Hilary Putnam, Mario de Caro (Editor) - Naturalism, Realism, and Normativity-Harvard University Press (2016)

Naturalism, Realism, and Normativity
Unauthenticated
Download Date | 6/17/16 3:51 PM
Unauthenticated
Naturalism,

Realism, and

Normativity
Hilary Putnam
EDITED BY M A RIO DE C A RO
Cambridge, Massachusetts
London, England
2016
Unauthenticated
Copyright © 2016 by the President and Fellows of Harvard College
All rights reserved
Printed in the United States of Amer ica
First printing
Acknowledgments on pages 227–228 constitute an extension of this copyright page.
Library of Congress Cataloging-in-Publication Data

Names: Putnam, Hilary, author. | De Caro, Mario, editor.
Title: Naturalism, realism, and normativity / Hilary Putnam ; edited by Mario De Caro.
Description: Cambridge, Massachusetts : Harvard University Press, 2016. | Includes
index.
Identifiers: LCCN 2015033370 | ISBN 9780674659698 (alk. paper)
Subjects: LCSH: Realism. | Naturalism. | Perception (Philosophy) | Normativity
(Ethics)
Classification: LCC B835 .P86 2016 | DDC 191— dc23 LC record available at
http:// lccn.loc.gov/2015033370
Unauthenticated
Contents
Introduction: Putnam’s Philosophy and Metaphilosophy

by Mario De Caro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
I . L I B E R A L N AT U R A L I S M A N D N O R M AT I V I T Y
1. Naturalism, Realism, and Normativity . . . . . . . . . . . . . . . . . . . . . . 21
2. On Bernard Williams’s “Philosophy

as a Humanistic Discipline” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .44
3. What Evolutionary Theory Doesn’t Tell Us about Ethics . . . . . . . 56
II . RE ALISM AND ONTOLO GY
4. Sosa on Internal Realism and Conceptual Relativity . . . . . . . . . . . 67
5. Richard Boyd on Scientific Realism . . . . . . . . . . . . . . . . . . . . . . . . . 90
I I I . R E A L I S M A N D V E R I F I C AT I O N I S M
6. Hans Reichenbach: Realist and Verificationist. . . . . . . . . . . . . . . . . 99
7. Between Scylla and Charybdis: Does Dummett Have

a Way Through? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
8. When “Evidence Transcendence” Is Not Malign . . . . . . . . . . . . . 128
Unauthenticated
vi Contents
I V. N A I V E R E A L I S M , S E N S AT I O N , A N D A P P E R C E P T I O N
9. Sensation and Apperception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
10. Perception without Sense Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
11. “Naive Realism” and Qualia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
V. L O O K I N G B A C K
12. The Development of Externalist Semantics . . . . . . . . . . . . . . . . . 199
13. Sixty-Five Years of Philosophy: A Participant’s Thoughts and

Experiences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
Unauthenticated
Unauthenticated
Unauthenticated
Introduction
PUTNAM’S PHILOSOPHY AND METAPHILOSOPHY
by Mario De Caro
T h e t i t l e of t h is volu m e , which is taken from the

fi rst essay collected here, is meant to direct the reader’s attention to
three main philosophical questions with which Hilary Putnam has been
dealing for very many years. The questions grow out of, on the one
hand, the conclusions reached in Putnam’s early papers, collected in the
first two volumes of his Philosophical Papers,1 in which he criticized log-
ical positivist accounts of the meaning of scientific terms and the na-
ture of scientific theories, in part because they were “incompatible with
a rather minimal scientific realism,”2 and, on the other hand, Putnam’s
return to a more nuanced version of those early realist views after a
fourteen-year (1976–1990) swerve in a direction that, notwithstanding
his original intentions, substantially leaned toward antirealism (Put-
nam’s “internal realist” years).3 The questions thus presuppose Put-
nam’s realism (prescinding from the aforementioned “swerve”) as well
1. Mathematics, Matter and Method, vol. 1, and Mind, Language and Reality, vol. 2 of Phil-
osophical Papers (Cambridge: Cambridge University Press, 1975).
2. As Putnam says in “What Theories Are Not,” a lecture delivered to the 1960 Interna-
tional Conference on Logic, Methodology and Philosophy of Science at Stanford Univer-
sity, and collected in Putnam’s Mathematics, Matter and Method, 215–227. “Rather minimal
scientific realism” occurs on p. 224. (This is the fi rst time Putnam recalls calling his posi-
tion “scientific realism.”)
3. The clearer exposition of Putnam’s internal realism is in his Reason, Truth, and History
(Cambridge: Cambridge University Press, 1981).
1
Unauthenticated
2 Introduction
as his more recent interest in what he calls the “entanglement” of fact

and value.4
1. Assuming that one accepts the arguments that Putnam has

given for many years, that antirealism, in its many forms, is an
unacceptable view, what form of realism should one embrace?
2. Assuming that the philosophies that appeal to the supernatural
are not acceptable, what form of naturalism should one endorse?
3. Assuming that normativity is an essential component of our
rational and moral practices, should or could we assume that, at
least in principle, it is reducible to the non-normative compo-
nents of such practices?
All these questions, besides being of primary philosophical interest,

presuppose deep metaphilosophical reflections regarding the nature,
role, goals, and methods of philosophy. Let’s examine them in turn.
1. Realism
In 1983 Michael Devitt wrote an article very unsympathetically titled
“Realism and the Renegade Putnam.”5 In that article, Devitt— a stu-
dent of Putnam’s during his graduate years at Harvard— complained
that, by opting for “internal realism,” Putnam became a “renegade”
from philosophical realism, which he had previously defended in the
papers collected in the Philosophical Papers volumes.
Indeed, in 1990 and subsequently, Putnam granted that internal
realism— a view based on an epistemic conception of truth—leaned
too much in an idealistic direction.6 But while acknowledging that in-
ternal realism was a mistake, one can easily see that Devitt’s harsh
judgment was quite unfair. As a matter of fact, Putnam never com-
pletely renounced realism. First of all, in his entire career (including his
4. Putnam used the term “entanglement” in The Collapse of the Fact / Value Dichotomy and
Other Essays (Cambridge, Mass.: Harvard University Press, 2002), but the idea, if not the
term, was already present in his Renewing Philosophy (Cambridge, Mass.: Harvard Univer-
sity Press, 1992) and in his subsequent essays on Dewey’s philosophy, such as those pub-
lished in Words and Life, ed. James Conant (Cambridge, Mass.: Harvard University Press,
1994).
5. Michael Devitt, “Realism and the Renegade Putnam: A Critical Study of Meaning
and the Moral Sciences, Noûs 17, no. 2 (1983): 291–301.
6. See Chapters 4, 7, and 8 in this work for some discussions on internal realism.
Unauthenticated
Introduction 3
internal realist period) Putnam maintained that the theoretical terms of

our best scientific practice do refer to real entities, even if such entities
are in principle unobservable (i.e., electrons and black holes), and this
means that he always rejected all forms of scientific antirealism, such
as instrumentalism, conventionalism, operationalism, and relativ-
ism.7 Second, when he moved away from his early physicalistically ori-
ented realism toward internal realism, Putnam was motivated by his
desire to shape a satisfying philosophical realism—that is, a realism
able simultaneously to (1) accept the approximate and revisable cor-
rectness of the scientific worldview and (2) accept the approximate and
revisable correctness of the commonsense worldview, which Putnam
then thought (and still thinks) was threatened by the reductionist con-
ceptions of reality. In this light, at least since 1976 Putnam has rejected
all positions that are unable (or worse, do not even try) to do full jus-
tice, at the same time, to science and common sense. Arguably, Put-
nam’s painstaking and uninterrupted efforts to shape a version of
realism able to acknowledge the partial and revisable but very real
verisimilitude of both the commonsense and the scientific images of
the world, without privileging one over the other, is one of his most
relevant bequests to the next generations of philosophers.
Before entering in more detail into Putnam’s views on these issues,
some remarks can be useful. First of all, it should be noted that the
problem of realism does not have an “all or nothing” form. No serious
philosopher has ever been a complete realist or a complete antirealist:
even Alexius Meinong, perhaps the most realist of all philosophers, de-
nied that a round square could be real; and even bishop Berkeley, a
champion of antirealism in regard to the material world, was an arch-
realist as to the mind (in particular the divine mind). In a word, all rel-
evant philosophical conceptions are a mix of realistic and antirealist
components: one can be a realist in one sense, without being a realist
in another sense.8
Moreover, as has frequently been noted, there is not one form of re-
alism, but many, which can be grouped in three main families:
7. See Hilary Putnam, “On Not Writing Off Scientific Realism,” in Philosophy in an Age
of Science: Physics, Mathematics, and Skepticism, ed. Mario De Caro and David Macarthur
(Cambridge, Mass.: Harvard University Press, 2012), 91–108.
8. The themes of this section of the introduction are analyzed more in depth in Mario
De Caro, “Realism, Common Sense, and Science,” The Monist 98 (2015): 197–214.
Unauthenticated
4 Introduction
1. Ontological realism. The views that belong to this family claim

the reality of certain sorts of things: concrete (trees, galaxies, or
atoms) or abstract (numbers, universals, or disembodied minds),
properties (redness, goodness, electric charges, or freedom of
the will) or events (the Big Bang, transubstantiation, the Moon
landing). Moreover, the theories of the ontological-realist family
can be general (such as when they claim the existence of the
external world or of time) or particular (when they claim the
existence of specific things or classes of things).9
2. Semantic realism. In a taxonomy due to the late Michael Dum-
mett,10 the views of this family claim that the meaning of a
sentence is given by the conditions under which it is true, and
that truth requires the reality of the entities to which the terms
of that sentence refer (whereas semantic antirealist views claim
that the meaning of a sentence is given by the conditions under
which speakers are justified in asserting it).
3. Epistemological realism. The views of this family claim that there
may be real facts and truths that are, in principle, unknowable (a
possibility that is denied by epistemological antirealists).
Often the discussions of these three forms of realism are intercon-

nected— and this is certainly the case for Putnam—but this is not
necessarily so. However, besides this formal taxonomy, the dif ferent
versions of realism can also be distinguished in virtue of their substan-
tial claims. In particular, today the two most important versions of re-
alism (in the general sense of the term), taken in their purest forms, are
(1) commonsense realism, the view that acknowledges as real only the
things we can experience directly (with our senses, such as Eddington’s
famous commonsense table)11 or indirectly (by using the instruments
that extend our senses, such as microscopes or telescopes), and (2) hyper-
scientific realism, the view that claims that (1) we should accept as real only
9. When discussing the independent existence of an entity, one should not pose the
question in genetic or causal terms. The table in front of me obviously exists because
someone built it; however, once it is built, the table exists independently of whoever built it.
10. The most complete discussion of metaphysical realism and antirealism by Dummett
is in his The Logical Basis of Metaphysics (Cambridge, Mass.: Harvard University Press, 1991).
11. Eddington spoke of “two tables,” the commonsense table and the physicist’s table, in
The Nature of Physical Reality (Cambridge: Cambridge University Press, 1929). On this, see
Putnam, “On Not Writing Off Scientific Realism.”
Unauthenticated
Introduction 5
the entities and events (both observable and unobservable) to which our
best scientific theories are ontologically committed, and none of these
has the properties that our commonsense objects are supposed to have
(Eddington’s physicists’ table was not solid, not smooth, etc.). When
taken in their purest forms these two forms of realism are incompat-
ible because each of them is drastically antirealist in the field in which
the other assumes a firmly realist attitude.
It could be argued that the origin of this opposition can be tracked
back to Aristotle’s defense of the commonsense view of the world against
Plato’s peculiar idealism (which actually was a strong form of realism
regarding ideal forms!). More safely, however, one can claim that the
discussion opposing commonsense realism to hyperscientific realism
became explicit when the scientific revolution obtained a foothold. As
early as 1597, Jacopo Mazzoni— a phi losopher who was a colleague
and a good friend of Galileo in Pisa and influenced his intellectual
development—wrote an important book in which, among other things,
he opposed the Aristotelianism of his own time, which defended the
commonsensical, perception-based view of the world, in favor of a ver-
sion of Platonism according to which the natural world is mathemati-
cally structured.12 A few years later, expanding on that view, Galileo
offered the basics of the modern scientific worldview by claiming that
only the mathematizable properties of the bodies investigated by physics
are real, while the so-called secondary qualities—the qualitative prop-
erties that we get through perception and commonsensically attribute
to the external bodies—have no independent existence:
As far as concerns the subject in which these tastes, smells, colours,

etc. appear to us as inhering, they are nothing other than mere
names, and they inhere only in the sentient body. Consequently,
if the animal were removed, all these qualities would disappear and
be annihilated.13
12. Jacopo Mazzoni, In Universam Platonis et Aristotelis Philosophiam Praeludia, sive de

comparatione Platonis et Aristotelis (Venezia, 1597); English translation in Alexander Koyré,
“Galileo and Plato,” Journal of the History of Ideas 5 (1943): 400–428, quotation on p. 421.
13. Galileo Galilei, in Le Opere di Galileo Galilei, repr. ed. Antonio Favaro, Edizione Na-
zionale, vol. 6 (Firenze: G. Barbèra, 1968), 197–372; The Assayer, trans. S. Drake, in The
Controversy of the Comets of 1618 (Philadelphia: University of Pennsylvania Press, 1960),
151–336. It should be noted that for Galileo “mathematizable” meant “geometrizable.” On
Unauthenticated
6 Introduction
But Galileo did not only defend a rigorous mathematical realism that
relegated secondary properties to the realm of the subjective and cred-
ited as real only primary properties; he also claimed that atoms, the un-
observable entities whose existence he postulated in order to explain
the changes of material states, were real.14 What we have here called
“hyperscientific realism” was born.
In the next centuries, hyperscientific realism was not unchallenged
(idealists and empiricists, for example, rejected it), but it remained an
important presence in Western culture. In his The Crisis of European
Sciences and Transcendental Phenomenology (1936), Edmund Husserl of-
fered an influential genealogical account of the antirealist view of the
commonsense world produced by modern science. In his opinion, Gal-
ileo was responsible for
the surreptitious substitution of the mathematically substruc-

tured world of idealities for the only real world, the one that is
actually given through perception, that is ever experienced and
experienceable— our everyday life-world. This substitution was
promptly passed on to his successors, the physicists of all suc-
ceeding centuries.15
According to Husserl, then, the only real world is the “life-world,”

that is, the world of human experience in which the secondary quali-
ties do belong to the external objects in which we commonsensically
locate them, and values and meanings may be objective.16 In Husserl’s
view, the life-world is the “forgotten meaning-fundament of natural sci-
Galileo’s scientific realism (which in his case was a form of Platonism), see Mario De Caro,
“On Galileo’s Platonism, Again,” in Hypotheses and Perspectives within History and Philosophy
of Science. Hommage to Alexandre Koyré 1964–2014, ed. Raffaele Pisano, Joseph Agassi, and
Daria Drozdova (Dordrecht, the Netherlands: Springer, forthcoming).
14. In the last years of his life, Galileo might have abandoned atomism in favor of a con-
tinuous conception of space: see Michele Camerota, “Galileo, Lucrezio e l’atomismo,” in
Lucrezio, la natura e la scienza, ed. Marco Beretta and Francesco Citti (Firenze: Biblioteca di
Nuncius / Istituto e Museo di storia della scienza, Leo S. Olschki, 2008).
15. Edmund Husserl, Die Krisis der europäischen Wissenschaften und die transzendentale
Phänomenologie: Eine Einleitung in die phänomenologische Philosophie, 1936; published in En-
glish as The Crisis of European Sciences and Transcendental Phenomenology, trans. David Carr
(Evanston, Ill.: Northwestern University Press, 1970), 48 (emphasis added).
16. Husserl’s “life-world” is a refi nement of the everyday, subjective experience of indi-
viduals, obtained through the “phenomenological reduction.” See Dermot Moran, Hus-
Unauthenticated
Introduction 7
ence,”17 and scientific concepts are mere idealizations with practical

uses, such as measurement and prediction, but do not refer to any un-
observable real ity. In this light, science can only be interpreted
operationalistically—that is, in antirealistic terms.
Husserl’s Crisis did not offer the first exposition of a pure form of re-
alism centered on the ontological primacy of the ordinary image of the
world (Pierre Duhem, for example, was reported to have called himself
“the unceasing apostle of common sense”);18 but certainly it was one of
the purest, for its explicit refusal to accept scientific ontology. Many
continental philosophers have followed Husserl in this direction, but
not only continental philosophers have done so.
The philosopher of science Bas van Fraassen, a leading analytic phi-
losopher, offers an excellent example in this sense. His “constructive
empiricism” (which should be not confused with the most common ver-
sions of empiricism)19 explicitly attempts at reconciling commonsense
realism and scientific antirealism. On the one hand, van Fraassen takes
as completely unproblematic the reference of “language . . . to trees and
mountains, people and books.”20 On the other hand, he is a strict an-
tirealist with respect to scientific theories that appeal to unobservable
entities, even when they produce accurate explanations and predic-
tions of observable evidence.21 Van Fraassen’s view represents one of
the most consistent expressions of commonsense realism because it
limits the scope of our ontology to what is observable and, accord-
ingly, assumes an antirealist position with regard to scientific theories
insofar as they make reference to unobservable entities.
Putnam is very sympathetic to the positive side of commonsense re-
alism as long as it wants to protect the ordinary view of the world from
the reductionist and eliminationism campaigns that are so common
serl’s Crisis of the European Sciences and Transcendental Phenomenology: An Introduction (Cam-
bridge: Cambridge University Press, 2012).
17. Husserl, Crisis, 48.
18. R. N. D Martin, Pierre Duhem: Philosophy and History in the Work of a Believing Physi-
cist (La Salle, Ill.: Open Court, 1991).
19. Van Fraassen’s constructive empiricism is very dif ferent from the more traditional
empiricism of, say, Hume and the neopositivist philosophers, according to which our
knowledge of the external world coincides with the knowledge of our sense data.
20. Bas van Fraassen, “On McMullin’s Appreciation of Realism,” Philosophy of Science 70
(2003): 479–491, quotation from p. 480.
21. Bas van Fraassen, The Scientifi c Image (Oxford: Oxford University Press, 1980).
Unauthenticated
8 Introduction
nowadays. He, however, strongly disagrees with the negative side of

commonsense realism—let us call it “hypercommmonsense realism”—
with its antirealism concerning science. In this perspective, he has
been a tireless proponent of the so-called no-miracles argument, which
is based on the idea that the only way of explicating the great explana-
tory and predictive success of the best theories of modern science is to
acknowledge that these theories are true (or approximately true) in re-
gard to the natural world and refer to real entities, even when those are
unobservable.22 From the point of view of antirealism, on the contrary,
the fact that science works so well in offering comprehensive explana-
tions and extremely precise predictions of observable phenomena is an
inexplicable mystery, if not a sheer miracle. Consequently, for Putnam
we should take our best scientific theories as approximately true and
the entities those theories refer to as real. Unsurprisingly, antirealists
have tried to attack the miracle argument in various ways, but Putnam
and others have responded to their arguments in convincing ways.23
In maintaining that scientific theories can be true or approximately
true, Putnam insists that they are not necessarily verifiable under epi-
stemically ideal conditions, as he instead believed during his internal
realist period.24 A convincing supporting example of why truth is not
epistemically constrained is the conjecture “There is no life outside the
earth,” which, if true, would be unverifiable even in ideal epistemic con-
ditions. Therefore, according to Putnam (apart from the internal re-
alist period), scientific theories can be true (or approximately true) even
if we are never able to verify them.
22. Hilary Putnam, “Do True Assertions Correspond to Reality?” in Language and Re-
ality, vol. 2 of Philosophical Papers (Cambridge, Mass.: Cambridge University Press, 1975),
70–84.
23. The main charges against the miracle argument are that it is based on an inference
to the best explanation (which allegedly is either doubtful in itself or not applicable at
this meta- explanatory level); that dif ferent theories can account for the same sets empir-
ical data; and that the history of science shows that false theories (such as Ptolemaic as-
tronomy) can offer excellent predictions. Against these arguments, see Stathis Psillos,
“The Scope and Limits of the No Miracles Argument,” in Explanation, Prediction, and
Confi rmation, ed. Dennis Dieks et al. (Dordrecht, the Netherlands: Springer, 2011), 23–35;
Hilary Putnam, “On Not Writing Off Scientific Realism”; Mario De Caro, “Review of
J. Ritchie, Understanding Naturalism,” International Journal of Philosophical Studies 19 (2011):
527–531; Mario Alai, “Novel Predictions and the No Miracle Argument,” Erkenntnis 79
(2014): 297–326.
24. See Chapter 5 here.
Unauthenticated
Introduction 9
But the story of contemporary realism gets more complicated. In

fact, many scientific realists today take the extreme stance that we
called hyperscientific realism, claiming not only that scientific re-
alism is correct, but also that the real ity described and explained by
science is the only real ity there is. Consequently, according to those
phi losophers scientific realism is the only realism worth defending.
This is not a view that Putnam shares. For him, scientific naturalism
and commonsense realism are to be reconciled, and this leads him
to the belief that we have to work within the framework of “liberal
naturalism.”
2. Naturalism
The vast majority of scientific realists, including Putnam, defend phil-
osophical naturalism—the conception that stresses the indispensability
of the scientific worldview and rejects any appeal to the supernatural.25
Among these philosophers, however, many combine naturalism with
“hyperscientific realism,” the previously mentioned view according to
which every element of the commonsense worldview that cannot be re-
duced to the scientific worldview is to be rejected. The resulting concep-
tion has been called “scientific naturalism” or “strict naturalism,” and
Putnam has relentlessly criticized it for the last four decades.
In this battle, Putnam has not been alone: the list of philosophers
who have tried to shape more tolerant forms of naturalism—forms that
have been grouped under the label “liberal naturalism”26 —includes
Donald Davidson, John McDowell, Jennifer Hornsby, P. F. Strawson,
Barry Stroud, and Tim Scanlon and is growing fast. However, liberal
naturalism is still a minority view because most of the philosophical
scene in analytic metaphysics continues to be occupied by scientific
naturalism.
It is ironical that, at least in part, scientific naturalism stemmed from
Husserl’s Crisis. In fact, Wilfrid Sellars, one of the founders of scien-
tific naturalism, found inspiration in that book (with the mediation of

26. Mario De Caro and David Macarthur, Naturalism in Question (Cambridge, Mass.:
Harvard University Press, 2004) and Naturalism and Normativity (New York: Columbia
University Press, 2010).
Unauthenticated
10 Introduction
his mentor Marvin Farber, who had been a student of Husserl’s in

Freiburg) for his famous distinction between the “manifest” and the
“scientific image of the world” and for his “stereoscopic vision” (i.e., a
unified view of those two images).27 Sellars, however, interpreted the
opposition between the two worlds in a way that was diametrically op-
posed to Husserl’s. While Husserl was a realist about common sense
and an antirealist about science, Sellars took the opposite stance, which
he expressed with the famous neo-Protagorean motto:
Speaking as a philosopher, I am quite prepared to say that the

common sense world of physical objects in Space and Time is
unreal—that is, that there are no such things. Or, to put it less para-
doxically, that in the dimension of describing and explaining the
world, science is the measure of all things, of what is that it is, and
of what is not that it is not.28
Sellars and the other founding father of scientific naturalism, V. W.

