism is
merely verbal, since the acceptance of one rather than the other would
make no observable difference to our scientific theory. Quine sees no
need to explain why some particular neural state gives rise to a
particular feeling: any such explanation could only be in terms of
previous physical causes, and even if we found such an explanation we
would not know that we had done so.

4. In response to Ned Block's questioning, Quine admits that the

possibility of an inverted spectrum situation is meaningful, since be has
allowed that we could have evidence sufficient to identify certain
sensations with certain neural states. Then, if A's sensation of seeing red
were identical to some neural state identical to B's sensation of seeing
green., we would have observational evidence for an inverted spectrum.

5. Martin Davies asks Quine to respond to Frank Jackson's

knowledge argument: Mary knows everything there is to know about
science, but she has gained this knowledge in a black and white
environment. Surely, the argument goes on, her knowledge is
incomplete; she lacks the experience of what it is like to see red. Quine
sees this situation as meaningless, since there is no experimental test to
set up. All we can say about Mary is that she has spotted a new

28
subjective symptom, namely an ability to see red; but such an ability is
not knowledge.

6. Quine views functionalism as helpful in providing a partial

explanation of the mind-it helps to simplify and systematize things. As
such, Quine believes it does not conflict with physicalism, eliminati vism
or behaviourism; but an exclusive subscription to functionalism would
force us to give up the objective of ultimate reduction to physical
reality.

7. Quine expresses sympathy with Donald Davidson's theory of

anomalous monism, the view that every particular datable mental
occurrence, or event can be identified with a particular physical event,
but that there is a shortage of psychophysical laws connecting the mental
and the physical. Quine discusses this theory in relation to the
propositional attitudes. He accepts the attribution of propositional
attitudes de dicto, where we are ascribing the belief by putting it in the
subject's language as best we can, and he makes sense of this by treating
the objects of belief as the actual linguistic forms. the sentences. He sees
the virtue of this to be that it preserves extensionality; there is no
problem of quantifying in because we do not quantify in. However,
Quine sees a vagueness in the verbs themselves, the attitudinals like
"believes"-there is a lack of connection between these verbs and our
overall physical theory. Quine sees this as a case of anomalous monism.

8. Paul Horwich pursues Quine' s outline of his treatment of

propositional attitudes de dicto. He argues that this treatment depends on
the notion of synonomy, since "Tom said that snow is white" would be
rendered as "Tom made an utterance which means the same as my
sentence 'snow is white "'. Quine responds that this is not the case, but
that his treatment does depend on translation, and should be regarded as
subsidiary to a chosen manual of translation. H owever, he
accommodates this problem in practice by empathy-we attribute a
belief by considering what belief we would have if we were in the
subject's position. Quine emphasizes that such empathy does not have to
be perfect; we only need to empathize to the extent that is relevant, and
hence it is possible to attribute propositional attitudes even to
languageless creatures. Quine prefers his principle of empathy to
Davidson's principle of charity (the view that we should always
interpret the subject in the way that makes it most likely that what he
says is true), since Davidson's principle will neglect tribal differences of
belief. He emphasizes the vagueness of propositional attitudes, and
points out that attribution of them is bound to grade off in respect of
objectivity. Belief in general, as an integrated concept, resists absorption
into science.

29
9. Ned Block returns to· Quine's view that reductionism and
eliminativism are essentially the same with regard to subjective
experience. He views this as strange, since in so many other areas of
science it makes a great difference whether we reduce or repudiate a
theory. Quine argues that in all cases the choice of whether to reduce or
to eliminate is decided on the basis of efficiency and economy of theory.
In the case of subjective experience, there is no observational distinction
to be drawn between the two approaches, and so it is arbitrary which
one to take. He agrees that it is intuitively wrong to say that there is no
such thing as subjective experience, but points out that this does not
mean that it is wrong.