Quine (who insisted on the thesis that only our “first-grade conceptual
system”—that is, physics—is in charge of describing reality as it really
is),29 generated a “puritanical” ontological attitude according to which
only scientifically describable phenomena are real.30 As a consequence,
scientific naturalists face what has been called a “location problem” or
“placement problem”31—that is, they have to show how to reduce the
features of the commonsense world to the scientifically acceptable fea-
tures or to show that that they are mere illusions. Unfortunately, as
Putnam sardonically put it once, “None of these ontological reductions
27. Wilfrid Sellars, “Philosophy and the Scientific Image of Man,” in Frontiers of Science
and Philosophy, ed. R. Colodny (Pittsburgh: University of Pittsburgh Press, 1962), 37–78.
On the influx of Husserl on Sellars, see Moran, Husserl’s Crisis, 292–293.
28. Wilfrid Sellars, Empiricism and the Philosophy of Mind (1956; Cambridge, Mass.: Har-
vard University Press, 1997), 83 (emphasis added).
29. W. V. Quine, Word and Object (New York: Columbia University Press, 1960), and
“Epistemology Naturalized,” in Ontological Relativity and Other Essays (New York: Co-
lumbia University Press, 1969), 69–90.
30. This term is due to Stephen Stich, Deconstructing the Mind (New York: Oxford Uni-
versity Press, 1996), 199.
31. Frank Jackson, From Metaphysics to Ethics (Oxford: Oxford University Press, 1998),
1–5; Huw Price, “Naturalism without Representationalism,” in De Caro and Macarthur,
Naturalism in Question, 71–105.
Unauthenticated
Introduction 11
gets believed by anyone except the proponent of the account and one
or two of his friends and / or students.”32
Putnamian liberal naturalism, instead, has a pluralistic attitude both
in ontology and epistemology. In his view, not all the real features of
the world can be reduced to the scientifically describable features, and
the natural sciences are not the only genuine source of knowledge to
which all the other apparent sources should hand over their epistemic
pretensions. Obviously, this approach gives rise to what could be called
a “reconciliation problem”—the problem of showing how the dif ferent
kinds of features of the world can all be real without conceptual ten-
sion or even contradiction.33 On this problem, Putnam has a number
of interest ing things to say.
In the last few years, he has started to make his reflections on liberal
naturalism explicit,34 but many impor tant, if scattered, reflections in
that direction can also be found in his previous works. Here is a list of
the most important features that, according to Putnam, a satisfying lib-
eral naturalism should encompass.
1. Liberal naturalism should incorporate scientific realism. It

should assume a realistic attitude toward the unobservable
entities postulated by our best scientific theories and consider
those theories as true (or as approximately true, pending the
appearance of better theories). This means that, in regard to the
realism / antirealism debate concerning science, liberal natural-
ists should assume neither an instrumentalist attitude nor a
quietist one (i.e., a noncommitting attitude regarding the
realism / antirealism debate on science).35
2. The different sciences should not be interpreted monistically, as if
they were one thing in content, method, and goals. From a philo-
sophical point of view, pluralism is the right intellectual attitude
in dealing with the sciences (but not only with the sciences).36
32. Hilary Putnam, “The Content and Appeal of ‘Naturalism’,” in De Caro and Macar-
thur, Naturalism in Question, 59–70, quotation from p. 62.
33. Mario De Caro, “Putnam’s Liberal Naturalism,” in Themes from Putnam, ed. Mi-
chael Frauchiger (Berlin: Ontos Verlag, forthcoming).
35. Putnam, “On Not Writing Off Scientific Naturalism,” and Chapter 5 here.
36. Putnam, “The Diversity of the Sciences: Global versus Local Methodological Ap-
proaches,” in Metaphysics and Morality: Essays in Honor of J. J. C. Smart, ed. Philip Pettit,
Unauthenticated
12 Introduction
3. It has to be acknowledged that both the scientific and the

commonsense view of the world are only approximately correct
and, consequently, are constantly revised. For Putnam, the key
concept in this regard is that of fallibilism, which he takes from
the pragmatist tradition: any of our beliefs could be wrong, but
(contrary to what the skeptic claims) we need context-specific
reasons to start doubting them.37
4. Contrary to a common opinion, not all objective knowledge
should be seen as object-based. In particular, there is no need of
assuming that mathematical or ethical statements refer to
peculiar, abstract objects (such as sets or the Good) in order to
be true or, in the case of ethics, for being better or worse than
other competing claims.
5. It is our epistemic duty to try to resolve the contradictions and
conceptual tensions between the commonsense and the scien-
tific worldviews, but science does not necessarily have the last
word here, even if it is frequently the case that it does. For
example, Copernican astronomy initially posed many problems
for common sense, and the first attempts to formulate the
differential and integral calculus raised so many contradictions
and inconsistencies that the whole enterprise looked common-
sensically unacceptable; but the attempt to resolve those incon-
sistencies led to enormous mathematical progress.
6. Liberal naturalism, like all other forms of naturalism, should
not make any supernatural assumptions. However, contrary to
another very common opinion and in the spirit of the prag-
matist tradition, according to Putnam this does not mean that
all religious experience should be condemned as irrational
or futile (not all religious people believe in divine dictation
or in divine properties that violate the laws of nature, for
example).38
Richard Sylan, and Jean Norman (Oxford: Basil Blackwell, 1987), 137–153; repr. “The Di-
versity of the Sciences,” in Words and Life, ed. James Conant (Cambridge, Mass.: Harvard
University Press, 1990), 463–480.
37. Putnam, Pragmatism: An Open Question (Oxford: Blackwell, 1995), passim.
38. Hilary Putnam, “Wittgenstein on Religious Belief,” in On Community, ed. Leroy
Rouner (Notre Dame, Ind.: University of Notre Dame Press, 1991), 56–75; repr. in Putnam,
Renewing Philosophy (Cambridge, Mass.: Harvard University Press, 1992), 134–157.
Unauthenticated
Introduction 13
7. As to the epistemological status of philosophy, it should be noted

that philosophy has a double face: one is the scientific face
(which interacts with the natural and social sciences), and the
other is the moral face (which “interrogates our lives and our
cultures as they have been up to now, and that challenges us to
reform both”).39 Both are essential. Pluralism plays a role also in
metaphilosophy, then.
8. The commonsense worldview is not in need of legitimation
from science, although commonsense beliefs that turn out to
be incompatible with scientific fi ndings have to be given up.
(Sometimes science does have the last word.) However,
phi losophers who claim that belief in the real existence of
colors or of solidity (for example, the real existence of pink
ice cubes) or the existence of referring is incompatible with
science are simply wrong. Color realism and realism about
representation40 are respectable scientific positions according
to Putnam, as well as commonsense positions, even if color
predicates and the notion of representation are not reducible
to predicates of physics. The placement problem is a false
problem. There is no need to reduce all the ordinary features
of the world to features describable in the vocabulary of the
exact sciences.
9. Regarding the reconciliation problem (how the commonsense
and the scientific views of the world can coexist without
one being prior on the other or delegitimatizing it), it should
be noted that, contrary to a view that is common nowadays in
many philosophical quarters, the world cannot be entirely
described by physics— obviously not by present physics, but
arguably not even by an ideal physics. In this sense, a favorite
example of Putnam’s is literary criticism: how could a physical
theory do ever so much as discuss the many aspects of Shake-
speare’s plays that sensitive readers and critics discern? As
Putnam wrote,
39. Putnam, “Science and Philosophy,” in De Caro and Macarthur, eds., Naturalism and
Normativity, 89–99.
40. Putnam cites Tyler Burge’s Origins of Objectivity (Oxford: Oxford University Press,
2010) in defense of the scientific legitimacy of realism about representation in the title essay
of this volume.
Unauthenticated
14 Introduction
The world cannot be completely described in the language

game of theoretical physics, not because there are regions
in which physics is false, but because, to use Aristotelian
language, the world has many levels of forms, and there is
no realistic possibility of reducing them all to the level of
fundamental physics.41
The consequence of this is “conceptual pluralism,” the necessity

of using a plurality of mutually irreducible but not incompatible
conceptual systems for dealing with the different levels of reality.
It is important to note, however, that conceptual pluralism must
not be confused with the different phenomenon of “cognitive
equivalence.”42 The latter is linked to the cases, not uncommon
in science, in which the same phenomena can be described with
apparently contradictory conceptual schemes that can indeed be
systematically translated into each other (such as when one
describes space-time points as individuals or as mere limits).
10. A satisfying liberal naturalist view should assume that among

the dif ferent levels of reality there is a relationship of superve-
nience. However, as Putnam writes in regard to human capaci-
ties (but these words could be applied to dif ferent cases),
there is no one simple answer to the question of whether
our agential capacities are locally supervenient (superve-
nient on just the relevant brain-states) or globally superve-
nient (supervenient on factors external to the brain, and
even to the organism), because it depends on which
agential capacities one is talking about, even if we restrict
the issue to perceptual capacities.43
11. Finally, a notion that one should use with caution if one does
not want to end up with the reconciliation problem is that of
41. Hilary Putnam, “From Quantum Mechanics to Ethics and Back Again,” in Philos-
ophy in an Age of Science, 51–71.
42. Sometimes Putnam instead of talking of “cognitive equivalence” talks of “concep-
tual relativity,” but this term could be misleading since it can be interpreted as an opening
to ontological relativity (which Putnam does not want to offer at all).
43. Hilary Putnam, “Reply to Stephen White,” European Journal of Analytic Philosophy 4,
no. 2 (2008): 29–32, quotation from p. 29.
Unauthenticated
Introduction 15
causation. According to Putnam, causation is an intentional,

context-relative notion—what the correct answer to the
question “What caused the phenomenon x?” is depends on the
specific interests of the inquirers. If I write a sonnet, what is
the cause of that action of mine? My love for poetry? Some
activations of my neurons? The specific cultural history of my
society? The education I received? The physical arrangement
of microparticles in my environment? All these potential
explanations point to dif ferent potential causes; and each of
these explanations can be correct (or incorrect) depending on
what interests one has when one asks the question “Why?”
Causation in fundamental physics is often simply not the
relevant notion.
3. Normativity
Normativity and ethics are subjects discussed in Chapter 1, while
Chapter 3 concerns the bearing of evolutionary theory on ethics, but
neither normativity nor ethics are large topics in the other chapters of
the present volume. However, Putnam has written extensively on them,
and a word on those writings may help to clarify the discussion in those
two chapters.44 Recurrent themes in Putnam’s past writings are:
1. That values grow out of human needs, and are historic products,
but are not therefore arbitrary or merely culturally relative. (See
Chapter 3.)
2. That neither epistemic values nor ethical values can be reduced
to a single principle. (In particular, “seek successful prediction”
and “try to falsify your theories” are completely insufficient by
themselves.)45
44. Among those writings, in addition to two volumes, The Collapse of the Fact / Value
Dichotomy (Cambridge, Mass.: Harvard University Press, 2002) and Ethics without Ontology
(Cambridge, Mass.: Harvard University Press, 2004), particularly useful for readers ap-
proaching this side of Putnam’s thought anew might be his essays on pragmatism and
ethics, particularly “Pragmatism and Moral Objectivity” and (with Ruth Anna Putnam)
“Dewey’s Logic: Epistemology as Hypothesis,” in Words and Life, ed. James Conant (Cam-
bridge, Mass.: Harvard University Press, 1994), 151–181 and 198–220, respectively.
45. Putnam argued this in “The Philosophers of Science’s Evasion of Values,” collected
in The Collapse of the Fact / Value Dichotomy.
Unauthenticated
16 Introduction
3. That there is a sometimes useful distinction between factual

judgments and value judgments, but philosophers have gone
astray by elevating it to the status of a metaphysical dichotomy.46
4. That both factual judgments and value judgments require
support by reasons (of both kinds). The notion of a reason does
not require the kind of reduction that hard scientific naturalism
seeks. (See Chapter 1.)
The aspect of Putnam’s point of view that may be least familiar is

his likening of epistemic values to aesthetic values. In his essay “The
Philosophers of Science’s Evasion of Values,” Putnam recalls a conver-
sation he had with Popper’s friend Jacob Bronowski in which Bronowski
said that he once told Popper, “You would not claim that scientists test
every falsifiable theory if as many crazy theories crossed your desk as
cross mine!”47 Epistemic values such as coherence, simplicity, and even el-
egance play an essential role in deciding which theories to test. But all
the familiar arguments for emotivism in ethics would, if they were right,
condemn judgments of coherence and elegance to the wastebasket of
mere subjectivity.
In order to appreciate this point, let us close by going back to a his-
torical case discussed earlier, that of Galileo.
It is known that, when he opted for Copernican astronomy over Ptol-
emaic astronomy, Galileo did not have any purely “factual” support in
his favor: only later did he refer to his astronomical observations (which,
by the way, at that time did not offer a final confirmation of Coperni-
canism), and to the famously wrong theory that the tides depend on
the rotation of the earth.
The real reason for which Galileo accepted Copernicanism was
metaphysical, and it incorporated a strong aesthetic reason, namely,
that the Copernican system was much simpler than the Ptolemaic
one. (“Simplex sigillum veri,” “Simplicity is the seal of truth,” was a
famous medieval adage that Galileo certainly accepted.) Accordingly,
in his famous Dialogue on the Two Main Systems of the World, he wrote
that, following Copernicus, he had become aware that the Ptolemaic
system was
46. Putnam argues this in The Collapse of the Fact / Value Dichotomy, 135–145.
47. Putnam, “Philosophers of Science’s Evasion,” 142.
Unauthenticated
Introduction 17
a monstrous chimera composed of mutually disproportionate

members, incompatible as a whole. Thus however well the astron-
omer might be satisfied merely as a calculator [whose only role is to
predict the future positions of planets], there was no satisfaction
and peace for the astronomer as a scientist [the scientist-philosopher,
who searches for the real laws of the universe]. And since he very
well understood that although the appearances might be saved by
means of assumptions essentially false in nature [false because dis-
proportionate and adverse to geometrical simplicity], it would be
very much better if he could derive them from true suppositions.48
And it is worth noticing that it was for the same metaphysical-aesthetic

reason—the preference of geometrical simplicity over complexity—that
later Galileo rejected Kepler’s correct idea of the elliptical orbits of the
planets.
The examples of the role that aesthetic factors play in science (and
in mathematics) are innumerable. With this sort of example in mind,
Putnam argues that values, including this sort of quasi-aesthetic eval-
uation, which obviously have a normative character, play a crucial role
also in science.
However, an objection to this argument that one might certainly ex-
pect is that one has to distinguish the context of discovery from the
context of justification; and that, in this way, one can separate what is
justificatorily irrelevant (such as Galileo’s aesthetic bias in favor of sim-
plicity) from the real factual content of a scientific theory. But Putnam
would dispute the claim that the reasons that justify accepting a theory
only depend on the factual content of the theory. If considerations of a
more holistic and even a “quasi-aesthetic” character, such as the simplicity
of the theory, are the reasons that, together with the observations, lead us
to accept certain theories and not even bother to test certain rivals, then if
those reasons must be regarded as purely “subjective”—that is, as not
reasons at all— our acceptance of the science we have must all be re-
garded as wholly subjective. And Putnam convincingly has claimed this
should be seen as a reductio of skepticism about epistemic values.
48. Galileo Galilei, Dialogo sopra i due massimi sistemi del mondo, 1632; published in En-
glish as Dialogue Concerning the Two Chief World Systems, Ptolemaic & Copernican, trans.
Stillman Drake (1953; Berkeley: University of California Press, 1967), 341 (emphasis added).
Unauthenticated
18 Introduction
Nowadays philosophers tend to be specialized in very few and fre-

quently very minor issues, which they scrutinize at the smallest degree.
Hilary Putnam does not belong to this group. His way of doing philos-
ophy rather resembles the way of operating of the thinkers of the classic
tradition. He only deals with big questions, in all philosophical fields,
and unceasingly tries to refi ne his answers to those questions and make
them consistent and harmonious. The general framework of his phi-
losophy is the attempt to reconcile the ordinary and the scientific
views of the world, without prioritizing one over the other, in a natu-
ralistic spirit respectful of the fundamental and irreducible role that
normativity plays in our lives.
We should all be grateful to Putnam not only for the innumerable
and invaluable contributions he has offered in many fields, but also
because he represents a shining example of what philosophy has been
in its most glorious moments and should always continue to be.49
49. I thank Massimo Dell’Utri, Andrea Lavazza, David Macarthur, Massimo Marraffa,
Matteo Morganti, Barry Smith, and Stephen White for many useful exchanges on the is-
sues dealt with in this introduction. A special thank goes to Hilary Putnam, with whom I
have had the honor to discuss these issues for several years.
Unauthenticated
D1
Naturalism
The terms “naturalism” and “realism” are useful (and even indispens-
able on occasion), but each of them has been applied to a wide range of
dif ferent and even incompatible positions. I very much like the term
“liberal naturalism,” which I first encountered in an important collection
of essays edited by Mario De Caro and David Macarthur entitled Nat-
uralism in Question. In my essay in that collection,1 I pointed out that
while Boyd, Casper, and Trout’s The Philosophy of Science admirably takes
the trouble to tell us what the editors mean by “naturalism” (and their
naturalism is certainly not “liberal”),2 their definition is disjunctive, and
thus offers us two possible meanings of “naturalism” rather than one.
According to Boyd and colleagues, naturalism is “[t]he view that all
phenomena are subject to natural laws, and / or [sic] that the methods
of the natural sciences are applicable in every area of inquiry,” and I
argued that both disjuncts are extremely vague. In what sense does a
“naturalist” have to believe that the “phenomenon” of Shakespeare’s
1. Hilary Putnam, “The Content and Appeal of ‘Naturalism,’ ” in Naturalism in Ques-

tion, ed. Mario De Caro and David Macarthur (Cambridge, Mass.: Harvard University
Press, 2004), 59–70.
2. Richard Boyd, Philip Casper, and J. D. Trout, eds., The Philosophy of Science (Cam-
bridge, Mass.: MIT Press, 1991).
21
Unauthenticated
22 LIBERAL NATURALISM AND NORMATIVITY
writing Julius Caesar was “subject to natural laws”? Is it enough to be-

lieve that writing that play didn’t violate any natu ral laws? Surely that
makes it too easy to be a “naturalist”! Or should one believe that in
our “first-grade” conceptual system (W. V. Quine)3—or in “the abso-
lute conception of the world” (Bernard Williams)4 —there is no fact of
the matter as to what a single sentence of that play means, but the pro-
duction of those marks on paper was a physical process whose physical
explanation exhausts what there is to be known about the play? Or
must one rather try to “naturalize” questions about meaning, and so
on? And what exactly does it mean to “apply” the supposed “methods
of the natural sciences” to Julius Caesar? In fact, the sort of hard natu-
ralism advocated by Boyd and Caspar’s volume chooses to say nothing
about such subjects as literary criticism, and even the “Cornell Moral
Realists,” a group of which Boyd is a leading member, regard talk of
obligation (“ought”) as belonging to the domain of the noncognitive.
What makes them moral realists is the odd claim that “good” refers to
a natural kind, on all four feet with biological kinds. Instead of seizing
either horn of this dilemma, De Caro and Macarthur pointed to and
celebrated the existence of a number of philosophical positions that are
“naturalist” in the sense that John Dewey, the most famous naturalist
philosopher of his period, was—namely, in the sense of rejecting all
appeals to supernatural entities in philosophy while simultaneously re-
jecting the positivist demand that aesthetic and ethical concepts be
reduced to the concepts of the natural sciences or expecting that they
could, or eventually will be, so reduced. Nor do they accept the posi-
tivist view that history will eventually become a “science.” In their
introduction to Naturalism in Question, De Caro and Macarthur
emphasized that the liberal naturalism they advocate doesn’t regard
normative utterances as somehow “second grade” or merely “expres-
sive,” but neither does it countenance a Platonic realm of normative
facts independent of human practices and needs. At the same time, it
does not countenance Moorean quasi-mystical faculties of moral intu-
ition. All this I like very much.
3. See W. V. Quine, Word and Object (New York: Columbia University Press, 1960), and
“Epistemology Naturalized,” in Ontological Relativity and Other Essays (New York: Co-
lumbia University Press, 1969), 69–90.
4. Bernard Williams, Descartes: the Project of Pure Enquiry (Harmondsworth: Penguin,
1978).
Unauthenticated
Naturalism, Realism, and Normativity 23
I don’t know whether Tyler Burge would ever employ the term “lib-
eral naturalism,” but I do know that his Origins of Objectivity is a great
book.5 It takes indispensability for good scientific explanations (the sci-
ences in question being vision science and related parts of psychology,
in particular) to be sufficient reason for accepting an unreduced notion
of representation in psychology and philosophy of mind, and I agree;
at the same time, what Burge writes about “naturalization projects in
philosophy” is surely in the spirit of “liberal naturalism”:
Promoters of “naturalizing” projects are often driven, I think, by

misconceptions of science. These misconceptions breed miscon-
ceptions of mind. The notion of representation—of reference or
attribution that can be correct or incorrect and that helps type-
individuate kinds of psychological states—is entrenched not only
in commonsense explanation but in scientific explanation in psy-
chology. There is nothing unnatural or supernatural about such
explanation. Some of the relevant psychology is well-supported,
mathematically rigorous, mature science. There is no basis, even
a prima facie one, to the worry that psychological notions are in-
vitations to mystery or miracle. Even if there were such basis, the
role that these notions play in powerful empirical science would
undermine it . . . I know of no good ground for thinking that . . .
[psychologists’] explanatory claims must be twisted into the mold
of biological or information-theoretic explanation, or any other
explanation in the natural sciences, in order to be explanatorily
successful.6
The reason I say that this is in the spirit of liberal naturalism is not
just Burge’s rejection of reductionism; some hard naturalists are quite
proud of being “nonreductive materialists” or the like. But even when
hard naturalists eschew the demand that reference and representation
talk be reduced to physical vocabulary, they feel the need to explain ref-
erence and representation in terms of “causal relations.”7 Even if the
5. Tyler Burge, Origins of Objectivity (Oxford: Clarendon Press, 2010).

6. Burge, Origins of Objectivity, 296–297.
7. Richard Boyd, “What of Pragmatism with the World Here?” in Reading Putnam, ed.
Maria Baghramian (Abingdon, United Kingdom: Routledge, 2013), 39–94.
Unauthenticated
explanations are, as I believe, pure hand-waving, the spirit, if not the

details, is obviously reductionist.
However, liberal naturalism, as represented by De Caro and Macar-
thur’s two volumes (the already mentioned Naturalism in Question and
Naturalism and Normativity),8 covers a wide range of positions, in-
cluding ones I disagree with. For example, one contributor to Natu-
ralism in Question, Huw Price, rejects the whole idea that reference is a
relation between linguistic items and worldly objects,9 an idea (he calls
it “representationalism”) that is an essential part of my own realism. An-
other contributor (and editor of the volume) advocates a version of
Wittgensteinian “quietism,”10 which I also reject.11 Thus, although I am
indeed a “liberal naturalist,” that label does not say very much about
my philosophy. To supplement it I need to say more—for example, I am
a liberal naturalist, but I am also a realist in my metaphysics and a re-
alist with respect to the normative. This essay will attempt to sketch
the outlines of a liberal naturalist metaphysically realist and normatively
realist position in philosophy.
Metaphysically Realist
Readers who know of my philosophy from Wikipedia may be surprised
to find me saying that my position is “metaphysically realist.” Had I not
consistently attacked metaphysical realism from “Realism and Reason”
(1978) on?12 Well, not exactly. In that lecture and for a number of years
afterward, I used “metaphysical realism” (sometimes with capital let-
ters) as a term for a specific position whose main feature was the insis-
tence that the world can be divided into (mind-independent) objects and
properties in exactly one way. But the use of that term was a mistake
on my part. It was a mistake because, although I repeatedly explained
8. Mario De Caro and David Macarthur, eds., Naturalism and Normativity (New York:
Columbia University Press, 2010).
9. Huw Price, “Naturalism without Representationalism,” in De Caro and Macarthur,
Naturalism in Question, 71–105.
10. See David Macarthur, “Putnam, Pragmatism, and the Fate of Metaphysics,” Euro-
pean Journal of Analytic Philosophy 4, no. 2 (2008): 33–46.
11. Hilary Putnam, “Reply to David Macarthur,” European Journal of Analytic Philosophy
4, no. 2 (2008): 47–49.
12. Hilary Putnam, “Realism and Reason,” in Meaning and the Moral Sciences (London:
Routledge and Kegan Paul, 1978), 123–138.
Unauthenticated
what I meant by it, based on the natural understanding of the phrase

“metaphysical realism,” it refers to a broad family of positions and not
just to the one position I used it to refer to. In effect I was saying that by
refuting the one philosophical view I called by that name I was ipso
facto refuting anything that deserved to be called metaphysical realism,
and that was not the case.
In addition, in my Dewey Lectures,13 which were the best known
source for my post–“internal realist” views, I was still under the influ-
ence of Wittgensteinian quietism; although this is an aspect of Witt-
genstein’s thought that I grew increasingly unsympathetic to, it was not
until my closing address to my eightieth birthday conference in Dublin
that I was willing to say, in those words, that I am a metaphysical re-
alist without capital letters, a realist in my metaphysics, but not a “meta-
physical realist” in the technical sense I gave to that term in “Realism
and Reason” and subsequent publications.14 At the same time I was also
willing to say,
Let me say for the record that I utterly and totally reject all ver-
sions of the “end of philosophy” story, whether they come from
Wittgensteinians, Rortians, Heideggerians, Derridians, or who-
ever. Philosophy was not a mistake; not in Parmenides’s time, not
in Plato’s time, not in Aristotle’s or Descartes’s or Hume’s or Kant’s
times, and not in our time. As Etienne Gilson put it, “philosophy
always buries its undertakers.”15
But what is it to be a metaphysical realist? In an elegant paper titled

“Confessions of a Hard-Core, Unsophisticated Metaphysical Realist,”16
Maudlin describes his reaction to first reading the lecture in which I
first used the term “internal realism.” In that lecture I had written:
13. Hilary Putnam, “Sense, Nonsense, and the Senses: An Inquiry into the Powers of
the Human Mind,” Journal of Philosophy 91 (1994): 445–517; reprinted in Hilary Putnam,
The Threefold Cord: Mind, Body and World (New York: Columbia University Press, 1999),
3–70.
14. Hilary Putnam, “From Quantum Mechanics to Ethics and Back Again,” in Bagh-
ramian, Reading Putnam, 19–36.
15. Putnam, “From Quantum Mechanics to Ethics and Back Again,” 68.
16. Tim Maudlin, “Confessions of a Hardcore, Unsophisticated Metaphysical Realist,”
in The Philosophy of Hilary Putnam, ed. R. E. Auxier, D. R. Anderson, and L. A. Hahn (Chi-
cago: Open Court, 2015), 487–501.
Unauthenticated
The most important consequence of metaphysical realism is that

truth is supposed to be radically nonepistemic— we might be “brains
in a vat” and so the theory that is “ideal” from the point of view of
operational utility, inner beauty and elegance, plausibility, sim-
plicity, “conservatism,” etc. might be false. “Verified” (in any oper-
ational sense) does not imply “true,” on the metaphysical realist
picture, even in the ideal limit.
It is this feature that distinguishes metaphysical realism, as I am
using the term, from the mere belief that there is an ideal theory
(Peircean realism), or, more weakly, that an ideal theory is a regu-
lative ideal presupposed by the notions “true” and “objective” as
they have classically been understood. And it is this feature that I
shall attack.17
And Maudlin’s reaction:
We now have something quite concrete to discuss. The metaphys-

ical realist thinks that an operationally ideal theory, a theory ideal
as far as we can tell, might actually be false. This is a thesis that ap-
pears to be clear and concise. It is also a thesis that I take to be
true and took to be true as a graduate student. So the most en-
during impact “Realism and Reason” had on me, despite its in-
tent, was to make me a confirmed metaphysical realist.18
The thesis that a theory “ideal as far as we can tell” might actually be
false simply amounts to the claim that a statement that we are episte-
mically entitled to accept as part of our best theory of the world might
be false, and there might be no way of verifying that it is false. And I
agree with Maudlin that there are such statements.
Hence, if we understand metaphysical realism as he does, as simply
making that claim, metaphysical realism is true. For example, there
might be very strong reasons to believe that there are intelligent extra-
terrestrials, including radio signals or whatever, but the signals could
be the product of quantum mechanical fluctuations of immensely low
probability. And metaphysical realism in Maudlin’s sense is part of my
17. Putnam, “Realism and Reason,” 123–138. The passage quoted is at p. 125.
18. Maudlin, “Confessions,” 485.
Unauthenticated
current view. Truth is not the same thing as warranted assertability

under ideal conditions. “Realism and Reason” was wrong.
But a responsible metaphysical realist needs to say something
about what truth is and not simply what it isn’t, and what follows is an
attempt to do that, to say something (of course, not every thing) about
what truth is.
Truth: Tarski’s Contribution and Its Interpretation

In “The Concept of Truth in Formalized Languages,”19 Tarski immea-
surably advanced our grasp of the formal logic of the predicate “True
sentence of L,” where L is a formalized language, and by doing so he
provided powerful new tools for mathematical logicians (something that
is underappreciated by philosophers) as well as material for philosophers
to assimilate and interpret. The tools for mathematical logicians that I
refer to include techniques that were later adapted to formalize the
statement that a set is defi nable over a given collection of sets with
parameters from that collection and quantifiers ranging over that col-
lection (tools that Gödel [1940] employed in the construction of the
hierarchy of “constructible sets” that he used to prove the consistency
of the Axiom of Choice and the Generalized Continuum Hypothesis20).
These tools have been employed in many contexts ever since: the ma-
terial for philosophers to assimilate and interpret includes the T-schema
or truth schema in section l of “The Concept of Truth in Formalized
Languages” (not to be confused with Convention T), Convention T,
and the “truth definitions,” as they have come to be called, that are re-
quired to meet that criterion of adequacy. (In both the German and the
original Polish of “The Concept of Truth in Formalized Languages,”
“Convention T” is called a “criterion” and not a “convention.” “Con-
vention T” comes from the 1956 English translation, T being the first
letter of the English word “true” and W being the fi rst letter of the
19. The original publication was in Polish in 1933, and the German translation appeared
in 1935. The 1956 English translation was reprinted as Alfred Tarski, “The Concept of
Truth in Formalized Languages,” in Logic, Semantics, Metamathematics: Papers from 1923
to 1938, ed. John Corcoran, trans. J. H. Woodger (Indianapolis: Hackett, 1983 [1956]),
152–278.
20. Kurt Gödel, The Consistency of the Continuum Hypothesis, Annals of Mathematics
Studies 3 (Princeton, N.J.: Princeton University Press, 1940).
Unauthenticated
German word “wahr.”) Although a minority of philosophers, including

Popper, saw “The Concept of Truth in Formalized Languages” as
putting forward a correspondence theory of truth, other philosophers
have linked it to “deflationary” theories of truth, while the particular
techniques Tarski employed led Hartry Field (correctly in my opinion)
to see Tarski as fi nding an intimate link between reference (or, in
Tarski’s—and Field’s—terminology, “denotation”) and truth,21 an in-
sight that Field unfortunately linked to an unworkable “naturalization
project” (to use Burge’s term) and later abandoned completely. The next
two sections elaborate on these remarks.
Tarski’s Criterion of Adequacy

Following Tarski, let us pretend that we have a particular interpreted
formal language in mind, and let us call it Bob. I choose an obvious
proper name, Bob, and not L, because L looks like a variable— a vari-
able for quantifying over formal languages— and it is essential to Tar-
ski’s approach that truth definitions are constructed one language at a
time. Tarski (rightly) does not attempt to construct a definition of “true
in L” when L is a variable over all languages, and doing so would imme-
diately lead to paradoxes. We will assume that Bob has a finite number
of primitive descriptive predicates and individual constants, and so on.
To avoid any appearance of circularity or begging the question, we will
also assume that none of these is interpreted as referring to semantic
properties or entities, such as truth or reference or “correspondence” or
facts or states of affairs. Let Meta-Bob be a second formal language
that contains the whole vocabulary of Bob, one in which it is possible to
quantify over inscriptions (sequences of symbols) in the language
Bob. In addition, permit Meta-Bob to be set-theoretically more pow-
erful than Bob in the ways needed for Tarski’s purposes, but without
having any primitive descriptive predicates or names other than those
of Bob itself. A “truth predicate” for Bob is a formula FORM(x) of
Meta-Bob with one free variable, say x. I avoid the custom of abbrevi-
ating the formula as “True-in-Bob” (or even worse, as “True-in-L”)
because doing so obscures the very issues I wish to raise. Of course, in
21. Hartry Field, “Tarsky’s Theory of Truth,” Journal of Philosophy 69, no. 13 (1972):
347–375.
Unauthenticated
ordinary language there is such a predicate as “true in Bob,” with no

hyphens, but that predicate involves the ordinary language notion of
truth, which Tarski claimed is inconsistent. That is the reason why we
need a formal replacement for it. “Defining truth in Bob,” in Tarski’s
sense, means (1) finding a FORM(x) that is true of all and only those
inscriptions x in Bob that are true in Bob and (2) reformulating this re-
quirement in a way that does not involve using the everyday language
word “true” or any related notion. Tarski’s solution to this latter desider-
atum, defining “material adequacy” of a truth definition without using
the word “true,” is the famous Convention T, which we state by saying:
(T) FORM(x) is a materially adequate truth predicate for Bob just

in case the result of replacing s by an arbitrary sentence of Bob and
s* by a structural-descriptive name of that sentence in:
s ⇔ FORM(s*)
is a theorem of Meta-Bob.
Tarski’s notion of a “structural-descriptive name” of an inscription

is easiest to explain via an example: if s is “Snow is white,” then the
formal name s* is “The letter capital S followed by the letter n followed
by the letter o followed by the letter w followed by space followed by the
letter i followed by the letter s followed by space followed by the letter w
followed by the letter h followed by the letter i followed by the letter t
followed by the letter e.”22
Thus, pretending that “Snow is white” is a sentence of Bob: Conven-
tion T requires that FORM(x) be defined in Meta-Bob in such a way that
“Snow is white” ⇔ FORM(the letter capital S followed by the letter n
followed by the letter o followed by the letter w followed by space fol-
lowed by the letter i followed by the letter s followed by space followed
by the letter w followed by the letter h followed by the letter i fol-
lowed by the letter t followed by the letter e) is provable from Meta-Bob’s
axioms.
Readers who have gone through life believing that Convention T is
a Davidsonian T-sentence or that it is an instance of “disquotation”
will be shocked to observe that “true” does not occur in Convention T
22. Tarski, “Concept of Truth,” 57.
Unauthenticated
(although abbreviating FORM(x) as “True-in-Bob(x)” can make this

hard to see, which is why I didn’t do that)!
Field’s Reconstruction of Tarski: The Link between

Reference and Truth
In his deservedly often-cited “Tarski’s Theory of Truth,”23 Hartry Field
argued that Tarski failed to reduce the notion of truth “to other non-
semantical notions” but that what Tarski actually did was reduce the
notion to other semantical notions.
This is an interest ing reading, although I am not certain that Tarski
had any such aim. What supports Field’s reading is that, in “The Con-
cept of Truth in Formalized Languages,” Tarski begins by describing
the ordinary language notion of truth, and that this is where Tarski puts
forward the T-schema.24 An instance of the latter is the famous “Snow
is white” is true if snow is white; what perhaps counts against it is that at
the end of that section (section l) Tarski seems to give up the project of
clarifying the ordinary language notion of truth entirely:
If these observations [concerning the inconsistency of the ordinary

language notion—HP] are correct, the very possibility of a consistent
use of the expression “true sentence” which is in harmony with the laws
of logic and the spirit of everyday language seems to be very question-
able, and consequently the same doubt attaches to the very possibility of
constructing a correct definition of this expression.25
In any case, one of Tarski’s purposes, which he achieved, was certainly

to find a way to define in Meta-Bob a predicate that has the same exten-
sion as the ordinary language semantical predicate “true in Bob” (no
matter which interpreted formalized language “Bob” might be); as I put
it once, the predicate(s) FORM(x) we get by Tarski’s method do not ex-
press the same concept as the semantical predicate “true in Bob.”26 In
23. Hartry Field, “Tarsky’s Theory of Truth.”

24. Schema (2) in Tarski, “Concept of Truth,” 155.
25. Tarski, “Concept of Truth,” 165 (emphasis in original).
26. Hilary Putnam, “A Comparison of Something with Something Else,” New Literary
History 17, no. 1 (1985): 61–79; reprinted in Hilary Putnam, Words and Life, ed. James
Conant (Cambridge, Mass.: Harvard University Press, 1994), 330–350.
Unauthenticated
Fregean language, FORM(x) has the right Bedeutung, but not the right
Sinn. But by giving two of his papers the titles “The Concept of Truth
in Formalized Languages” and “The Semantical Concept of Truth” [em-
phasis added], Tarski certainly gave the impression that his formal re-
sults somehow do capture the Sinn of “true in Bob” (that is, that they
capture the concept of truth) and not just the extension of that everyday
language semantical predicate.
A Technical Section
To explain all this, I need to say something about how one defi nes
FORM(x) following Tarski’s recipe. Simplifying outrageously by pre-
tending that all the predicates of Bob are monadic (the fact that this is
in general not the case is the reason that all the mathematical ingenuity
in Tarski’s paper was required), the easiest way to do this is to imagine
that we construct Meta-Bob in two stages. I shall also, less outrageously,
simplify by assuming Bob contains no names, only predicates, and (since
the universal quantifier can be defined in terms of the existential quan-
tifier [(x)F(x) ⇔ ¬(Ex) ¬F(x)]), I shall assume the primitive notation of
Bob has only the existential quantifier.
Stage 1. First, (if necessary) we add to the vocabulary of Bob enough
apparatus to do logical syntax (but not semantics!), that is, logical means
for quantifying over strings of symbols, such as the letter capital S fol-
lowed by the letter n followed by the letter o followed by the letter w,
and over finite sequences of such strings. We also ensure it is possible
to define such syntactic predicates as “x (a string of symbols) is a wff
(well-formed formula) of Bob” and “x (a string of wffs) is a proof in Bob,”
“x is a theorem of Bob,” and so on.
Stage 2. We temporarily add to Meta-Bob a primitive predicate Ref(x, y),
which we will read as “x (a wff of Bob with one free variable) refers to y.”
The interpretation of Ref should be clear from the following example:
if x is “F(v)” and a is an entity in the range of the quantifiers of Bob,
then [Ref(x, a) ⇔ F(a)].
Next, we start to create an inductive definition of “Ref” by adding to
the axioms (the axioms of Bob plus whatever axioms we need to do syntax
and set theory in Meta-Bob) as follows: for each atomic predicate—say,
“Glub(x)”—of Bob, we add as an axiom ( y)[Ref(“Glub” followed by
“(” followed by x followed by “),” y) ⇔ F( y)]. In words, “Glub(x)” refers
Unauthenticated
to y if and only if Glub( y), and similarly for all the other atomic predi-
cates of Bob. Note that there are only fi nitely many atomic predicates
in what Tarski calls “formalized languages,” and hence only fi nitely
many of these axioms. These axioms are the basis clauses of the induc-
tive definition.
Stage 3. We add further axioms recursively extending the definition
of “Ref” to molecular formulas, for example, an axiom saying that if x
is the result of writing the disjunction symbol “∨” between inscriptions
w and z, then Ref(x, y) ⇔ Ref(x, w) ∨ Ref(x, z), and an axiom saying that
if x is the result of writing the negation symbol “¬” before w, then
Ref(x, y) ⇔ ¬Ref(x, w).
And (important!) we add to the axioms the statement, duly formal-
ized, that for all w, if v is a variable, and w is z preceded by an existen-
tial quantifier (Ev), and z contains no free variables dif ferent from v,
then {Ref(w,1) ⇔ (Ex)Ref(z, x)} & {(Ref(w, 0) ⇔ ¬(Ex)(Ref(z, x)}. This
amounts to adopting the convention that a sentence (a wff with no free
variable) “refers to” 1 if it is true and to 0 if it isn’t.
Note that even if Bob’s primitive predicates were all monadic, Bob
would in general still contain polyadic defined predicates, for example,
“F(x) ∨ y = 0” and “F(x) & ¬F( y)”; therefore, we need a further clause in
our inductive definition for existential quantification of a polyadic for-
mula, but formulating such a clause requires more of Tarski’s technique
than we shall explain. Suffice it say that this technique involves con-
structing Ref as a relation between formulas and sequences (in fact, even-
tually constant infinite sequences) of objects.
Stage 4. Now comes the crucial stage. Given that “Ref” has been
inductively defined, it is a straightforward application of techniques
familiar to logicians from Frege on to turn the inductive defi nition
of “Ref(x, y)” into an explicit definition of a two-place predicate, say,
SAT( y, x) [read: “y satisfies x,” employing Tarski’s terminology]27 pro-
vided Meta-Bob has strong enough set theoretic axioms; if it doesn’t, we
just add them. And voilà, we have a predicate of Meta-Bob whose defi-
nition involves only the atomic predicates of Bob and (if necessary) the
27. Instead of speaking of a formula as referring to (a sequence of) objects, Tarski speaks
of (the sequence of) objects as satisfying the formula; I have imitated Tarski’s language here,
which is why “Ref(x, y)” has suddenly become “SAT( y, x).” Note that “Ref(x, y)” gets
dropped from the notation of Meta-Bob in the course of my “stages.”
Unauthenticated
primitive predicates of our favorite set theory but no nonlogical predi-

cates that weren’t already in Bob— a predicate that provably satisfies all
the clauses in the inductive definition of Ref!28
Therefore, we can drop Ref from our list of primitive predicates,
drop the axioms that contained it, and use “SAT( y, x)” instead of
“Ref(x, y),” since we know that all the old axioms for Ref are theorems
of Meta-Bob, once we simply replace “Ref(x, y)” with “SAT( y, x).” And
again— voilà!—it can be verified that if we take FORM(x) to be “x is a
wff with no free variables & SAT(1, x),” then for each sentence of Bob,
the result of substituting it for s and its structural-descriptive name for
“s*” in “FORM(s*) ⇔ s” will be a theorem of Meta-Bob. Convention T
is satisfied!
What Field Observed

What Field observed, and what I hoped to bring out clearly by breaking
up the definition of SAT (Tarski’s formal predicate for “satisfaction”)
into stages in the way I did, was that (1) the “truth predicate” [in the
case of Bob, the predicate FORM(x)] is defined in terms of the “refer-
ence predicate” SAT, and (2) the construction can be said to mimic the
process of defi ning truth in Bob in terms of reference in Bob (i.e.,
reference restricted to formulas of Bob); (3) reference in Bob (Field’s
“denotation”) is defined—or rather SAT, a relation coextensive with the
converse of reference in Bob is defined—by turning an inductive defi-
nition into an explicit definition by well-known means. Finally, (4) the
basis clauses in the inductive definition “define” reference in the case of
atomic (undefined) predicates or rather specify the extension of reference
in the case of such predicates by a finite list of cases. From this, Field
concluded that Tarski reduced reference to what Field called “primitive
denotation” (denotation for atomic predicates) but that he failed to re-
duce “primitive denotation” to nonsemantical notions (which, on Field’s
reading of Tarski’s article was Tarski’s aim), but only listed the exten-
sions of the primitive predicates. Thus, according to Field, what re-
mains to be done after Tarski is to “naturalize” primitive denotation,
that is, to defi ne reference for atomic predicates in terms acceptable to
28. Note that inductive defi nition depends only on the logical and syntactical axioms we
added, but not on any “semantical” axioms!
Unauthenticated
a materialist philosopher. A finite list of clauses such as “in Bob, ‘w’

followed by ‘h’ followed by ‘i’ followed by ‘t’ followed be ‘e’ refers to all
and only white objects” is not a reduction at all, let alone one accept-
able to a “physicalist” like Field.
Philosophical Morals of the Above