B. Linguistic Behaviourism

1. Quine explains how his linguistic behaviourism is independent of

psychological behaviourism. Linguistic behaviourism is unavoidable in
Quine' s opinion because language is learned only by the observation of
others' behaviour. Quine holds that although there is a lot of internal,
non-behavioural activity which underlies our use of language, it is only
the external behaviour which we have to go on when learning the
language. Psychological behaviourism may or may not be correct, but it
is irrelevant to Quine's philosophy of language.

2. Martin Davies asks Quine whether he still thinks that there is a

problem of making empirical sense of the distinction between
extensionally equivalent grammars. Quine does not view this as he did
in "Methodological Reflections on Current Linguistic Theory". He notes
that Morris Halle suggested to him a way of behaviourally
discriminating between extensionally equivalent grammars: the best
grammar would be the one whose rules most closely resembled those
acquired by a child, and we observe these by studying the child's
successive mistakes and corrections in his acquisition of a language.
Quine also feels that data gained by observing the decay of speech (for
example after a brain in jury) would be relevant to constructing the best
grammar.

3. Davies attacks this response, asking Quine ifhe is not laying down
ahead of the development of science a boundary determining what kind
of evidence is even potentially relevant when choosing between .
grammars. Quine disagrees; he states that other sources of evidence
could be relevant, but at present he sees only these two-it is the job of
science to discover whether or not this is the case. Quine explains that
the evidence gained from observing, say, a Japanese speaking child
could be relevant to discovering the grammar of an English speaking
child, but that such evidence should be treated with caution.

. 30
4. Quine views it as a common mistake in linguistics to attribute
universality to things which do not hold for all languages. For example,
grammarians studying an exotic language will use the term "pronoun",
but the term may have applications in such languages different from our
own. These differences do not matter as long as they are recognized, but
Quine feels that all too often they are not. He holds that there is a great
deal that is universal in language (for exaple innate readiness), but
what he thinks is required is some particular-language-independent
criterion for what constitutes a linguistic universal.

5. Paul Horwich clarifies Quine's view that there are two separate
but closely linked projects. One is linguistics, the aim of which is to
systematize a purely behavioural phenomenon; the other is the study of
the causes of this phenomenon, the neural activity in the brain.

C. Cognitive Science

1. Ned Block points out that, in the field of psychology, cognitive

science has prospered while behaviourism has faded. He suggests that
Quine, given his views on the supremacy of science, should perhaps
abandon psychological behaviourism because of this fact. Quine
disagrees with this, remarking that we should not follow science
slavishly, we should follow it critically and exploit it. Philosophy should
not be subservient to science, it should merely recognize that there is no
empirical access to information th�t is more dependable than science
itself. Quine states again that psychological behaviourism is irrelevant to
his work, it is the concern of psychologists in the field.

2. Quine does not disapprove of cognitive psychology as a science,

and in fact very much respects the work being done in language­
learning and semantics. Wliat he dislikes about cognitive science is its
focus on a theory of communication of ideas, such ideas being meant to
represent meanings. Quine emphasizes that behaviourism is
fundamentally important, even in psychology, since behavioural
evidence is all we have available for study. However, much of cognitive
science is for Quine a viable way of construing that evidence. He states
that he would even approve of the theory of communication of ideas, if
it were (I) clarified and (2) linked conjecturally with neurology by
some promising hypotheses.

3. Quine has at times made speculative empirical claims about

psychology in order to reinforce his philosophy of language. Ned Block
argues that these views may be false, and he wonders whether their
falsity would really affect Quine' s views in philosophy of language.

31
Quine responds that his speculation in psychology is mirrored by many
psychologists' ignorance in logical analysis, and what is required is
more interaction between the two fields of study. For example, a
developmental psychologist may write about the first occurrence of
divided reference in a child's history, without being aware of the
sophistication of the argument surrounding the topic. He goes on to
argue that what speculation he does make in psychology is very unlikely
to be disagreed with. However, Quine adds that, like all science, his
philosophy is fallible, and it might ensue that psychologists produce
enough evidence to force him to abandon some of his theories.

This program ends by returning briefly to philosophy of mind when the

panel explores Quine's views on the topics of mental causation and
consciousness. He concludes that consciousness remains for him a deep
mystery.

32