(1) Apart from section l, which ends on a massively pessimistic note,
there is no attempt in Tarski’s “The Concept of Truth in Formalized
Languages” to reduce any semantical notion to a nonsemantical notion
(unless it be a failed attempt, which is what Field thought). What there
is, is a defi nition of a predicate FORM(x) that is coextensive with
“true in Bob,” but one that defi nes a property the sentence “snow is
white” has, for example, in a possible world in which snow is white and
the word “white” means black! (This is so because nothing in the defi-
nition of FORM(x) refers to the use or meaning of expressions.) As a
conceptual analysis of “true,” the predicate FORM(x) fails miserably.29
People would not have had so much difficulty in grasping this undeni-
able fact if Tarski had titled his conception “The Look No Semantics
Conception of Truth.”30 Of course, this is not to impugn the mathemat-
ical importance of the “material equivalence”—by which Tarski must
mean the coextensiveness— of “true” (in the relevant language Bob) and
the truth predicate [our FORM(x)] or the philosophical significance of
Tarski’s work, which can only be extracted if we are clear on what it
did and what it did not accomplish.
(2) “Correspondence theory of truth” is the name traditionally given
to theories of the form: “a statement [or thought or belief or proposition,
etc., depending on the particular philosopher] is true if and only if the
statement [or thought, etc.] corresponds to a fact [or to reality or to some
appropriate piece of reality, depending again on the particular philoso-
pher’s metaphysical views].” The correspondence theory is discussed in
section 1 of Tarski’s article (where the T-schema [not Convention T!] is
stated), but the section ends on a negative note about the whole project
of clarifying the notion of truth in everyday language, as we noted
29. In 1953, Carnap suggested to me a way of meeting this objection. I describe Carnap’s
objection (which depended on defining Bob by Bob’s “semantical rules”) and show why it fails
in Hilary Putnam, Representation and Reality (Cambridge, Mass.: MIT Press, 1988), 61–67.
30. I put it this way in “Comparison of Something with Something Else.”
Unauthenticated
above. However, Convention T itself does not (as I have emphasized)

contain the word “true.” FORM(x) is a one-place predicate of Meta-Bob,
and we have specified that Meta-Bob does not contain any words like
“true” nor any words like “fact.” Likewise, Convention T does not con-
tain any words like “true” nor any words like “fact.” Thus there is no
way in which Meta-Bob, or Convention T, or the “truth predicate”
FORM can even express the correspondence theory of truth.
(3) “Deflationary theory of truth” is the name given to the more re-
cent, twentieth-century theory that the notion of truth is wholly captured
by one or the other of the two following disquotation principles:
(D) To call a statement (or sometimes a “proposition,” rarely a “sen-

tence” as in Tarski’s famous article) “true” is simply to affirm the
statement.
OR SOMETIMES,
(D′) The statement “S is true”31 is equivalent to the statement S.
Since Convention T does not mention truth, as I have been empha-

sizing, it obviously does not state either D or D′. But it is plausible that
it presupposes D′ for the following reason: Tarski’s reader is supposed to
see that if all of the conditionals that are required to be theorems of
Meta-Bob by Convention T are theorems—and if we assume that the the-
orems of Meta-Bob are true, which cannot be taken for granted because
a mathematical theory may contain false statements without being
inconsistent—then each sentence of the form
FORM(s*) ⇔ s
is true— and, using the T- Schema
true(s*) ⇔ s,
it follows that
FORM(s*) ⇔ true(s*).
Thus, it is plausible that the disquotation principle is presupposed

by Tarski’s claim that Convention T is a correct (“accurate,” in the
31. A technical problem: in D′ is S a variable over statements? Or is D′ to be understood

with some sort of systematic ambiguity? The literature discusses this problem extensively,
and there are dif ferent proposals.
Unauthenticated
original Polish version) condition for the “material adequacy” of a for-

mula like FORM(x) as a truth predicate for Bob. In any case, the idea of
disquotation easily arises from a study of Tarski’s great paper. But there
is all the difference in the world between accepting a disquotation
principle and accepting the claim that such a principle captures com-
pletely what one has to know about truth. The latter is the thesis of
deflationism. I conclude that Tarski is not committed to that thesis any
more than he is committed to the correspondence theory of truth.
(4) Just as it is plausible to see a disquotation principle as presupposed
by Convention T, even if Tarski did not state one, it is plausible to see
the fact that the extension of “true in Bob” is determined by the extension of
“denotes in Bob” as driving the entire strategy of defining the desired
truth predicate [FORM(x)] inductively in terms of a predicate [in our
simplified version of Tarski, above, SAT( y, x)] that is constructed to
have precisely the extension of the everyday language predicate “x refers
to y in Bob.”
In sum, and this is something I regard as of great importance, Tarski’s
formal methods intuitively draw on and presuppose not just one property of
truth, the T- schema, or disquotation, but on that property and the further
property that the extension of “true” depends on the extension of “refers.” The
concepts of truth and of reference are intimately related, and his entire proce-
dure exploits the relation, as Field saw in 1972.
Realism Is Incompatible with Deflationism

If one says “Asteroids [or daisies or marsupials] exist,” but one’s account
of what it is to understand these speech sounds (or if one writes it instead
of saying it out loud, if one’s account of what it is to understand such a
string of symbols) does not mention any connection whatsoever between
those “vocables” (or those “strings”) and asteroids or daisies or marsu-
pials or objects and properties in terms of which such entities can be
described, then I, for one, fail to see how what one says can be under-
stood in a realist way. This is a point I debated with Michael Devitt32 as
well as with Rorty,33 and I will not repeat all that here. What follows is
32. See Hilary Putnam, “Comment on Michael Devitt,” in Baghramian, Reading Putnam,
121–126.
33. See Hilary Putnam, “Richard Rorty on Real ity and Justification,” in Rorty and His
Critics, ed. Robert Brandom (Oxford: Blackwell, 2000), 81–87.
Unauthenticated
addressed to a reader who “gets it” and agrees. Of course, a nonrealist

can be a deflationist and simply refuse to understand sentences about
such things in a realist way. For a logical positivist to know the meaning
of a sentence is just to know its method of verification, and the method
of verification is to be described in terms of tests we can perform.
Moreover, saying that certain sentences are causally connected to as-
teroids (or daisies or marsupials) isn’t enough to capture the way in
which truth-apt assertions about real objects relate to the world. If I say
that there are marsupials in Australia, I intend my utterance to be re-
lated to marsupials and not to anything else that the event of my making
that utterance may have been caused by, for example, textbooks or zoos.
In short, when I say it, I am referring to marsupials, and that fact is not
captured by pointing out that my saying it was causally connected to
marsupials.
If this is often underappreciated by deflationists, it seems to me that
missing it comes from missing the point with which I closed the pre-
vious section and will repeat here:
Tarski’s formal methods intuitively draw on and presuppose not just one
property of truth, the T-schema, or disquotation, but on that property and the
further property that the extension of “true” depends on the extension of “re-
fers” (and on the possible extensions of “refers,” if the logical vocabulary includes
modal operators.34 Tarski did not consider such languages.) The concepts of
truth and of reference are intimately interrelated.
If deflationists regularly fail to mention the interdependence of truth
and reference, they do, however, recognize the need for some account
of meaning or at least of the sameness of meaning. After all, I can speak
of true sentences in a language that is not properly contained in my own
language. The sentence “ ‘Schnee ist weiss’ is true if and only if Schnee
ist weiss” is not a well-formed sentence in either English or German.
The standard form of disquotation in this case (a generalization of
34. If a modal primitive is added to the language, say the symbol ◊, then the appro-
priate clause will read: a satisfies ◊F just in case ◊(a satisfies F). Here is a word example:
take F(x) to be (Ey)(x loves y). Interpret ◊ as physical possibility (or, alternatively, so cio-
log ical possibility), and take a to be Alice. Then “a satisfies ◊F” says that Alice satisfies “it
is possible x loves somebody,” “◊(a satisfies F)” says that it is possible that Alice satisfies “x
loves somebody,” and these two formulas have the same truth condition, namely, that in some
possible world there is a person whom Alice loves. In mathematical jargon, “satisfies” com-
mutes with ◊. [In my view, “in some possible world there is” means that it is possible that
there is a world in which there is; modal logic does not presuppose the actual existence of
possible worlds.]
Unauthenticated
Tarski’s T-schema) is to say that if I am using English (or a formalized

version thereof) as a metalanguage (for a part of German that is free of
semantical words and includes the sentence “Schnee ist weiss”), then
the appropriate T-sentence is:
“Schnee ist weiss” is true in German if snow is white.
And more generally, for any sentence s in the part of German in

question,
(T) “s” is true in German if . . .
where the three dots are to be replaced by the translation of the sen-
tence s in English.
That the notion of translation is needed for disquotation and therefore
needed by deflationists (since their thesis is that grasp of disquotation is
all that is needed for an understanding of truth) is widely recognized. But
what I have not seen discussed by deflationists, let alone taken seriously,
is the thought that translating sentences presupposes knowing what their de-
scriptive constituents refer to. It is an illusion that disquotation does not
presuppose the relation of reference.
The idea that one can speak of translation as if it didn’t involve
reference can’t, of course, just stem from verificationism, as it did in
Carnap, although Horwich, a leading deflationist, once suggested
that sameness of meaning is just sameness of “use” and coupled this
with an explic itly verificationist account of “use.” He wrote, “The
communal disposition to use a word in a certain way should not be
regarded as simply the disposition to treat certain sentences as defi-
nitely and permanently acceptable and others not. In addition, there
are dispositions to sanction various levels of confidence [cashed out as
“betting behavior”] in the truth of certain sentences—where the ap-
propriate degrees of belief are a function of observable circumstances.”35
Degrees of confi rmation were also identified with betting behav ior
by Carnap.
35. Paul Horwich, “Wittgenstein and Kripke on the Nature of Meaning,” Mind and
Language 2 (Summer 1990): 105–121 (emphasis added).
Unauthenticated
Field (even after his conversion to deflationism) is no verifica-

tionist. But given that Field was a student of Quine’s (as well as of
myself ), I suspect that Quine’s account of “radical translation” in
Word and Object may well have influenced him, and in the next sec-
tion I shall say why I think Quine’s account of radical translation was
mistaken (something Burge does in detail in chapter 7 of Origins of
Objectivity).
Quine’s Unrealistic Account of Linguistic

Development and Translation
In Quine’s view, originally spelled out in Word and Object and devel-
oped and sometimes modified subsequently (in ways that do not affect
the criticisms that follow), the translator who has no prior knowledge
of the semantics of the language she is trying to figure out and chil-
dren who are learning their first language have nothing to go on ex-
cept sounds that their informants (parents, in the case of the child) are
disposed to produce in response to stimulations of their nerve endings
(their “speech dispositions,” a concept Quine takes from Skinner’s be-
haviorist psychology, as chapter 1 of Word and Object makes clear). To
describe one of those dispositions as, in Quine’s famous example, “a dis-
position to say ‘gavagai’ when the speaker sees a rabbit,” as opposed to “a
disposition to say ‘gavagai’ when the speaker sees a collection of unde-
tached rabbit parts” is regarded as an arbitrary imposition. The child,
according to Quine’s developmental account,36 does not discriminate
rabbits as opposed to undetached rabbit parts, or “rabbithood-
exemplified,” in any determinate way, and the linguist’s assumption that
her “jungle language” informant does so is likewise unwarranted.
It is easy to miss the fact that a very substantial position on the psy-
chology of perception, child development, and language learning is pre-
supposed by this Quinean account, one that supports Quine’s whole
indeterminacy thesis. As Burge points out, in Quine’s picture the sci-
ences of linguistics and psychology float free of connections with bio-
logical accounts of the relations between the basic needs and activities
36. W. V. Quine, “Speaking of Objects,” Proceedings and Addresses of the American Philo-
sophical Association 31 (1958): 5–22; reprinted in W. V. Quine, Ontological Relativity and Other
Essays (New York: Columbia University Press, 1969), 1–25.
Unauthenticated
of animals and the ecological facts about their environments.37 That

picture also ignores psychological explanations of perception, which
rely on biological accounts and feedback into them. No biological ac-
count describes the environment as containing undetached rabbit parts,
and undetached rabbit parts do not figure in any scientific explanation
in perceptual psychology either. When I say this, I am not turning my
back on “liberal naturalism” and opting for “scientific naturalism”; I am
rather opting for a liberal naturalist understanding of what natural-scientific
explanations are, one that takes seriously what the natural sciences ac-
tually do, and does not require them to conform to “physicalist” scru-
ples (by avoiding notions like representation). Although Quine may
have abandoned logical positivism, his insistence in Word and Object that
the account of “speech dispositions” begin with proximal stimuli (“nerve
endings”) and his view that children cannot be said to perceive objects
until they have learned the apparatus of quantification are reminiscent
of the positivist picture, in which a “constructional system” is used to
interpret bare sense data. There is a difference between linguistic ref-
erence, which presupposes complex cognitive capacities, and primitive
perceptual representation, but the idea that organisms cannot discrim-
inate objects prior to acquiring those cognitive and linguistic capacities
is misguided. Linguistic reference grows out of perceptual representa-
tion, as Burge rightly argues, and if we can accept that, we need not be
frightened into denying that truth of linguistic representations is a form
of accurate representation. Moreover, if we see reference as growing out of
perceptual transactions between organisms and things in their environ-
ments, we also won’t be tempted to deny that reference is a relation
between words and things, as Sellars at times did. Deflationism about
reference typically goes with deflationism about truth. If saying that
representation is a relation between organisms (and states of organisms
and, derivatively, bits of language) and real things, properties, and events
is “representationalism,” then representationalism is no sin!
Normativity in Ethics
When philosophers teach ethics courses, they naturally emphasize the
places at which the major theories and thinkers disagree. Less often do
37. Burge, Origins of Objectivity, 211–212.
Unauthenticated
they try to see how insights from dif ferent schools of thought can be
combined. Yet it is especially impor tant to do that in ethics because
when the subject matter is how to live, as John Dewey emphasized in his
sections of Dewey and Tufts’s monumental Ethics,38 no one school has
a monopoly on insight. Utilitarianism is inadequate as an account of
our ethical lives, but it does identify a question that all too often needs
be put to someone who proposes a significant rule of action, law, or
policy: “Would it contribute to the satisfaction of more people’s desires
or interests than the alternatives?” At the same time, pragmatists such
as Dewey point out that desires and interests may be partial and selfish
and that even when they are not, they may be based on bad reasoning
or bad information. Desires and interests are not simply givens whose
satisfaction needs to be maximized (as too many economists have as-
sumed); they need to be subjected to intelligent and informed criticism.39
And Kantians have long pointed out that what satisfies the desires and
interests of even a substantial majority may be wrong because it tramples
on the rights of a minority.
Although, I don’t agree with Scanlon’s claim that morality can be
founded on the single principle that “an act is wrong if its performance
under the circumstances would be disallowed by any set of principles
for the general regulation of behav ior that no one could reasonably
reject as a basis for informed, unforced general agreement,” 40 I do see
that principle as a laudable attempt to state what is right in Kant’s cat-
egorical imperative in a naturalistic spirit—that is, to state it in a way
that does not depend on Kant’s untenable belief that respect for the
moral law has a “supersensible” source (as contrasted to our “inclina-
tions” that have a natural source). Like Dewey,41 I see the categorical
38. John Dewey, Ethics, vol. 5: 1908, in The Middle Works, 1899–1924, ed. Jo Ann Boyd-
ston (Carbondale: Southern Illinois University Press, 1978), 283–284.
39. See Hilary Putnam, “Dewey’s Central Insight,” in John Dewey’s Educational Philos-
ophy in International Perspective: A New Democracy for the Twenty- First Century, ed. L. A.
Hickman and Giuseppe Spadafora (Carbondale: Southern Illinois University Press,
2009), 7–21.
40. I criticize that idea in “Capabilities and Two Ethical Theories,” collected in Hilary
Putnam, Philosophy in an Age of Science: Physics, Mathematics, and Skepticism, ed. by Mario De
Caro and David Macarthur (Cambridge, Mass.: Harvard University Press, 2012), 299–311.
The quotation from Scanlon is from What We Owe to Each Other (Cambridge, Mass.: Har-
vard University Press, 1998).
41. Dewey, Ethics, 283–284.
Unauthenticated
imperative (as well as Scanlon’s principle, just quoted) as “a method for

realizing the full meaning of a proposed course of action,”42 but not as
the foundation on which morality rests. Morality rests on a complex and
evolving set of human needs and interests, but those needs and inter-
ests are neither biological givens nor the product of mere contingency.
As long as free inquiry and discussion survive, they will continue to be
forced to meet the test of rational scrutiny, to answer the demand for
good reasons.
What makes Scanlon’s naturalism clearly a liberal naturalism is the
willingness to use the notion of a reason without any accompanying
“naturalization project,” and I am delighted that he chose to follow
What We Owe to Each Other with a set of Locke Lectures entitled Being
Realistic about Reasons devoted to the legitimacy and importance of doing
just that.
In Closing
Part of what I have been saying might be summarized thus: the fact
that something— perceptual representation or reference or truth or
intentionality or reasons— can’t be “naturalized” in the way that “physi-
calists” demand doesn’t make those things “non-natural” or “queer” or
suspiciously close to “supernatural.” It is true that the notion of a reason,
for example, is not the subject matter of a special science, but that notion
is presupposed by all science as well as by fields like history and politics
and criticism (including philosophical criticism) that are not sciences,
because in all of them one has to decide what there is reason to con-
sider, and “elegance” figures in the reasons scientists give for testing
certain theories at all. They are not scientific notions, but the activity
of science presupposes a reasonable command of them.43
42. Concerning the Categorical Imperative, Dewey wrote, “As a method (though not of
course the only one) of realizing the full meaning of a proposed course of action, nothing
could be better than asking ourselves how we should like to be committed forever to its
principle: how we should like to have others committed to it, and to treat us according to
it? . . . In short, by generalizing a purpose, we make its general character evident” (Ethics,
283–284).
43. See Hilary Putnam, “The Philosophers of Science’s Evasion of Values,” in The Col-
lapse of the Fact-Value Dichotomy (Cambridge, Mass.: Harvard University Press, 2002),
135–145.
Unauthenticated
Science depends on what is not fully scientific at every point. And

both pragmatists and Wittgensteinians have rightly criticized the bad
habit of turning what are sometimes useful distinctions into untenable
dualisms, and that includes the dichotomy between normative and de-
scriptive language (a dichotomy that our thick ethical concepts—
including such ancient concepts as brave and wise—leap over without
the slightest sign of embarrassment) and also includes the dichotomy
between science and nonscience itself. One can learn from pragmatists
and Wittgensteinians and philosophers of so many other kinds without
becoming a card-carrying member of any philosophical sect. And that
is something I have always tried to do in my philosophical life.
Unauthenticated
D2
On Bernard Williams’s
“Philosophy as a Humanistic Discipline”
I n “P h i l osoph y as a Humanistic Discipline,”1 Bernard

Williams has misunderstood my criticisms of his position, and thus
ends up talking past me. Because I hope we can soon have an interchange
of views that is not marred by misunderstandings and mutual misin-
terpretations, I want to say where I think I have been misunderstood
and what my criticism actually was.
I shall quote and discuss several passages in William’s lecture.
Williams writes,
In the course of Putnam’s book (which contains a chapter called

“Bernard Williams and the Absolute Conception of the World”),
I myself am identified as someone who “views physics as giving us
the ultimate metaphysical truth.” Now I have never held any such
1. Bernard Williams, “Philosophy as a Humanistic Discipline,” Philosophy 75 (2000):

477–496. My criticisms of Williams, to some of which he replies in this lecture, were con-
tained in “Objectivity and the Science / Ethics Dichotomy,” collected in my Realism with a
Human Face, ed. James Conant (Cambridge, Mass.: Harvard University Press, 1990), 163–
178; “Bernard Williams and the Absolute Conception of the World,” a chapter in my Re-
newing Philosophy (Cambridge, Mass.: Harvard University Press, 1992), 80–107; and “Prag-
matism and Relativism: Universal Values and Traditional Ways of Life,” collected in Words
and Life, ed. James Conant (Cambridge, Mass.: Harvard University Press, 1994), 182–197.
[Unfortunately, Bernard Williams’s demise in 2003 put an unanticipated end to our de-
bates. Note added on July 15, 2015.]
44
Unauthenticated
On Bernard Williams’s “Philosophy as a Humanistic Discipline” 45
views, and I agree with Putnam in rejecting it. However, I have

entertained the idea that science might describe the world “as it is
in itself,” that is to say, give a representation of it which is to the
largest possible extent independent of the local perspectives or idio-
syncrasies of enquirers, a representation of the world, as I put it, “as
it is anyway.”2
Since I did not mean by “the ultimate metaphysical truth” any more
than “a description of the world as it is anyway,” there is already a mis-
understanding at work here. The reason I spoke of physics in the pas-
sage Williams quotes, and not more vaguely of “science” as Williams
does here, is that Williams himself has elsewhere said explicitly that the
notion of an “absolute conception” does not “look too pale” because we
have a conception of “what an adequate physics might look like.”3 I did
make— and will go on making, until and unless Williams himself cor-
rects me— one further assumption: I take Williams’s position to be that
the description of the world in terms of its nonperspectival properties
is a complete description of all of the world. It isn’t, that is to say, that
an exhaustive description of the world in terms of all the “absolute”
properties would only describe a part of the world, and there is another
part, the “perspectival part,” that would still remain to be described.
This means that if a semantical statement, for example, “John referred
to object X,” describes a “perspectival fact”— say, the fact that in such-
and-such a perspective, John referred to object X—then that whole fact,
including the perspective and John and the object X and the relation
between them—must somehow appear in the “absolute conception.”
And the absolute description is envisaged as being given in terms of the
fundamental magnitudes of physics! Thus, it seemed to me that Wil-
liams does need to somehow reduce semantical facts to purely physical
facts if he is not to become an outright denier of the real ity of the
semantical (an “eliminationist,” in the jargon of contemporary ana-
lytic philosophy).4 And in his book Descartes: The Project of Pure Enquiry
2. Williams, “Philosophy as a Humanistic Discipline,” 481.

3. Bernard Williams, Descartes: The Project of Pure Enquiry (Harmondsworth, United
Kingdom: Penguin, 1978), 247.
4. In current parlance, an “eliminationist” (Paul Churchland is the paradigm case) holds
that propositional attitudes (e.g., belief) and semantic relations (e.g., reference) don’t exist,
and that the idea that they do is a superstition (Churchland has compared it to belief in
Unauthenticated
he did suggest, if I do not misread him, that eliminationism with re-

spect to the semantical is probably the right line to take. For he wrote,
If the various sorts of considerations [Quinian and Davidsonian

considerations— HP] which have been summarily sketched here
are correct, then we have to give up not just dualism but the belief
in the determinacy of the mental. These considerations converge
on the conclusion that there are no fully determinate contents of
the world which are its psychological contents.5
Indeterminacy of psychological contents (in the sense of Quine and

Davidson) is indeterminacy of the semantical. And that indeterminacy,
if their arguments are accepted, must be far reaching indeed. (Quine
says, for example, that there is no fact of the matter as to whether
“Tabitha” refers to his cat Tabitha or to “the whole cosmos minus the
cat.”6)
It is true that Davidson, while giving full credit to Quine for the ar-
guments that allegedly establish “indeterminacy,” says that the extent
of indeterminacy is much reduced in his theory by his willingness to
make “a more far reaching application of the principle of charity” than
Quine.7 (Perhaps this is the reason that Williams wrote “no fully de-
terminate contents” and not simply “no determinate contents.”) But
charity is a maxim of interpretation—that is, of translation into the in-
terpreter’s home language— and it cannot bestow any additional deter-
minacy on the home language itself. It isn’t as if one could really argue
that reference is partially, even if not “fully,” determinate, except in the
witches, or in phlogiston). Compare “Activation Vectors vs. Propositional Attitudes: How

the Brain Represents Real ity,” in Churchland’s On the Contrary: Critical Essays, 1987–1997
(Cambridge, Mass.: MIT Press, 1998). [In my view, Quine’s claim that reference is indeter-
minate to the extent that there is no fact of the matter as to which object “Tabitha” refers to
is just eliminationism under another name.] I criticize the cited essay of Churchland’s in
“Truth, Activation Vectors, and Possession Conditions for Concepts,” Philosophy and Phe-
nomenological Research 52, no. 2 (1992): 431–447. Note that eliminationism is very dif ferent
from reductionism.
5. Williams, Descartes, 300.
6. W. V. Quine, The Pursuit of Truth (Cambridge, Mass.: Harvard University Press,
1990), 33.
7. Donald Davidson, Inquiries into Truth and Interpretation (Oxford: Clarendon Press,
1984), 228.
Unauthenticated
sense (which even Rorty could agree with) of “determinate relative to a

translation scheme”). That the vocabulary of physics (or a perfected
future physics) is rich enough to give a complete description of the world
is the view I (perhaps misleadingly) described by saying that Williams
views physics as giving us the ultimate metaphysical truth. Indeed, the
view that any single vocabulary could suffice to give a complete descrip-
tion of the world does seem to me a metaphysical fantasy.8 But if the
word “metaphysical” is what is leading to misunderstandings, I am
willing to omit it.
That it did lead to misunderstandings is evident, I think, when Wil-
liams writes,
[Putnam] supposes that the idea of an absolute conception of the

world must ultimately be motivated by the contradictory and in-
coherent aim of describing the world without describing it: as he
puts it, we cannot divide language into two parts, a part that de-
scribes the world “as it is anyway” and a part that describes our
conceptual contribution . . . My idea was not that you could con-
ceptualize the world without concepts.9
I never thought that this was Williams’s idea; I have far too much re-
spect for Williams’s intelligence to charge him with such an obvious
blunder. Nor did the assertion he quotes (“we cannot divide language
into two parts, a part that describes the world ‘as it is anyway’ and a part
that describes our conceptual contribution”) accuse him of making that
blunder. To think, as I said Williams thinks, that there is a part of
language that describes the world “as it is anyway” is to accuse him
of thinking that there are concepts that do this, and not to accuse
him of thinking that we can describe the world “as it is anyway” without
using concepts (whatever describing without using concepts might mean).
Our disagreement is over whether it makes sense to think that the concepts
of physics do this (assuming, as I said I will continue to assume until Wil-
liams corrects me, that the description of the world in terms of its “abso-
lute” properties is supposed to be a description of all of the world.)
8. See my “From Quantum Mechanics to Ethics and Back Again,” in Reading Putnam,
ed. Maria Baghramian (Abingdon, United Kingdom: Routledge, 2013), 19–36, esp. 28–30.
9. Williams, “Philosophy as a Humanistic Discipline,” 482.
Unauthenticated
Putnam’s basic argument against the idea of the absolute concep-

tion is that semantic relations are normative, and hence could not
figure in any purely scientific conception. But describing the world
involves deploying terms that have semantic relations to it; hence,
it seems, Putnam’s conclusion that the absolute conception is sup-
posed to describe the world without describing it.10
I have already said that there was no such “Putnam’s conclusion.”

Also, my basic argument was not that semantic relations are normative
(although there is a sense in which they are), “and hence could not figure
in any purely scientific conception,” but that Williams himself denies
that semantic relations are determinate (I have already quoted him on
this point). But if semantic relations are “perspectival” in the sense of
holding only relative to some scheme of interpretation or other, and if
[as Quine claims] there is no fact of the matter as to whether any scheme
of interpretation that correctly predicts the stimulus-meanings of whole
sentences is more right than any other [or, in Davidson’s case, whether
any collection of reference-assignments that comes out right on the
truth-conditions of whole sentences is more right than any other], then
“absoluteness”—a notion which Williams employs semantical terms to
define—is also “perspectival” in the sense of holding only relative to our
choice of one scheme of interpretation [that satisfies the constraint] or
another. I shall spell this point out in more detail in a moment. But,
broadly speaking, my argument was that unless “the absolute conception
of the world” is more than our projection, it cannot do the metaphysical
work (of supporting a sense in which scientific truth is less “perspectival”
than, say, semantical truth) that Williams wants it to do. Williams, I
claimed, needs an absolute notion of “absoluteness.” But his denial that
semantic relations could figure in any purely scientific conception—not
mine—leaves Williams with only a perspectival notion of absoluteness,
not an absolute one. As I put it in the chapter to which he refers, “Is Wil-
liams saying that it is just our local perspective that there is an absolute
conception? Even Rorty might be able to agree with that.”11
A few words more on Williams’s claim that “Putnam’s basic argu-
ment against the idea of the absolute conception is that semantic rela-
10. Williams, “Philosophy as a Humanistic Discipline,” 482–483.

11. Putnam, Renewing Philosophy, 101.
Unauthenticated
tions are normative, and hence could not figure in any purely scientific
conception.”
I myself have never affirmed or denied that there could be a “scien-
tific conception” of semantic relations. As I have explained elsewhere,12
I think the terms “science” and “scientific” are much too vague for that
question to make sense. But I take it that what Williams means by the
phrase “purely scientific conception” here is “absolute conception,” or,
perhaps, conception that is at least a good sketch of what an absolute
conception might be. (He says, and I have already quoted this, that we
already have a notion of what an “adequate physics” might be, and that
this fleshes out our understanding of what an “absolute conception”
could be; he has also said that “the world itself has only primary quali-
ties.”)13 Thus it seems clear that Williams himself provides all the
premises I needed to argue that, on his own view, semantic relations
do not figure at all in the absolute conception. Indeed, in the lecture to
which I am replying he wrote, “I take it as obvious that any attempt
to reduce semantic relations to concepts of physics is doomed.” So I do
not see why references to my view that “semantic relations are norma-
tive,” and the expression “Let us grant for the sake of the argument the
principle, which is certainly disputable, that if semantic relations are
normative, an account of them cannot figure in the absolute concep-
tion,” come into Williams’s discussion at all. Given what Williams says
about the absolute conception (e.g., that it contains only primary
qualities— and, presumably, what is definable in terms of primary
qualities using the apparatus of mathematical physics), and what he
says about semantic relations, it follows from Williams’s premises—not
just from mine—that “an account of them cannot figure in the abso-
lute conception.” And given his characterization of the absolute con-
ception, it is not hard to see that he is right. What would an account of
a semantical relation in the language of mathematical physics (or in
terms of “primary qualities”) be?
It is, of course, true that I believe that semantic relations are norma-
tive, not in the sense of believing that semantical statements are value
judgments, which would be absurd, but in the sense (and this far I agree
12. Compare with my “The Idea of Science,” Midwest Studies in Philosophy 15 (1990): 57–
64; collected in my Words and Life, 481–491.
13. Williams, Descartes, 247.
Unauthenticated
with Davidson) that there is no criterion for the correctness of state-

ments to the effect that words have either the same meaning or the
same reference above and beyond reasonable interpretative practice.
Judgments of reference and meaning, I have claimed, essentially in-
volve judgments of reasonableness; and the idea of reducing judgments
of reasonableness to exact science (in the sense in which mathematical
physics is an exact science) are nothing but scientistic fantasies.14
Certainly this is controversial, but it also is something on which, unless
I misunderstand him completely, Williams and I agree.
And a word to explain my statement that Williams employs seman-
tical terms to define “absoluteness.” Williams’s basic thought experi-
ment is the following: imagine that there are various tribes of intelligent
extraterrestrials who (independently of us, and without even knowing
of our existence) investigate the world scientifically. What they would
converge in believing—or better, what they would converge in be-
lieving not for accidental reasons, but because it is the way things in
fact are independently of what we think or believe—is the absolute con-
ception. [This is, roughly, Williams’s definition of “absolute concep-
tion.”]15 Now “A and B converge in believing C” means that, from a
certain point on, A and B both believe C. But if there is a fact of the
matter as to what the “psychological contents” (e.g., the beliefs) of A
and B are only relative to a scheme of interpretation, then there is a fact
of the matter as to whether A and B converge in believing C (or any-
thing else, for that matter) only relative to a scheme of interpretation.
“Converge” is a semantical term.
I now turn to a dif ferent issue— quite possibly the real issue between
us. Williams writes,
So why does Putnam assume, as he obviously does, that if there

were to be an absolute conception of the world, philosophy would
have to be part of it? I doubt that he was simply thrown by the
14. I have argued this in a number of places, most recently in the Threefold Cord: Mind,
Body and World (New York: Columbia University Press, 1999), 119–125, and in “Aristotle’s
Mind and the Contemporary Mind,” in Demetra Sfendoni-Mentzou, Jagdish Hattiangadi,
and D. M. Johnston, eds., Aristotle and Contemporary Science (New York: Peter Lang, 2000),
7–28.
15. Compare Bernard Williams’s Ethics and the Limits of Philosophy (London: Routledge,
1985), 136.
Unauthenticated
Hegelian implications of the word “absolute,” with their implica-

tion that if there is absolute knowledge, then philosophy possesses
it. What perhaps he does think is the conjunction of two things:
first that philosophy is as good as it gets, and is in no way inferior
to science, and, second, that if there were an absolute conception
of the world, a representation of it which was maximally indepen-
dent of perspective, that would be better than more perspectival
or locally conditioned representations of the world. Now the first
of these assumption is, as it were, half true; although philosophy
is worse than natural science at some things, such as discovering
the nature of the galaxies (or, if I was right about the absolute con-
ception, representing the world as it is in itself), it is far better
than natural science at other things, for instance making sense of
what we are trying to do in our intellectual activities. But the
second assumption I have ascribed to Putnam, that if there were an
absolute conception, it would somehow be better than more per-
spectival representations—that is simply false. Even if it were pos-
sible to give an account of the world that was minimally perspec-
tival, it would not be particularly ser viceable to us for many of our
purposes, such as making sense of our intellectual or other activi-
ties, or indeed in getting on with most of those activities. For those
purposes—in par ticular, in seeking to understand ourselves—we
need concepts and explanations which are rooted in out more local
practices, our culture and our history, and these cannot be replaced
by concepts which we might share with very different investigators
of the world.16
To take the second of the two assumptions that Williams ascribes to

me first, I don’t think I ever assumed that “if there were an absolute
conception of the world, a representation of it which was maximally in-
dependent of perspective, that would be better than more perspectival
or locally conditioned representations of the world.” And as for the
16. Williams, “Philosophy as a Humanistic Discipline,” 483–484. Williams continues

there: “The slippery word ‘we’ here means not the inclusive ‘we’ which brings together as a
purely abstract gathering any beings with whom human beings might conceivably commu-
nicate about the nature of the world. It means a contrastive ‘we,’ that is to say, humans as
contrasted with other possible beings; and in the case of many human practices, it may of
course mean groupings smaller than humanity as a whole.”
Unauthenticated
supposed first assumption, “that philosophy is as good as it gets, and is

in no way inferior to science”—that is at best misleadingly put. But let
me explain.
The two sorts of judgments that were chiefly at issue in “Bernard
Williams and the Absolute Conception of the World” were (1) seman-
tical judgments, and (2) ethical judgments. My claim that Williams
himself needs an absolute conception of absoluteness, not a perspectival
one, refers to semantical judgments.17 I take it that when he speaks of
“philosophy” in the long passage that I quoted, he is counting both sorts
of judgments as part of “philosophy.” Confining attention here to the
former sort, the question becomes: why, if I was not making the two
assumptions attributed to me, did I assume semantical statements must
be “absolute” if Williams’s view is to work?
I think I already answered this above, but it may make matters
clearer if I now discuss the assumption that I am supposed to make
“that philosophy is as good as it gets, and is in no way inferior to sci-
ence.” Consider a statement that might be part of the absolute con-
ception, namely, that the sun is approximately 93 million miles from
the earth. (Of course, it couldn’t be referred to as “the sun” and “the
earth” in the absolute conception; but the same problem arises with
any example I might give— and it is a problem for Williams, not for
me, I believe.18) And consider the semantical statement (which I take
to be part of what Williams here calls “philosophy”) that when I just
asked you to consider that statement, I was using “the earth” and
“the sun” to refer to a planet and the star around which it revolves,
respectively. What I believe is that the semantical statement states a
fact—the fact that the words “the sun” and “the earth” bear a seman-
tical relation to two things that are not marks and noises and do not
consist of marks and noises—which is just as much an objective fact
as the fact that the sun is approximately 93 million miles from the
earth.
I am unable to make coherent sense of Williams’s notion of a “per-
spectival” truth, but since he seems to depend on Quine and Davidson
17. I discuss what Williams says about ethics in Ethics without Ontology (Cambridge,
Mass.: Harvard University Press, 2004), 122–129.
18. Having gone Davidsonian, Williams cannot avail himself of the solution that “the
earth” and “the sun” are rigid designators and not descriptions in these judgments!
Unauthenticated
for his view that “psychological contents” (and hence meaning and
reference) are not “absolute,” let us see how this statement fares on
their respective philosophical views. In Quine’s view, to say that “the
earth” refers to a certain planet is “parochial” in the sense that (1) the
term “the earth” can be mapped on infi nitely many dif ferent things
via what Quine calls “proxy functions,” and (2) each of the resulting
mappings is as entitled to be regarded as “the” relation of reference as
any other. Of course, each of the infinitely many resulting models of
my language is such that (in the corresponding model of my metalan-
guage) the sentence “ ‘the earth’ refers to the earth” is true; but that
doesn’t mean that there is a fact of the matter as to which object the
term “the earth” in the above statement refers to (just as there is no
fact of the matter as to which object “Tabitha” refers to, even though
“ ‘Tabitha’ refers to Tabitha” remains true no matter what proxy
function we choose). It is only relative to a “translation scheme” that
I am talking about the earth at all. And in Davidson’s account of ref-
erence, the same thing happens, as I have already noted, notwith-
standing his claim that his greater willingness to appeal to charity “re-
duces” indeterminacy.19
There is an important difference between Quine and Davidson (but
I do not see how Williams will be able to take advantage of it). Davidson
does think that whole sentences have determinate truth conditions, and
truth, for Davidson, is not simply disquotational (as it is for Quine). It
is not that Davidson defines “true”; famously, he holds that it is a simple
and clear idea that needs no definition. But in his view there is a con-
nection between my sentence “The earth is 93 million miles from the
sun” and the world external to language, one that is captured precisely
by the truth condition for my sentence. And to be a thinker at all, ac-
cording to Davidson, I have to have such notions as “holds true” and
“is true”; other wise, I cannot even form concepts. So Davidson’s view
implies that all those extraterrestrial investigators that Williams hy-
pothesizes in his thought experiment must have the notion of truth;
that there is a difference between a false sentence and a true one is some-
thing they must recognize from the word “go.” So “true” would seem
to belong to the absolute conception—unless the reason they converge
19. Davidson, Essays on Truth and Interpretation (Oxford: Clarendon Press, 1984),
221–240.
Unauthenticated
on this belief isn’t that there is “anyway” a difference between a false

sentence and a true one. But I hope that will not be Williams’s way out
(the seas of language would be too high for me to sail, in that case). In
sum, it isn’t a question of “philosophy” (semantics, in this case) being
better or worse than physics; it is simply the case, I am arguing, that if
physics is to be “absolute” in Williams’s sense, then our talk about phys-
ical entities had better refer—had better objectively refer. And so at
least one semantical notion—“refer,” or if you believe Davidson,
“true”—had better be “absolute.” For Quine’s version of perspectivalism
with respect to the semantical utterly guts Quine’s so-called robust re-
alism of any content that Richard Rorty would have any reason to ob-
ject to.20
Isn’t this an objection to Davidson as much as to Williams? In a way
it is. For, given Davidson’s insistence that experience has only a causal
and no justificatory role with respect to our beliefs, I do not think Da-
vidson really does have a satisfactory answer to McDowell’s charge that
it is unintelligible, in Davidson’s picture, how sentences do have deter-
minate truth conditions.21 But here I have only been concerned to argue
that if Davidson has an answer, that answer depends on a kind of re-
alism with respect to the semantical concept of truth that seems incom-
patible with Williams’s identification of the “absolute” with the physics
of “primary qualities.”
I said that I discuss Williams on ethics in another place. But just to
avoid misunderstandings, let me say this much here: of course there
could be beings whose lives were such that our ethics was simply inap-
plicable to them. Very likely, they couldn’t even find our ethical notions
intelligible. But I don’t believe (1) that ethical notions are, in general,
descriptions (so the dichotomy that seems to be implicit in Williams’s
discussions—“absolute” description of objects in terms of their primary
qualities or just a “perspective” on such objects—seems to me to leave
out of account a huge range of judgments that are objective but not de-
scriptions (not even “perspectival” ones), including all purely mathe-
matical judgments); and (2) the fact that not every creature might need
20. I argue this in detail in “A Comparison of Something with Something Else,” col-
lected in Realism with a Human Face, 330–350.
21. John McDowell, Mind and World (Cambridge, Mass.: Harvard University Press,
1994), 129–161.
Unauthenticated
a certain concept doesn’t, I would argue, imply that judgments involving

that concept are only true in the “social world” of the creatures who
do need that concept. But these are views that I have defended
elsewhere.22
22. Although I disagree with some of Wittgenstein’s views about mathematics, that
mathematics is not a description (e.g., of intangible and necessarily existing “mathematical
objects”) is something I have long agreed with. See my “Mathematics without Foundations,”
collected in my Mathematics, Matter and Method, vol. 1 of Philosophical Papers (Cambridge:
Cambridge University Press, 1979), 43–59, and my “On Wittgenstein’s Philosophy of Math-
ematics,” Proceedings of the Aristotelian Society, Supplementary Volumes 70 (1996): 243–264. I
criticize the inference from the fact that a par ticu lar ethical concept is peculiar to a certain
society and reflects that society’s needs and interests to the conclusion that what is said with
the aid of that concept is only true “in that social world” in “Pragmatism and Relativism:
Universal Values and Traditional Ways of Life,” in Words and Life, 189–196.
Unauthenticated
D3
What Evolutionary Theory
Doesn’t Tell Us about Ethics
P e r h a ps, there has to be a skeptic at every academic

conference. I fear that today, it is my turn to play that not altogether
enviable role, for I shall defend the view that the answer to the ques-
tion to which the conference for which the present chapter was written
was devoted, namely “to what extent can we understand the roots and
complexity of ethical judgments from a Darwinian perspective?” is
“Not very much.” But first, a couple of caveats.
If the reference to “moral sentiment” in the title of the book from
which this chapter derives includes minimally conceptualized emo-
tions that are relevant to ethics,1 that is, certain emotions that have
not yet become entangled with concepts that have a complex cultural
role and a complex cultural history— emotions such as spontaneous
sympathy—then it is indeed plausible that evolutionary theory can
tell us something about them. It is also quite plausible that the origin
of altruistic behaviors is something that evolutionary theory can
shed light on. But one thing I would like to do today is to show you
how far short explanations of altruistic behav ior and of sympathy (as
manifested in, for example, groups of apes) fall of dealing with what
1. Hilary Putnam, Susan Neimann, and Jeffey Schloss, eds., Understanding Moral Senti-
ments from a Darwinian Perspective: An Exploration of the Roots and Complexity of Ethical Judg-
ment (New Brunswick, N.J.: Transaction, 2014).
56
Unauthenticated
What Evolutionary Theory Doesn’t Tell Us about Ethics 57
we have come to call ethics. (Of course, I do not deny that sympathy
and altruistic behaviors, even if they do not amount to ethics all by
themselves, are among the preconditions for the complex institution
that we call ethics. Indeed, they are probably among the preconditions
for the complex institution that we call science, and for the complex
institution that we call government—perhaps for all complex human
institutions.)
I speak of “what we have come to call ethics,”2 because my claim is
that ethics is something with a long cultural history. Ethics has devel-
oped and changed throughout recorded history (and doubtless had a
prehistory that is much longer than recorded history). I know, of course,
that stressing the role that culture plays in producing what we call
“ethics” is often considered tantamount to endorsing “moral relativism”
or even “nihilism”; that is, it is considered to entail that there is no
objective rightness or wrongness to ethical judgments, but that, I believe,
is a mistake. That ethics has a cultural and not only an evolutionary his-
tory, that it underwent a long process of development, and that it
wasn’t something that was “fixed in our genes” from the beginning does
not by itself imply anything about either the rationality or the objec-
tivity of ethical judgments.
To illustrate what I mean by the “cultural history” of ethics, I will offer
a much-too-brief account of a few moments in the development
of Western ethics; ideally, one would wish to consider also the history of
Chinese ethics over the same period of time, the history of Indian
ethics, and so on. (In The Idea of Justice,3 Amartya Sen draws on the
history of Indian ethics at a number of places.) At any rate, the term
“ethics” in the West assumes something like its present meaning with
Aristotle’s lectures, particularly the ones written up by his students and
preserved for us as the Nicomachean Ethics. Although Aristotelian ethics
is still relevant today and one of the leading approaches among Anglo-
phone ethicists, called “virtue ethics,” acknowledges Aristotle as its
forefather, there are many ways in which this concept of ethics is very
dif ferent from post-Enlightenment ethics.
2. I shall not distinguish, as some theorists do, between “ethics” and “morals”; indeed,
“moralis” functioned for Latin ethicists, such as Cicero who introduced it, as simply the
translation of the Greek “ethika.”
3. Amartya Sen, The Idea of Justice (Cambridge, Mass.: Harvard University Press, 2009).
Unauthenticated
Aristotelian ethics is considered to be a form of “virtue ethics”

because the central concern is to identify “virtues,” that is, traits that
we should wish citizens of a political community (in Aristotle’s time, a
Greek polis) to possess, both because those traits are good for the com-
munity and because they are seen as admirable in themselves, not to
formulate “rules” or “maxims” of conduct. It is noteworthy that “al-
truism” is not explicitly listed as one of the virtues, although public-
spiritedness (which is not exactly the same thing by any means) is highly
valued. Indeed, insofar as the Greeks saw something we might see as
“altruism” as a virtue, they certainly did not conceive of it in the way
in which, for example, Marc Hauser seems to conceive of it in the
primate experiments he describes in Moral Minds,4 that is, in the
Kantian-cum-Rawlsian way, according to which the “altruistic” person
(or animal) is disposed to act cooperatively and beneficially toward
others, regardless of their social ranks. It would be an understatement
to say that there is very little said in Aristotelian thought about showing
“altruism” toward one’s slaves. (We do, of course, fi nd the usual slave-
holders’ paternalistic claim that those who are slaves by nature are
better off being slaves; and perhaps, Hauser would consider this an ex-
pression of altruism.) On the other hand, both theoretical intelligence
and practical intelligence (sophrosyne) are important virtues for Greek
thought. Note that, while we do not consider theoretical intelligence a
moral virtue at all, it certainly is one for both Plato and Aristotle. It is
considered a part of the highest form of human flourishing, and human
flourishing is a central topic of Greek ethics. By “sophrosyne” is meant
the sort of discriminating intelligence that enables one not simply to
follow rules “mechanically,” but to discern what is appropriate to a par-
ticular situation. It is not surprising that in societies that frequently
found themselves at war, and where all the males were expected to
defend the polis in armed combat when necessary, the “manly” virtue
of courage (carefully distinguished by Plato and Aristotle from foolhar-
diness, or mere fearlessness) was regarded as a central virtue. Modera-
tion was also a central virtue, prudence was a central virtue, and for
Aristotle and Plato and the Greek and Roman moralists that came
4. Marc Hauser, Moral Minds: How Nature Designed Our Universal Sense of Right and
Wrong (New York: Ecco, 2006). I prescind from the problems that surfaced about Professor
Hauser’s research.
Unauthenticated
after them, knowing what the good life consists in, particularly knowing
that the highest goods in life are definitely not the accumulation of
wealth, fame, or sensual pleasure, were the central topics.
Fast-forwarding more than two millennia and coming to the Enlight-
enment, we find that equality has become a central topic.5 The great
historical changes that culminated in the American and French Revo-
lutions were influenced by the idea of government as a “social contract”
into which all the citizens enter, and after those revolutions, it eventually
became the dominant model in the West (and powerfully influenced
such leaders as Sun Yat-sen and Mahatma Gandhi in the East as well).
It is true, of course, that as late as World War I, many of the leading
states of Eu rope were still monarchies, albeit “constitutional monar-
chies”; nonetheless, the idea that a just government rests on the con-
sent of the governed has become the dominant moral-political idea in
the West and, at least at the level of lip ser vice, in the world.
If we look more closely at the process by which democratic and egal-
itarian values acquired their present centrality, we can also see that
changes in epistemology, many of them connected with changes in our
understanding of how the natu ral sciences should proceed, played a
large role in changing our conceptions of ethics.
What I have in mind is this: according to Aristotle— and Aristotle
was the most influential philosopher of science for almost two
millennia— science reveals necessary truths about the world. Aristotle
indeed valued observation, and he even spoke of “induction.” But, in the
end, observation and induction are to lead us to necessary truths, and
necessary truths, once properly perceived, are supposed to be “apo-
dictic,” unrevisable. This idea was, of course, challenged by empiricists
from the seventeenth century on. In our own time, the overthrow of
the idea of apodictic truth was part of the overthrow of the analytic–
synthetic dichotomy, and more broadly of the a priori–a posteriori di-
chotomy, that Quine deserves so much credit for.
Why do I mention equality, democracy, and the a priori–a posteriori
dichotomy virtually in the same breath? The reason is this: as early as
Aristotle and Plato, the question of the equality or inequality of men
5. In this connection, see my Spinoza Lectures, “Enlightenment and Pragmatism,” col-

lected in my Ethics without Ontology (Cambridge, Mass.: Harvard University Press, 2004),
89–129.
Unauthenticated
and women, particularly the question whether women should be the po-
litical equals of men, was one about which there was disagreement. In
his Republic, Plato famously argued6 that women should be eligible to
be “Guardians,” that is, members of the ruling elite of his ideal state.
Aristotle, however, “justified” the prejudices of his time (and his “justi-
fication” was widely accepted—by men, of course—for more than two
thousand years) by arguing that women should not be eligible because
they lacked the higher form of intelligence, intellectual nous. Today,
such a claim would be regarded as an empirical claim and as a claim
that has been falsified. But in Aristotle’s time and after, it was accepted
(by men!) as something more or less self-evident. Originally, this
“justification” was accepted (by those men who saw the need for a jus-
tification) largely on Aristotle’s authority, and eventually it came to be
regarded as one of those self-evident truths that reason simply comes
to perceive after appropriate examination of the phenomena. And once
a supposed truth about nature had achieved that status in ancient and
medieval science, it was viewed as unnecessary and inappropriate to
look for “evidence” for or against.
If the whole history of ethics were simply a history of different “codes”
of conduct supported by claims of the “superiority” or “inferiority” of
dif ferent groups of people, or the “superiority or “inferiority” of dif-
ferent ways of life, that were not themselves subject to rational challenge, then
it could indeed be argued that the fact that ethical codes are cultural
products supports “ethical relativism,” or even what Alex Rosenberg has
called “nihilism,” but that argument goes too fast. For with the Enlight-
enment came not only the valorization of equality but also the valori-
zation of the new post-Galilean science. And with the valorization of
modern science came the idea that claims about nature, including claims
about human nature, are empirical claims that can be challenged and
investigated.
It is true that so-called “racial science” exerted a baneful influence
well into the last century, and in the eyes of the German Nazis pro-
vided a “justification” for the Holocaust; but today we do know that the
claims of “racial science” were both empirically false and profoundly
unscientific in their supposed methodology. Generalizing this case, we
see that the presuppositions upon which a supposedly ethical form of behavior
6. See also Eurthyphro; see my discussion in Ethics without Ontology, 89–92.
Unauthenticated
are based can themselves frequently be rationally criticized, and this is enough
to show that the “nihilistic” picture of ethics as just a matter of “condi-
tioning” by one’s culture, “conditioning” to value behaviors that are
rationally arbitrary, is a naive oversimplification. We are, of course,
brought up in cultures, and we acquire beliefs from our cultures, but
it is possible to ask whether those beliefs are reasonable or unreason-
able. For example, the idea that the highest type of human being is a
brave male warrior, and that society should therefore be led by brave
male warriors—the idea that the standard of human value should be
bravery in battle—was one that had already come to seem unreasonable
by the time of Plato and Aristotle (which is obviously not to say that they
didn’t value bravery in battle).
Of course, a “nihilist” may retort, “Sure, but what is ‘reasonable’
and ‘unreasonable’ is itself just a matter of conditioning.” But if judg-
ments of the reasonable and unreasonable are themselves outside of
the domain of rational criticism, then science itself must lie outside the
domain of rational criticism. For science is not just a matter of checking
“predictions,” as the positivists seemed to think; it is also a matter of
deciding which hypotheses are worth the time and expense of exper-
imental testing, as the great pragmatist C. S. Peirce already empha-
sized, and such decisions, as Walsh and I have argued in a number of
papers,7 presuppose epistemic values. Indeed, even when a theory is
tested experimentally, the decision as to which experiments to trust
and which to regard as dubious requires judgments of “coherence” and
“plausibility,” which resemble aesthetic judgments. (This is particu-
larly true of evolutionary theories, by the way, which are always open
to the suspicion of being just-so stories.) If all values are regarded as
matters of arbitrary conditioning, then science itself should be regarded
“nihilistically” as well.
I don’t mean to suggest that virtue ethics and social contract theory
exhaust all the dif ferent sorts of ethical concerns we have; obviously
they don’t. Rather, my picture is this: ethics, as we know it, rests on
a certain number of human interests. Some of those interests, for ex-
ample, the interest in a stable polis, large or small, are virtually universal
among human communities, and perhaps evolutionary psychology can
7. These papers are collected in Hilary Putnam and Vivian Walsh, The End of Value-Free
Economy (Abingdon, United Kingdom: Routledge, 2014).
Unauthenticated
tell something about them. Others, however, came to the fore only
after the old claims of the innate superiority of one “race” over an-
other, or one gender over another, or warriors over nonwarriors crum-
bled in the face of reasonable criticism. In short, I believe that there are
learning processes in history, and that the rise of ethical systems that
value equality and not inequality and that see just governments as
resting upon the consent of the governed rather than on the divine
right of kings (or the innate superiority of philosophers) represents
progress. If there is evolution here, it is cultural evolution, and cultural
evolution, as has often been remarked, is “Lamarckian”; in cultural
history, there is an “inheritance,” that is to say a transmission, of ac-
quired characteristics, and those characteristics are not themselves
written into our genes.
As I pointed out above, certain of the interests that became central
to ethics in the West did so as the result of enormous political trans-
formations, transformations that are still continuing or striving to
continue in many parts of the world (for example, as I write these
words, in Iran), as the demand for the rule of law and the idea of the
“consent of the governed” become fundamental and acquire the aspi-
ration to universality.
The interest that my colleague and good friend Thomas Scanlon sees
as constitutive of ethics, the interest in being guided by a set of rules
that are accepted by all with an interest in a just society, and that no
one with that interest can reasonably object to, I see not as the founda-
tion of ethics but as something that has come to be one of the basic in-
terests of ethics. But there are others. I do believe that Scanlon has well
described how one sort of ethical claim can have motivating force in
any community that shares one of the basic interests of morality. But I
also believe that morality has a number of basic interests, including re-
spect for the humanity in the Other, equality of moral rights and re-
sponsibilities, compassion for suffering, and concern to promote human
well-being, and not only the desire to be governed by principles for
which one can give one another reasons, although that too is one of
them. Even though these interests sometimes conflict, I believe that,
on the whole and over time, promoting any one of them will require
promoting the others. Precisely for this reason, a philosopher who suc-
cumbs to the temptation to see ethics as standing on a single “founda-
tion” can always write a book “showing” that all of ethics “derives” from
Unauthenticated
that interest—indeed many such books have already been written.8 But
I believe this temptation should be resisted.
To sum up, then: I am suggesting that the history of ethics is what
I just called “Lamarckian” and not “Darwinian.” It is the history of
the evolution of a set of distinctive interests that themselves presup-
pose certain historical developments; in par ticular, they more and
more presuppose the development of democratic societies. I do not
agree that these interests are arbitrary; I don’t believe that the interest
in democracy is an “arbitrary” interest; on the contrary, I agree with
de Tocqueville that even if “the great democratic revolution” that oc-
curred in his time was “a new thing” it also represented “the most
continuous, ancient and permanent tendency known in history.” 9 But
if Tocqueville is right, and the aspiration to democracy is a very old
one, the realization of that aspiration, even to the imperfect extent
that it has been realized today, is a very recent phenomenon as histor-
ical time goes. Certainly, there were those who aspired to see the
equality of men and women and the equality of all races and an end to
theocracy, nonconstitutional monarchy, and all forms of dictatorship
in the very distant past, but it required new means of criticism, scien-
tific and philosophical criticism, to overthrow the arguments for in-
equality that dominated for so long. In short, we have in ethics a vast
complex of institutions and a vast complex of concerns, and one doesn’t
even begin to scratch the surface of that complex if one confi nes one-
self to talking about “altruism.”
Before closing, however, I need to say something about the
“Chomskian” view proposed by Marc Hauser in Moral Minds. According
to that view, our species has an “innate moral grammar,” and all the
historical developments that I have described would be just so many
dif ferent realizations of that universal moral grammar. According to
Hauser, the grammar itself can be accounted for by evolutionary
psychology. There are at least two major difficulties with this theory,
in my view.
8. In Ethics without Ontology, I said of the ethicists who succumb to this temptation that
“it is as if they wanted to see ethics as a noble statue standing at the top of a single pillar”
[Nelson’s Column, perhaps], and I added that “I see ethics as a table with many legs, which
wobbles a lot [ because the floor is not even] but is very hard to turn over” (28).
9. Alexis de Tocqueville, Democracy in Amer ica, trans. George Lawrence (Chicago: En-
cyclopedia Britannica, 1990), 1.
Unauthenticated
The first, and most serious, is that “universal grammar, in Chomsky’s

sense, has to provide for all possible concepts. (Chomsky likens the “lan-
guage organ” to a switch box, and he tells us that when the “switches” in
a child’s brain are appropriately set, “the child has command of a partic-
ular language and knows the facts of the language, that a particular ex-
pression has a particular meaning and so forth.”)10 Now, if all possible
concepts were definable from a reasonably small number of concepts that
might have been selected for by evolution—say, observation concepts—
then this might be compatible with evolutionary psychology. But the pic-
ture of all our concepts as reducible to observation concepts is precisely
what Quine demolished (and Chomsky himself certainly wants no part
of this positivist picture). Even if Hauser only wants all moral concepts to
be reducible to some determinate number, any attempt to specify such a
reduction basis will be refutable by showing that the proposed defini-
tions of the innumerable different moral concepts that different cultures
actually have in terms of the alleged “basic” ones are riddled with coun-
terexamples. (That proposed “analyses” of concepts have counterexam-
ples is something “analytic philosophers” have become very good at
showing, which is why the very term “analytic” philosophy has become
somewhat of a misnomer, as Jerry Fodor has pointed out.)
A second problem, perhaps less serious in the eyes of those who, like
Steven Pinker and possibly Hauser, hold on to the idea that concepts
are innate, is that Chomsky himself does not think that evolutionary
explanations of the existence of the “language organ” work; he famously
thinks it is simply serendipitous that we have the language organ we
have.11 The whole idea of innate language is, needless to say, highly
controversial, and I continue to represent the negative pole of the con-
troversy. But evolutionary psychology already labors under the unfair
accusation of being a mere collection of just-so stories. It must not
burden itself with the science-fiction idea of a “language organ” with a
code for all possible concepts.
10. Noam Chomsky, Language and the Problems of Knowledge (Cambridge, Mass.: MIT
Press, 1988), 62–63.
11. However, if, as Chomsky claims in Language and the Problems of Knowledge, all pos-
sible “par tic u lar meanings” are defi nable in terms of the positions of a fi nite number of
“switches” in the “language organ” (62–63), then it would seem as if something like the
explicit defi nability of all concepts from some fairly limited basic vocabulary is still presup-
posed by Chomsky’s picture.
Unauthenticated
D4
Sosa on Internal Realism
and Conceptual Relativity
1
E r n e s t S os a’s 1993 paper, “Putnam’s Pragmatic Real-
ism,” is one of the most important of the many papers that were pub-
lished analyzing and criticizing the “internal realism” that I defended
in books and papers published between 1976 and 1989.2 Not only is
the argumentation outstanding, as is the case in all of Ernest Sosa’s
work, but the scope of the discussion is extraordinarily comprehensive.
1. Ernest Sosa, “Putnam’s Pragmatic Realism,” Journal of Philosophy 90 (1993): 605–626.

2. My “internal realism” was fi rst announced in “Realism and Reason,” my presidential
address to the Eastern Division of the American Philosophical Association (Boston, Mass.,
December 29, 1976), reprinted in my Meaning and the Moral Sciences (London: Routledge
and Kegan Paul, 1978), 123–138; the position was further elaborated and developed in
Reason, Truth and History (Cambridge: Cambridge University Press, 1981), in the papers
collected as Realism and Reason, vol. 3 of Philosophical Papers (Cambridge: Cambridge Uni-
versity Press, 1983), and in The Many Faces of Realism (LaSalle, Ill.: Open Court, 1987), and
the concluding chapter of Representation and Reality (Cambridge, Mass.: MIT Press, 1988).
I fi rst renounced the identification of truth with “idealized rational acceptability,” which
had been a central element of that position (although I retained— and still retain— another
element that I called “conceptual relativity”), in my reply to Simon Blackburn at the Gif-
ford Conference on my philosophy at the University of St. Andrews (November 1990); the
proceedings, including that reply, are published as Reading Putnam, ed. P. Clarke and R.
Hale (Oxford: Basil Blackwell, 1993), and I repeated that renunciation in more detail in my
reply to David Anderson in the issue of Philosophical Topics 20, no. 1 (1992) devoted to my
philosophy. The position that I sketched in the reply to Anderson was later developed as my
Dewey Lectures, “Sense, Nonsense and the Senses,” Journal of Philosophy 91 (1994): 445–
517, collected as part I of The Threefold Cord (New York: Columbia University Press, 1999).
67
Unauthenticated
68 REALISM AND ONTOLOGY
If I did not reply to the paper at the time, the reason is that I had only
relatively recently given up some of the views that Sosa criticized, and
was still in the process of working out my present position. But I wel-
come the opportunity that I very much value to explain just which of
Sosa’s criticisms I accept and which I believe that I can meet. I look for-
ward very much to learning his response!
The Model-Theoretic Argument

Sosa sets the stage for his discussion of my “internal realism” with a
quotation from a paper of Donald Davidson’s.3 By internal realism I
had in mind not just that the truth of sentences or utterances is relative
to a language. That much, Davidson observed, was “familiar and triv-
ially correct.” But, he continued, “Putnam seems to have more in
mind—for example that a sentence of yours and a sentence of mine may
contradict each other, and yet each be true ‘for the speaker.’ ” It is hard
to think in what language this position can be coherently, much less
persuasively, expressed.” 4 Sosa opens his paper by asking, “What
arguments might lead to such a view?”5 and picks four that, as he says,
“stand out.” “First, the “model-theoretic’ argument; second, the argu-
ment from the nonobjectivity of reference and of the sort of causation
involved in contemporary accounts of reference; third, the argument
from the unlikelihood of scientific convergence on a finished science
that provides an objective and absolute conception of reality; and, fi-
nally, the argument from the nonabsoluteness of objecthood and of
existence.”6
Before I discuss what Sosa has to say about these arguments, permit
me to object strongly to Davidson’s report of my view. The place where
3. Donald Davidson, “The Structure and Content of Truth,” Journal of Philosophy 87

(1990): 279–328.
4. What is strange about this paper of Davidson’s is that I had heard Davidson present
these arguments (at a conference at the Universidad Autonoma de Mexico) several years
earlier and replied to them not only on the spot but also in print in “Truth and Convention:
On Davidson’s Refutation of Conceptual Relativism,” Dialectica 41 (1987): 69–77; collected
as “Truth and Convention,” in my Realism with a Human Face, ed. James Conant (Cam-
bridge, Mass.: Harvard University Press, 1990), 96–104. Davidson’s paper ignored my re-
buttal entirely.
5. Sosa, “Putnam’s Pragmatic Realism,” 606.
Unauthenticated
Sosa on Internal Realism and Conceptual Relativity 69
I stated the view Davidson was attacking (the view I called “conceptual
relativity”) was my first Carus Lecture,7 and nothing like the assertion
that “a sentence of yours and a sentence of mine may contradict each
other, and yet each be true ‘for the speaker’ ” occurs in that lecture. (In
fact, neither the notion of contradiction nor the notion true for the speaker
occurs.) What does occur is an example: the example of a world in which
the number of objects will be said to be three if one disallows mereolog-
ical sums as objects and seven if one allows them, and a discussion in
which I claimed that it is absurd to suppose that there is some kind of
metaphysical fact of the matter as to whether “mereological sums really
exist.” We can extend our language so that we speak of such things, and
in the sense of “exist” and “object” we thereby create, it will be true that
“there are such objects as mereological sums,” or we can refuse to extend
our language in that way; and both procedures are legitimate. I thereby
drew the conclusion that the logical primitives themselves, and in particular
the notions of object and existence, have a multitude of dif ferent uses rather
than one absolute “meaning.” What this implies is that the sentence “There
are seven objects (in the relevant world)” (spoken by a “Polish logician”
who has introduced mereological sums into his language) and the sen-
tence “There are three objects (in the relevant world)” (spoken by my
fictitious “Carnap”) do not “contradict” each other. Moreover, as I ex-
plained in the follow-up paper “Truth and Convention,” it does not help
to say that the difference in the use of “object” in the two languages is a
“difference in meaning” either—especially if the criterion for sameness /
difference of meaning is supposed to be “translation practice.” But I
shall return to this issue below.
In contrast, the model-theoretic argument was not an argument for
conceptual relativity, but rather an argument for the identification of
truth with “idealized rational acceptability.”8 In his essay, Sosa decided
to “set the controversy [provoked by that argument] aside, as one with
little prospect of any new progress or insight beyond what is already
contained in the literature about it,” and “to discuss instead, and in turn,
7. These were published as The Many Faces of Realism. “Conceptual relativity” is ex-
plained on pp. 16–21.
8. My internal realist notion of truth coincided with the notion that Crispin Wright
calls “superassertibility” in his Truth and Objectivity (Cambridge, Mass.: Harvard Univer-
sity Press, 1992) (although that is not how he interprets me!), if we prescind from possible
differences over the notion of “assertibility.”
Unauthenticated
the other three arguments that sustain Putnam’s pragmatic realism.”9

Since Sosa does not discuss the model-theoretic argument, I shall re-
frain from discussing it as well.10
The “Second Argument”

The second of the four arguments attributed to me by Sosa he describes
as “the argument from the nonobjectivity of reference and of the sort of
causation involved in contemporary accounts of reference.”11 He gives
what he calls “a thumbnail sketch” of this argument (which he attri-
butes to my “Why There Isn’t a Ready-Made World”)12 as follows:13
1. Truth depends on, and is constituted by, reference (at least in

part).
2. Reference depends on, and is constituted by, causation (at least
partly).
3. Causation is radically perspectival.
4. Reference is radically perspectival (from 2, 3).
5. Truth is radically perspectival (from 1, 4).
6. Reality is “internal” to one’s perspective (from 5).
As I will now explain, I do not recognize this argument in my writing.

Nevertheless, what Sosa says in this section of his paper,14 and indeed
in each of the sections that follow, is impor tant to consider carefully.
Before I consider what Sosa says about the argument he attributes
to me, I wish to say briefly what I thought I was doing in the papers in
which he perceives this argument. The philosophers I was debating
against subscribe to the kind of scientism that Bernard Williams de-
fended in Descartes: The Project of Pure Enquiry and later in Ethics and
the Limits of Philosophy. That means that they thought the description
of the world as it is in itself, mind-independently (Williams calls it
9. Note that Sosa here uses “pragmatic realism” as synonymous with “internal realism.”
I would today describe myself happily as a pragmatic realist, but not as an “internal
realist.”
10. My reasons for giving up “internal realism” are given in detail in “Sense, Nonsense
and the Senses.”
11. Putnam, “Sense, Nonsense and the Senses,” 606.
12. In Putnam, Realism and Reason, 205–228.
13. Putnam, “Sense, Nonsense and the Senses,” 607.
14. Sosa, “Putnam’s Pragmatic Realism,” 607–608.
Unauthenticated
“the absolute conception”) could be given in the language of a per-

fected natu ral science, indeed, in physicalistic terms. In strikingly
similar language, in fact, both Michael Devitt and Clark Glymour
asserted that reference is some kind of “causal connection.” And they
conceived of “causal connection” as something that holds between
(mind-independent) physical events mind-independently. In short, they
held that the semantical (or, as I prefer to say, the “intentional”) can be reduced
to the nonsemantical.
In the context of this debate, it was natural that I should attack this
reductionist claim. In “Why There Isn’t a Ready-Made World,” how-
ever, I envisaged only two possible positions: either my own “internal
realism,” with its “verificationist semantics,” or a materialist version of
metaphysical realism (a dichotomy I blush at today). In Sosa’s recon-
struction of that essay, however, point 3 reads “Causation is radically
perspectival” and what follows from 3, together with 2, is “Reference is
radically perspectival.” But I did not argue in that paper that the context-
and-interest relativity of “the cause” leads to a context-and-interest
relativity of “refers.”15 (“Internal realism,” which I did argue for, does
not automatically imply a relativity of “refers.”) What I did argue is that
“the cause” means, roughly, “that part of the total cause that may rea-
sonably, given the interests appropriate to the context, be regarded as
the bringer about as opposed to a background condition,” and hence
that concept “the cause” involves something intentional.
Thus, from my point of view, Sosa’s point 4 should have read: “Ref-
erence is intentional,” his 5 should have read “Truth is intentional” and
his 6 should not have been there at all.
One philosopher who saw that that is what I was arguing and who
was deeply concerned to rebut my arguments is Jerry Fodor, who wrote:
I have helped myself to the notion of . . . one event being the cause
of another. I have therefore to claim that whatever the right
15. It might be thought, after all, that all the interest-relativity is already allowed for in
the fact that reference is relative to a context, and that there will be no possibility of dif-
ferent answers to “What does word W refer to?” once the context in which W was used has
been specified. Actually, I do not think this is the case: I think that the answer to this
question may depend on who is asking and for what reason and not only on the context of
the user of the word W in question. But this is something I did not argue in the paper
Sosa refers to.
Unauthenticated
unpacking of [this concept] may be, it doesn’t smuggle in inten-

tional / semantic notions.16
That was precisely what my argument was designed to show: that “the
cause” does “smuggle in intentional / semantic notions,” and hence the only
way that materialist metaphysical realists have suggested for avoiding
the model-theoretic argument doesn’t work.
Given his reading of my argument, however, I find it natural that Sosa
writes:
Perhaps it is true that our concepts of reference and truth are

ineliminably perspectival. Even so it still would not follow that
reality itself could not be largely as it is independently of us and
our thought, in the sense that plenty of real ity could not have
existed propertied and interrelated very extensively just as it is in
fact propertied and interrelated even if we had never existed to
have any thoughts, and even if no other finite thinkers had existed
to take our place. What is more, our perspectival references and
truths may be seen to derive necessarily from absolute and unper-
spectival reality.17
But I never claimed that “truth is perspectival,” as far as I can see,

although how we express it is certainly dependent on perspective.
Moreover, no antirealist (including my former self ) denies that,
for example, the moon exists “independently” of our perceptions
and thoughts in the sense of having existed before there were per-
ceptions and thoughts. Its existence was not caused by perceptions
and thoughts.18 To pretend that advocates of antirealist (or “veri-
ficationist”) semantics deny the causal independence of the moon or
the solar system from human thoughts and perceptions is to misde-
scribe their position.
16. Jerry Fodor, Psychosemantics (Cambridge, Mass.: MIT Press, 1988), 126.
18. Sosa, “Putnam’s Pragmatic Realism,” 609. In Realism with a Human Face— a work
Sosa cites in his essay— I write, “It is a part of [our image of the world] itself that the world
is not the product of our will—or our dispositions to talk in certain ways either” (29). Com-
pare Sosa’s own distinction between existing relative to a conceptual scheme and existing in
virtue of that conceptual scheme (“Putnam’s Pragmatic Realism,” 621).
Unauthenticated
Sosa continues his analysis and criticism of my arguments from the

“nonobjectivity of reference” [Sosa’s terminology, not mine]19 in a section
titled “II. Objectivity, Absoluteness and the Many Faces of Realism.”20
Here again he repeats the charge of fallacy, this time in connection with
the criticisms I made many years later of Bernard Williams’s position:
What the metaphysical realist is committed to holding is that there

is an in-itself reality independent of our minds and even of our ex-
istence, and that we can talk about such reality and its constitu-
ents by virtue of correspondence relations between our language
(and / or our minds) on the one hand, and things-in-themselves
and their intrinsic properties (including their relations) on the
other. This does not commit the metaphysical realist to holding
that reference itself (or correspondence, or causal explanation) is
among the objective properties constitutive of in-itself reality.21
What Sosa claimed is that I overlooked the possibility that there

could be an “in itself” real ity [in Williams’s sense] even if reference,
truth, correspondence, and so on, are all “perspectival” in the sense of
being relative to interests (and other features of the contexts in which
we think and speak). However, Williams’s position is that the descrip-
tion of the world in terms of its nonperspectival properties is a complete
description. As I explain in Chapter 2, that means that if it is a fact that
some speaker referred to object X from such-and-such a perspective,
then that whole fact, including the perspective and the object X and the
relation between the speaker and X, must somehow appear in the “abso-
lute conception,” and so Williams does need to somehow reduce facts
that involve intentionality to pure physical facts if he is not to become
a sheer eliminationist with respect to the intentional. I think it is because
19. I do not myself see why the fact that something is interest-relative need mean that it
is nonobjective. As Richard Boyd (surely no “antirealist”!) once remarked, reference can be
interest-relative, but that fact that a word refers to X given certain interests is perfectly
objective.
21. Sosa cites my “Objectivity and the Science / Ethics Distinction” (Putnam, Realism
with a Human Face, 174) as the locus of the alleged fallacy. Note that this was published in
1990, and written after I had given up “internal realism.” Its purpose was to defend the ob-
jectivity of ethics, not to argue that “real ity is perspectival,” however that might be
understood.
Unauthenticated
Sosa himself does not feel the appeal of the idea of reducing the inten-
tional to the nonintentional (nor the appeal of denying its existence al-
together) that my physicalist opponents are so attracted to that he
(charitably) misdescribes Williams’s views.
One sees such a charitable misdescription when Sosa writes,
There is nevertheless an argument open to Williams’s view if the

latter includes commitment to “objectivism,” which is defined by
Putnam . . . as the view that what really has a place in objective reality
is only what is included in the ontology and the ideology of “finished sci-
ence,” only what the absolute conception recognizes.22
Sosa immediately adds “It is not at all clear that Williams himself would
accept objectivism.” In fact, not only is it clear that Williams accepts
objectivism, but he wrote a whole book the gravamen of which is that
objectivism is the lasting element of truth in Descartes philosophy!23
“Finished Science”
As Sosa points out, in my criticism of Bernard Williams’s views I
argued that
there is no evidence at all for the claim . . . that science converges

to a single theory. We simply do not have the evidence to justify
speculation as to whether or not science is “destined’ to converge
to some one defi nite theoretical picture . . . Mathematics and
physics, as well as ethics and history and politics, show our con-
ceptual choices; the world is not going to impose a single language
upon us, no matter what we choose to talk about.24
(As examples of questions on which we cannot expect science to dic-

tate a single answer I listed whether stones are identical with mere-
22. Sosa, “Putnam’s Pragmatic Realism,” 610. Here Sosa is quoting from The Many Faces
of Realism, p. 4 (emphasis added).
23. Bernard Williams, Descartes: The Project of Pure Enquiry (Harmondsworth, United
Kingdom: Penguin, 1978). See Chapter 2 in this collection for a description of Williams’s
agreements with it.
Unauthenticated
ological sums of particle-time-slices and whether points are individ-

uals or limits.) In his essay, Sosa reconstructs this argument as
follows:25
a. There is no real possibility of a finished science.

b. Things-in-themselves are by defi nition the things in the
ontology of a finished science, and intrinsic properties are by
definition those in the ideology of finished science.26
c. Hence, there is no possibility that that there are things-in-
themselves with intrinsic, objective properties.
I do have to concede that I argued badly.

What I should have said was that the question “Do mereological
sums exist?” is not a scientifi c question at all, and not that science
isn’t going to “converge” to an answer. “Science” couldn’t care less
whether we quantify over mereological sums or not, or whether we
take points to be individuals or (as Whitehead and Russell did) to be
limits, or, to shift to a mathematical example, whether we take sets
as primitive (and identify functions with sets of ordered pairs), or
we take functions and numbers as primitive (and identify sets with
“characteristic functions,” as is customary in recursion theory), or
take functions and numbers and sets as all primitive. Even if science
“converges” it isn’t going to converge to one single “ontology” and “ide-
ology.” (But that doesn’t mean there aren’t other senses in which it
may well converge.)
I was mistaken to write as if “one defi nite theoretical picture”
required one single ontology and one single ideology (i.e., as if theories
did not have a number of alternative versions— a point that I myself
stressed in other writings). I believe that this mistake accounts for So-
sa’s attribution of “a” to me. But “b” is a proposition (indeed, the most
important one) that Bernard Williams argues for. I absolutely do not see
why Sosa thinks that I agree with it. I have argued against “things-in-
themselves” in various places, but always, I believe, in the context of
some debate, and then the term was to be understood as the particular

26. Sosa also reads the following “defi nition” of subjective into my writings (p. 612): “ϕ is
a subjective property = Df ↗ is postulated by a par tic u lar language or conceptual scheme.”
This would commit me to the view that all the properties we ever talk about are
“subjective”!
Unauthenticated
opponent (who might not be a realist at all) understood it. What I reject
is not the idea of mind-independent things (in the sense of things caus-
ally independent of the mind), but (1) the idea that there is one single
metaphysically privileged use of “thing” (or “object,” or “entity”), and
(2) the idea that there is a fact of the matter as to such questions as “Is
a table identical with the mereological sum of its time-slices?” But I
would not express this by saying “There are no things in themselves,”
because I don’t think any of the metaphysical uses of the notion I have
seen to date are intelligible.
The situation is similar with respect to “intrinsic properties.” I would
not define them in terms of Williams’s concept of finished science (or
rather, asymptotically approachable finished science). Actually, I would
say that when people talk about “intrinsic properties” they generally
suppose them to be essential properties in the Aristotelian sense
(properties without which something would not be the thing that it is),
and also supposes them to be interest-independent. And I don’t think
that there is a definite set of properties possessed by, say, dogs that are the
“intrinsic properties” of dogs interest-independently. What is “essential”
to being a dog from the point of view of a molecular biologist is not
what is “essential” from the point of view of an evolutionary biolo-
gist, nor what is “essential” from the point of view of someone who is
interested in dogs as pets. (I argue this in detail in “Aristotle after
Wittgenstein”).27
So I throw the ball back to Sosa in the following sense: I say,
“Ernie, you want to read me as defi ning these metaphysical notions
(‘thing-in-itself,’ ‘intrinsic property’), in fact defi ning them the way
Bernard Williams did, and then asserting ‘there are no things-in-
themselves with intrinsic properties.’ But I don’t think these notions
are intelligible (as used by metaphysicians), nor do I think that all the
dif ferent (unfortunate) ways they have been used are captured by Wil-
liams’s definition. I don’t want to either assert or deny the thesis that
‘there are things-in-themselves with intrinsic properties.’ So do we have
any remaining disagreement about this issue?” If the answer is “yes,” I
suspect the remaining disagreement(s) will come up in what I shall say
now about conceptual relativity.
27. Collected in Hilary Putnam, Words and Life, ed. James Conant (Cambridge, Mass.:
Harvard University Press, 1994), 62–81.
Unauthenticated
Conceptual Relativity
As Sosa describes, in The Many Faces of Realism, I used the following
example (I quote Sosa’s presentation of the example, which is quite
accurate):28
Suppose a world with just three individuals x1, x2, x3. Such a world
is held by some “mereologists” to have in it a total of seven things
or entities or objects, namely, x1, x2, x3, x1 + x2, x1 + x3, x2 + x3,
x1 + x2 + x3. Antimereologists by contrast prefer the more austere
ontology that recognizes only the three individuals as objects that
really exist in that world. Talk of the existence of x1 + x2 and its ilk
is just convenient abbreviation of a more complex discourse that
refers to nothing but individuals. Thus, suppose x1 is wholly red
and x2 is wholly black. And consider
1. There is an object that is partly red and partly black.

2. There is an object that is red and an object that is black.
For the antimereologist, statement 1 is not true, if we assume that

x3 is also wholly red or wholly black; it is at best a convenient way
of abbreviating the likes of 2.
Sosa goes on to quote my response (which, as he correctly points

out, was in agreement with Carnap’s views on similar questions),
namely, that “the question is one of the choice of a language. On
some days it may be more convenient to use [antimereological] lan-
guage: . . . on other days it may be convenient to use [mereological]
language.”29
Explaining this answer, Sosa writes:30
Take the question,

How many objects with a volume of at least 6 cubic centimeters
are there in this container?

29. The quotation is from my “Truth and Convention: On Davidson’s Refutation of
Conceptual Relativism,” in Dialectica 41 (1987): 69–77; quotation from p. 75.
Unauthenticated
This question can have no absolute answer on the Carnap-Putnam

view, even in a case where the container contains a vacuum except
for three marbles each with a volume of 6 cubic centimeters. The
antimereologist may say
3. There are three objects in the box.
But the mereologist will reply
4. There are at least seven objects in the box.
The Carnap-Putnam line is now this: which statement we

accept—3 or 4— is a matter of linguistic convenience. The language
of mereology has criteria of existence and identity according to
which sums of individuals are objects. The language of antimere-
ology rejects such criteria, and may even claim that by its criteria
only individuals are objects.
Sosa begins the first of his criticisms with the following words: “There
is a valuable insight here, I believe, but I am puzzled by the linguistic
wrapping in which it is offered.”31
After saying “I am puzzled by the linguistic wrapping,” Sosa
continues,
After all, none of 1–4 mentions any language, or any piece of lan-
guage, nor does any of them say we shall or shall not or should or
should not use any language or bit of language. So I do not see how
our decision actually to use or not to use any or all of the sentences
1–4 can settle the question of whether what these sentences say is
true or false.
My reply to this objection is that what settles the question whether

what these sentences say is true or false is not merely our decision to use
(assert?) any or all of them, but our adoption of what Sosa himself called
“the criteria of existence and identity” of mereology or “the criteria of
31. Sosa, “Putnam’s Pragmatic Realism,” 615. However, there is one (possibly quite con-
sequential) change I would make in the above explanation of “the Carnap-Putnam line”: in
the last sentence, I would change the last clause to read “and may even claim that by its
criteria there are no such objects as the ‘sums’ x1 + x2, x1 + x3, x2 + x3, x1 + x2 + x3.”
Unauthenticated
existence and identity” of antimereology, together with certain empirical

facts. The way that works is as follows: if the Eiffel Tower does not exist
(and that, I agree with Sosa, is not a question with respect to which there
is any “conceptual relativity”) or if the Statue of Liberty does not exist
(ditto), then the mereological sum of the Eiffel Tower and the Statue of Liberty
also does not exist (no matter which of the criteria we adopt). But if they
do both exist, then if we adopt the mereological criteria of existence and iden-
tity, then we have adopted conventions of language that make it trivially correct
to say that the mereological sum of the Eiffel Tower and the Statue of Liberty
exist. The example itself was meant to illustrate precisely how there can
be a choice between different uses of “exist,” on some of which it is true
to say that mereological sums exist, while on others it is false.
[“But Hilary, how can you talk of conventions after Quine?” I can
imagine my old friend Burton Dreben (and not only Burton Dreben)
exclaiming. The answer here, as I explained long ago,32 is that while I
find the notion of convention indispensable, I do not explain it in terms
of the Carnapian notion of “analyticity.” What is and what is not a
matter of convention is something on which we may change our minds,
and empirical facts may turn out to be relevant. But I do not agree with
Quine that the notion is simply to be discarded.]
Perhaps anticipating some such response, Sosa immediately suggests
that a linguistic formulation of the doctrine of conceptual relativity
would render it trivial. He writes:
Here for a start is a possibility [i.e., a possible interpretation]:

LR1. In order to say anything you must adopt a language. So you
must “adopt a meaning” even for so basic a term as “object.” And
you might have adopted another. Thus you might adopt Carnap-
language (CL) or you might adopt Polish-logician language (PL).
What you say, i.e., the utterances you make, the sentences you
affirm, are not true or false absolutely, but are true or false only rel-
ative to a given language. Thus, if you say “There are three objects
in this box” your utterance or sentence may be true understood as a
statement of CL while it is false understood as a statement in PL.33
32. See Hilary Putnam, “Convention, a Theme in Philosophy,” in Realism and Reason,
vol. 3 of Philosophical Papers (Cambridge: Cambridge University Press, 1983), 170–183.
Unauthenticated
But under this interpretation linguistic relativity seems trivially true.

Who could deny that inscriptions of shapes and emissions of sounds
are not true or false independently of their meaning, independently of
all relativization to language or idiolect?
My reply to this consists of three points:
1. The speaker of, say, PL does not do anything that would ordi-
narily be called giving (or “adopting”) a meaning to the word
“object” (if this is not clear, substitute “entity”). When he says
that there are such objects (or such entities) as mereological
sums, he counts, at least for linguistic purposes, as simply using
“object” (“entity”) in the normal (Anglo-American) way. So the
trivial linguistic truth that the truth-value of our utterances
depends on the meanings we give to their words (or that our
linguistic community has already given them) is not the same as
the thesis of conceptual relativity that I affirmed above, unless
“meaning” is already being given a special philosophical
interpretation.
2. To see that it is not trivially true that if we adopt CL we thereby
make “There are seven objects in the box Sosa described [the
one with 3 marbles in a vacuum]” true, consider the question
from the standpoint of a metaphysical realist who does not believe
in the existence of mereological sums. (I called him “Professor
Antipode” in The Many Faces of Realism.) Obviously Professor
Antipode will say something like this: “I don’t mind your saying
that when you use the word ‘object’ you mean to include mereo-
logical sums as objects. But that doesn’t make ‘Mereological
sums exist’ true, any more than saying ‘When I use the word
“object” I mean to include leprechauns’ makes ‘Leprechauns
exist’ true.”
3. On the other hand, according to my own unmetaphysical sort of
realism, adopting the conventions of PL does make it true to say
(in PL) “Mereological sums exist,” and adopting the conventions
of (CL) makes it true to say “Only three objects exist” [in the
relevant world], and a fortiori that mereological sums do not
exist. Whether I am right in this claim or not is not an instance
of trivial linguistic conventionality, as Professor Antipode’s
argument shows.
Unauthenticated
Sosa’s “Nonlinguistic” Restatement

After expressing dissatisfaction in this way with “the linguistic turn
taken by Carnap and now Putnam,”34 Sosa moves to a more positive
assessment, writing, “Nevertheless, it still seems to me that there is a
valuable insight in Putnam’s now repeated appeal to the contrast be-
tween the Carnapian conceptual scheme and that of the Polish logician.
But, given our recent reflections, I would like to put the insight without
appeal to language or to any linguistic relativity.”
The insight, as expressed at the close of Sosa’s essay,35 is that by
extending my reasoning, we reach a set of options in contemporary
ontology that present us with “a rather troubling trilemma.”
To comment on these words, I need to explain Sosa’s trilemma, which
he presents with the aid of an example:
I am supposing a snowball to be constituted by a certain piece of

snow as constituent matter and the shape of (approximate) round-
ness as constituent form. That par tic u lar snowball exists at that
time because of the roundness of that par tic u lar piece of snow.
More, if at that time that piece of snow were to lose its roundness,
then at that time that snowball would go out of existence.
Compare now with our ordinary concept of a snowball, the con-
cept of a snowdiscall, defi ned as an entity constituted by a piece of
snow as matter and as form any shape between being round and
being discshaped. At any given time, therefore, any piece of snow
that constitutes a snowball constitutes a snowdiscall, but a piece
of snow might constitute a snowdiscall without then constituting
a snowball. For every round piece of snow is also in shape between
being discshaped and round (inclusive), but a discshaped piece of
snow is of course not round.
Any snowball SB must hence be constituted by a piece of
snow PS which also then constitutes a snowdiscall SD. Now, SB
is distinct (a dif ferent entity) from PS, since PS would survive
squashing and SB would not. By similar reasoning, SD is also
distinct from PS. And, again by similar reasoning, SB must also

Unauthenticated
be distinct from SD, since enough partial flattening of PS will

destroy SB but not SD. Now, there are infi nitely many shapes
S1, S2 . . . between roundness and flatness of a piece of snow,
and for each i, having a shape between flatness and Si would give
the form of a distinctive kind of entity to be compared with snow-
balls and discballs. Whenever a piece of snow constitutes a snowball,
therefore, it constitutes infi nitely many entities all sharing its
place with it.
Under a broadly Aristotelian conception, therefore, the barest
flutter of the smallest leaf hence creates and destroys infi nitely
many things, and ordinary reality suffers a sort of “explosion.”36
The first of the three responses to this threat of ontological “explo-

sion” that constitutes the “trilemma” Sosa himself calls “conceptual rel-
ativity,” and he explains it as follows:
Perhaps snowballs do exist relative to all actual conceptual schemes

ever, but not relative to all conceivable conceptual schemes. Just
as we are not willing to countenance the existence of snowdiscalls,
just so another culture might have been unwilling to countenance
snowballs. We do not countenance snowdiscalls because our con-
ceptual scheme does not give to the snowdiscall form (being in
shape between round and disc-shaped) the status required for it to
be a proper constitutive form of a separate sort of entity—at least
not with snow as underlying stuff.37
And Sosa points out that:
That would block the explosion of real ity, but the price is con-
ceptual relativity. Supervenient, constituted entities do not just
exist or not in themselves, free of any dependence on or rela-
tivity to conceptual scheme. What thus exists relative to one
conceptual scheme may not do so relative to another. In order for
such a sort of entity to exist relative to a conceptual scheme, that

Unauthenticated
conceptual scheme must recognize its constituent form as an ap-

propriate way for a separate sort of entity to be constituted.38
Sosa now considers a possible objection to this first response, which

he promptly rebuts:39 mustn’t we think of the existence of the framers and
users of the conceptual scheme as also relative to that conceptual scheme? “Are
we then not caught in a vicious circle?” And he replies that “existence
relative to a conceptual scheme is not equivalent to existence in virtue of
that conceptual scheme. Relative to scheme C the framers of C exist in
virtue of their constitutive matter and form, and in virtue of how these
satisfy certain criteria for existence and perdurance of such subjects
(among whom happen to be the framers themselves). There is hence
no vicious circularity.” And he sums up this first response (conceptual
relativity) thus:
The picture is then roughly this. Each of us acquires and develops

a view of things that includes criteria of existence and perdurance
for categories of objects. When we consider whether an object of
a certain sort exists, the specification of the sort will include the
relevant criteria of existence and perdurance. And when we cor-
rectly recognize that an object of that sort does exist, our claim is
elliptical for “. . . exists relative to this our conceptual scheme.”40
Comment: Both Ernest Sosa and Jennifer Case have noted that all
the examples I gave (and, I might add, that Sosa now gives) of “concep-
tual relativity” involve what Sosa called “recondite entities” (he added
“of controversial status,” but I have explained above why I don’t think
that is always the case). And Case went on to make an impor tant
suggestion:
Reading Davidson’s discussion of conceptual schemes as contra-

vening Putnam’s agenda requires overlooking the difference be-
tween natural languages and languages like Carnap’s and the Polish
Logician’s. For lack of a better term, let me call languages of the

Unauthenticated
latter variety “optional languages.” If having a conceptual scheme

is to be associated with having a language, it should be associated
with having an optional language. Modifying a remark of David-
son’s, we may say that where conceptual schemes differ, so do
optional languages.
It is not necessarily the case that where conceptual schemes
differ so do natural languages. Someone who has a single natural
language may have multiple optional languages and, therefore,
multiple conceptual schemes.41
If one looks at the matter this way, as I think we should, then one
will not say that “we” (or “our culture”) do not countenance snowdiscalls
or mereological sums. Our culture allows us to do dif ferent things in
dif ferent contexts, including introducing, if we want, an optional lan-
guage in which we quantify over snowdiscalls—and “snowdiscalls” are,
after all, no more unusual than, say, “the mereological sum of my nose
and the Eiffel Tower.”
Second, I don’t think it is happy to say that “mereological sums exist
relative to the Polish Logician’s language” (or “snowdiscalls exist
relative to the SD scheme”). If we use PL, in some context and for some
appropriate reason, then we should simply say, “There is an object that
is the mereological sum of my nose and the Eiffel Tower,” or “The
mereological sum of my nose and the Eiffel Tower exists.” We do not
have to relativize existence to PL in PL. What we have to do is make
clear which optional language we are speaking.
“But then what you count as the same sentence in English may have
dif ferent truth-conditions in dif ferent contexts!” That is the case
anyway! (“Tomatoes are vegetables” has dif ferent truth-conditions in
the mouth of a grocer and the mouth of a botanist, for example, and
none is disturbed by this.)
Third, I do not see why Sosa thinks he has “put the insight without
appeal to language or to any linguistic relativity.” Talk of criteria of ex-
istence and perdurance is metalinguistic talk on the face of it. (Perhaps
Sosa read me as advocating that we must always restate existence
claims in the formal mode, as Carnap did. But this was not the aspect
41. Jennifer Case, “On the Very Idea of a Conceptual Scheme,” Southern Journal of Phi-
losophy 35, no. 1 (1997): 1–18; quotation from p. 11.
Unauthenticated
of Carnap’s view that I endorsed. What I endorsed was his tolerant atti-
tude to a plurality of—to use Jennifer Case’s term— optional languages,
with dif ferent so-called ontologies.)
Fourth, as already mentioned, I think that—very importantly—
conceptual relativity should be our approach not only to questions of
existence and perdurance, but to many questions having to do with
cross-category identification (such as “Are points mere limits?,” or “Are
functions sets of ordered pairs?”).42
Living with the Explosion

A second approach to the problem illustrated by the snowdiscall example
is what Sosa calls “try[ing] to live with the explosion.” This is his term43
for just saying that all the objects in all the alternative conceptual
schemes (what I would now, taking Case’s suggestion, call all the alter-
native optional languages) are genuine elements of reality. There are,
we will say if we take this line, snowballs and snowdiscalls and whose
knows what else besides? (“Possible worlds?”)
As Sosa points out, however,
If we allow the satisfaction by any sequence S of any form F of the

appropriate polyadicity and logical form to count as a criterion of
existence for a corresponding sort of object, then reality right in
front of us, before us, and all around us is unimaginably richer and
more bizarre than we have ever imagined. And any way we shall still
face the problem of giving some explanation of why we focus so
narrowly on the objects we do attend to, whose criteria of existence
and perdurance we do recognize, to the exclusion of the plethora of
other objects all around and even in the very same place.44
Comment: “Trying to live with the explosion” is tremendously costly

for additional reasons having to do with the problem of criteria of
42. A possible fifth point: as I say in a passage that Sosa quotes, I think yet another ap-
proach to the kind of problem he raises with the snowdiscalls example is sortal identity. I
know that both Kripke and Quine shudder, but isn’t just this sort of example a good reason
to reconsider the shudders?
43. Sosa, “Putnam’s Pragmatic Realism,” 621
Unauthenticated
cross- category identity that I mentioned a moment ago. Consider the fact
that I have mentioned a couple of times in this essay that in recursion
theory and hierarchy theory (one of the branches of mathematics I
wrote quite a few articles in) we regularly take numbers as individuals,
and functions of numbers, functions of functions, functions of func-
tions of functions . . . (as in type theory, but extended through the
transfinite) as primitive. We do not take sets to be still additional enti-
ties, but “identify them” with characteristic functions. In another
branch of mathematics (set theory), we regularly take sets as primi-
tive, and identify functions with sets of ordered pairs. I have never
met a philosopher or a mathematician who thinks there is a “fact of the
matter” as to which is right! Yet they can’t both be right.45 According to
“conceptual relativity,” there is only the question of a choice of an “op-
tional language” here. But if we go for the “exploding reality” approach,
what do we do? Do we say there is a (possibly unknowable) “fact of the
matter” as to whether sets are characteristic functions or functions are
sets of ordered pairs? That, I must admit, seems crazy to me!46
Nor does the problem arise only in pure mathematics. Typically,
when we find a way of “interpreting” one version of a physical (or geo-
metrical) theory in another, the dif ferent alternative “translations” are
incompatible if taken at face value. For example, even if we decide that
points are limits, there are just as many ways of formalizing the notion
of a “limit” as there are of formalizing the notion of a “set.” Yet surely
the adoption of one or another way is a choice of a linguistic option,
and not a metaphysical claim.
The Last Option: Eliminativism

The third of the three options that constitutes Sosa’s trilemma is
“eliminativism”:
A third option is a disappearance or elimination theory that re-

fuses to countenance supervenient, constituted objects. But then
45. Assuming the Axiom of Foundation, anyway.

46. [Davidson and Quine have at times suggested a move that would come to this: as-
serting that there are sets1 and functions1 and sets2 and functions2 and that sets1 are (char-
acteristic) functions1 while functions2 are sets2 of ordered pairs (which are . . . ?). Any takers
for that one?!]
Unauthenticated
most if not all ordinary reality will be lost. Perhaps we shall allow
ourselves to continue to use its forms of speech “. . . but only as a
convenient abbreviation.” But in using those forms of speech, in
speaking of snowballs, chains, boxes, trees, hills, or even people,
we shall not believe ourselves to be seriously representing reality
and its contents. “As a convenience”: to whom and for what ends?
“As an abbreviation”: of what?47
Sosa recognizes that “with alternatives so grim, we are encouraged

to return to our relativistic reflections.” But now he raises an additional
worry— and a deep one:
Our conceptual scheme encompasses criteria of existence and

perdurance for the sorts of objects that it recognizes. Shall we
say now that a sort of object O exists (has existed, exists now, or
will exist) relative to a scheme C at t if and only if, at t, C recog-
nizes O by allowing the corresponding criteria? But surely there
are sorts of objects that our present conceptual scheme does not
recognize, such as artifacts yet uninvented and particles yet un-
discovered, to take only two obvious examples . . . What is it for
there to be such objects? Is it just the in-itself satisfaction of
constitutive forms by constitutive matters? That yields the ex-
plosion of real ity.48
Part (but only a part!) of the answer to this worry is that we should
not think of ourselves has having one conceptual scheme. We have a
language within which we can already introduce an indefinite number
of conceptual schemes or optional languages (and as that language de-
velops, there will be the possibility of still more). But we must not think
of all the optional ontologies as if they might be simply pooled: that leads
right back to “the explosion of real ity” (with the consequences noted a
moment ago). Yet reality does force us to recognize that we need, for
example, at least one optional language in which we can describe (for
example) quantum real ity. That there are a number of such optional

Unauthenticated
languages (and ways of “translating” back and forth between them) is

well known.49
To take Sosa’s two very different examples separately: we already have
in ordinary language the broad (and indispensable) category artifact.
So we can now say that there are kinds of artifacts that will probably be in-
vented that we do not now have names for. Saying that does not, of course,
permit us to answer such questions as What is the exact cardinal number
of uninvented artifacts?, even “in principle,” but that is a bad question
anyway if conceptual relativity is right—bad because there will doubt-
less be more than one (optional) way of counting artifacts. The ques-
tion does not become sensible by relativizing it to something called
“our” conceptual system because, as I am emphasizing, we use and need
to use many.
The problem of undiscovered sorts of particles submits to a similar
treatment. But what if Sosa were simply to ask “What of presently inde-
scribable but yet to be discovered sorts of physical entities?” With respect to
this I would say that to say that there are such is to say that reality is
not exhausted by what we can talk about in any precise way, and cer-
tainly not by our relatively precise optional languages of present-day
science. As I said in my Dewey Lectures, we renegotiate—and are forced
to renegotiate— our notion of reality as our language and our life
develop.
Sosa closes this penultimate part of his discussion by repeating his
three options, eliminativism, absolutism (the “explosion of reality”), and
conceptual relativism, and writes, “Right now I cannot decide which of
these is least disastrous. But is there any other option?”
Conclusion
I cannot resist quoting Sosa’s closing paragraph, which I very much ap-
preciate, in full:
49. Such “translations”—the technical term for them is “relative interpretations”—do

not, however, necessarily preserve what Quine has caused philosophers to call the “on-
tology” of a theory, nor are they certifiable as correct by linguists on the basis of what Da-
vidson refers to as “translation practice.” Rather, they are accepted because they preserve
explanations under passage from one version of a theory to another. For a detailed discus-
sion, see Putnam, “Truth and Convention”; and Putnam, “Equivalence,” in Realism and
Reason, 26–45.
Unauthenticated
Of the four Putnamiam arguments for pragmatic realism—the

model-theoretic argument; the argument from the perspectival
character of causation, reference and truth; the argument from ag-
nosticism regarding scientific convergence upon a finished sci-
ence; and the argument for conceptual relativity—this fourth and
last of them seems to me far the most powerful and persuasive. It
raises a threefold issue—the choice between eliminativism, abso-
lutism, and relativism— still wide open on the philosophical
agenda, and a most exciting issue before us today.50
Thank you, Ernie!

Unauthenticated
D5
Richard Boyd on Scientific Realism
I agree with many of the things Richard Boyd says.1

What’s more I’ve agreed with them for a long time, but Boyd misses
this because he reads me in a systematically wrong way. Boyd was so
upset by the things I wrote in my “internal realist” period that he has
been projecting those views onto my publications ever since, and he
shouldn’t do that. But even in that period, as I explained in “From
Quantum Mechanics to Ethics and Back Again,”2 I mistakenly be-
lieved that scientific realism, in Boyd’s sense,3 was compatible with my
“verificationist semantics.”4 I have long believed two principles that I
learned from Boyd, namely that theories in the mature sciences are typi-
1. This paper is a reply to Richard Boyd, “What of Pragmatism with the World Here?,”
in Reading Putnam, ed. Maria Baghramian (Abingdon, United Kingdom: Routledge, 2013),
39–94.
2. This is chapter 2 of my Philosophy in an Age of Science: Physics, Mathematics, and Skepti-
cism, ed. Mario De Caro and David Macarthur (Cambridge, Mass.: Harvard University
Press, 2012), 51–71.
3. I described and supported Boyd’s version of scientific realism in Meaning and the
Moral Sciences (London: Routledge and Kegan Paul, 1978), 20–22.
4. I identified “internal realism” with the thesis that truth is “idealized rational accept-
ability” in Reason, Truth and History (Cambridge: Cambridge University Press, 1981). I used
the term “verificationist semantics” in “Computational Psychology and Interpretation
Theory,” in my Realism and Reason, vol. 3 of Philosophical Papers (Cambridge: Cambridge
University Press, 1983), 139–154.
90
Unauthenticated
Richard Boyd on Scientific Realism 91
cally approximately true, and terms in the mature sciences typically refer, and
I am unrepentant in that respect.5 However, I also think that those
principles are not enough to refute the idea that the “right” semantics
is verificationist semantics (whether of Dummett’s kind, or of the kind
I advocated in Reason, Truth and History), because antirealists may claim
that it is possible to reinterpret the whole language, including all scien-
tific hypotheses expressible in it, and if these two principles represent
a scientific hypothesis, in a very liberal sense of “scientific” (as we both
claimed), then antirealists should claim that they can reinterpret them
too. And that was what I tried to do in my lecture titled “Realism and
Reason”6 and subsequently: accept scientific realism, but reinterpret it
to make it compatible with verifi cationist semantics. I have always been a
scientific realist, in Boyd’s sense, in spite of the misunderstanding, which
is very widespread, that I renounced scientific realism—that’s just
wrong. And I’ve never accepted Nelson Goodman’s idea that we “make”
the world.
Moreover, I even repented, in my Dewey Lectures,7 of saying that
“the mind and the world together make up the mind and the world.”8
And I’ve written a lot about Pragmatism, and in every single paper I
say that I don’t think that the pragmatist theories of truth— any of the
pragmatist theories of truth—Peirce’s, James’s or Dewey’s—were right.
(Fortunately, Dewey never really depended on the pragmatist theory of
truth. He mentions it in one footnote in the Logic,9 but what he is other-
wise concerned with is warranted assertibility, and that means that, for-
tunately, his belief that Peirce got the definition of truth right actually
5. I explain why I am unrepentant in “On Not Writing Off Scientific Realism,” in my

Philosophy in an Age of Science, 91–108.
6. “Realism and Reason,” Proceedings and Addresses of the American Philosophical Association
50 (1977): 483–498; reprinted in Meaning and the Moral Sciences, 123–140. This is where I
fi rst used the term “internal realism.”
7. Hilary Putnam, “Sense, Nonsense, and the Senses: An Inquiry into the Powers of the
Human Mind,” Journal of Philosophy 91 (1994): 445–517; reprinted in The Threefold Cord:
Mind, Body and World (New York: Columbia University Press, 1999), 3–70.
8. Hilary Putnam, Reason, Truth and History (Cambridge: Cambridge University Press,
1981), xi.
9. The footnote in question reads, “The best defi nition of truth from the logical stand-
point which is known to me is that by Peirce: ‘The opinion which is fated to be ultimately
agreed to by all who investigate is what we mean by the truth, and the object represented in
this opinion is the real’.” John Dewey, Logic: The Theory of Inquiry (New York: Henry Holt,
1938), 343n.
Unauthenticated
plays no real role in anything Dewey does. But in any case one can learn
from a philosopher without believing every thing he says, or even be-
lieving every thing he regards as tremendously impor tant. John Mc-
Dowell and I both believe we can learn a great deal from Kant, but
that doesn’t mean we don’t reject certain ideas from the first Critique
that Kant would have regarded as absolutely essential to his whole vi-
sion. It had better be the case that we can learn from dead philosophers,
’cause we’re all gonna be dead!)
Now, with respect to Ethics without Ontology: I did not mean to reject
the idea of metaphysics with a small “m.” Sometimes I describe what I
do as metaphysics, but I know that very often, because of who my
teachers were,10 and perhaps because of the bad influence of a certain
side of Wittgenstein, I do use the word “metaphysics” as a pejorative.
But I am very well aware that I have metaphysical views. And the point
of Ethics without Ontology was not that one shouldn’t have a metaphys-
ical view—a naturalist metaphysical view—of the nature of ethical judg-
ment and ethical practices. In fact, in Ethics without Ontology I tried to
present one, namely, I suggested that ethics rests on a certain set of
human interests. I compared ethics to a table on many legs, and I said
such a table wobbles a lot because the ground isn’t even, but it is very
hard to overturn. And I had something to say about what those inter-
ests are. Like Stanley Cavell,11 I think that ethics isn’t the only way,
but it is far preferable to other ways of settling conflicts. It’s not the only
way of settling conflicts, Cavell says, but the other ways are often inac-
cessible or brutal. And I don’t think that it should be a constraint on
the semantics of ethical terms that anyone who repudiates ethics or re-
jects ethics is irrational. A lot of bad arguments for noncognitivism
assume that cognitivists think that something like that is a constraint
on the semantics of ethical terms. They think that a cognitivist must
make a certain traditional philosophical Platonist view come out
right. But that isn’t what I think.
But let me return to the subject of ontology, and to my dissatisfac-
tion with Quine’s criterion of ontological commitment, and the implicit
10. Here I am thinking primarily of W. V. Quine and Hans Reichenbach.

11. See my “The Fact / Value Dichotomy and Its Critics,” in Philosophy in an Age of Sci-
ence, 283–298, for a discussion of the view of Cavell’s to which I referred.
Unauthenticated
assumption that goes with it, the assumption that “exist” is univocal.12
I wish to draw a distinction here that is foreign to Quine’s way of
thinking: the distinction between saying that an expression has dif-
ferent truth conditions in two or more dif ferent circumstances, and the
difference between saying it has “dif ferent meanings” in the sense that
is connected with translation practice (call it, “the usual linguistic
sense”).13 In the latter sense, it is (roughly) true that “exist” has a con-
stant meaning. I say that because, with Charles Travis, I think that a
word may have the same meaning in dif ferent contexts even though the
truth-evaluable content of what you say by using that word may vary
with the context. (“There is a lot of coffee on the table” may used to
say either that there is coffee to drink, or that coffee has been spilled
on the table; these uses go with dif ferent truth conditions, but it is not
the case that the words have meanings other than those given in a good
dictionary, and those meanings are the same in the two cases). The
truth-evaluable content of what we say using the word “exist” varies a
great deal, and in some contexts it is perfectly right to say that mereo-
logical sums exist, and in other contexts it is right to say there aren’t
really such things as mereological sums, that’s “just one way of talking,”
and we can say what we want to say without it, for example, by talking
about sets. I’m sure Carnap would have said it’s a matter of convention
whether or not we say that there are such things as mereological sums,
and my gut feeling is that that’s got to be right. The view that somehow
there is a deep problem as to whether mereological sums “really exist”
seems absurd. I know that Gideon Rosen thinks that the problem is so
hard that we’ll never know whether there are really mereological sums
or not. That seems to me crazy! And I would assume that Dick Boyd
12. There are deep tensions in Quine’s thinking: for example, between thinking of
“exist” as univocal, and admitting that it admits of both “objectual” and “nonobjectual”
interpretations. Moreover, even when Quine clearly thinks of “exist” as univocal, he equiv-
ocates (or, perhaps, he is torn) between thinking of the univocality of “exist” as something
trivial (what better standard of univocality do we have, he asks in effect, than how some-
thing looks when properly “regimented”?) and thinking of it as involving a substantive
claim, the claim that being committed to the existence of abstract entities (when quantifi-
cation over them cannot be explained away as “nonobjectual”) is doing something he calls
“positing” intangible objects, objects whose existence has to be inferred from the contribu-
tion that positing their existence makes to the success of prediction, particularly in physics.
13. For more on these two senses, see my Ethics without Ontology (Cambridge, Mass.:
Harvard University Press, 2004).
Unauthenticated
probably shares my intuition here. But it’s this idea that there’s just one
correct way of using “exist” and that way determines a unique answer
to the question “are mereological sums included in the furniture of the
universe or not?” that I reject. That’s the sort of “ontology” that I wrote
an obituary for.14
Thus, I am not the internal realist that Boyd thinks I still am, al-
though there are certainly deep issues about which we disagree. Some
of them go beyond the scope of this comment,15 but I will say some-
thing about whether truth is “correspondence.” I agree with Boyd that
true empirical statements correspond to states of affairs that actually
obtain. But I would not say that I have a “correspondence theory of
truth.” In The Threefold Cord: Mind, Body, and World, I defended what I
described as Fregean disquotation (not, please note, Fregean defl ation). I
believe that the disquotational property of “true” is an extremely impor-
tant one, as Frege and Tarski both realized, whatever the differences
between their theories, but neither thought that the disquotation prop-
erty is all there is to say about truth, which is the characteristic thesis
of what is called “deflation.”16 I think one problem with speaking of
truth as defined in terms of a correspondence theory of truth is, first,
that it is misleading. I believe that, in addition to the disquotation prop-
erty, it is a property of the notion of truth that to call a statement of
any kind—not only an empirical statement, but also a mathematical
statement, a statement of logic (e.g., such-and-such a schema is valid), an
ethical statement, and so on—true is to say that it has the sort of cor-
rectness appropriate to the kind of statement it is. When we speak of
“states of affairs” what we normally think of are empirical states of af-
14. Chapter 4 of Ethics without Ontology is titled “ ‘Ontology’: An Obituary.”

15. Examples of our disagreements: I believe that all the familiar moral predicates—
“oughts” as well as “good”— are truth-apt, while Boyd believes that “good” is the name of a
“natu ral kind,” but oughts are noncognitive; I believe that value judgments and descriptive
judgments are entangled, in the sense of presupposing one another, while Boyd rejects this
claim; Boyd styles himself a “materialist,” and I think that his claim (in “What of Pragma-
tism with the World Here?,” 63) that “reference is a real relation holding because of rele-
vant causal interactions between language users and the features of the world to which they
refer” either amounts to nothing more than a supervenience claim, in which case it does
nothing to show that reference is “physicalistically” acceptable, or else simply assumes that
the “relevant causal interactions” are describable at the level of physics, which is what needs
to be shown. [Note added July 4, 2015.]
16. Paul Horwich’s Truth (Oxford: Oxford University Press, 1990) is an excellent account
and vigorous defence of the “deflationist” position.
Unauthenticated
fairs, ways the universe can be.17 And correspondence to such a state
of affairs— one that actually obtains—is the standard of correctness for
empirical statements. Here Boyd and I agree, I believe. When we con-
fine attention to empirical statements, both correspondence and disquo-
tation are features of truth. But even this statement is misleading.
It is misleading because said just like that it could be read as implying
that there is one and the same kind of correspondence at stake no matter
what the empirical statement is, and no matter what the occasion of its
utterance may be. But that is not the case. “This piece of beef weighs
one pound” may “correspond to reality” by the standard of correspon-
dence appropriate to a butcher shop, but be extremely wrong by the
standard of laboratory science. And the kind and degree of correspon-
dence changes again when the statement is “John is very neurotic,”
or . . . The difficulty in giving a picture of our notion of truth (I doubt
we can give anything that deserves the title of a “theory”) is to do jus-
tice simultaneously both to the unity of the notion and to the plurality
of the correctness-conditions that go with it and give it content.
Moreover, when it comes to mathematical statements and logical
statements I am not sure I want to speak of their correctness conditions
as “correspondence,” but I don’t have time to unpack this remark here.18
Suffice it to say that if one must speak of correspondence, then let us
recognize that what sort of thing (in what Sellars once called “the widest
possible sense of thing”) a statement has to correspond to in order to
count as “true” varies from language game to language game. Never-
theless, the concept of truth does exhibit certain constant features; the
disquotation schema describes one.
I would also say that “true” is a logical word like “and,” “or,” or “not.”
That is why it can apply to every kind of sentence, without being in-
compatible with the fact that dif ferent sorts of sentences have dif ferent
sorts of truth-conditions. I can say a tautology is true without being
forced to regard the tautology as “describing” something or “corre-
sponding to a state of affairs,” or inventing a “vacuous state of affairs”
for tautologies to correspond to. Just as one can conjoin a descriptive
sentence and an evaluation with an “and” or an “if-then,” one can say
17. On the indispensability of such talk, see my “From Quantum Mechanics to Ethics
and Back Again,” in Baghramian, Reading Putnam, 19–36.
18. See chapter 3 of Ethics without Ontology for an explanation of why I say this.
Unauthenticated
that a descriptive sentence and an evaluation are both “true” without

implying that they have the same sort of truth-conditions. And, by the
way, my rejection of the fact / value dichotomy doesn’t mean that there’s
no use for a distinction between descriptions and evaluations. A meta-
physical dichotomy and a distinction are not the same thing. But if I
say, for example, “If you’d practiced the piano an hour every day for the
last year, you would have played better,” I am putting an “if___ then
___” between an empirical statement and a judgment of aesthetic merit.
And that’s perfectly fine.
In closing, what I’ve said doesn’t necessarily preclude correspondence-
talk. But I fear that “correspond” tends to suggest that truth depends
on a relation (moreover, one and the same relation) between whole state-
ments and something (reality? parts of reality? what sort of parts?), no
matter what sort of statement we are calling “true.” So I would prefer to
say that descriptive statements (note the restriction!) are connected to the
world via the reference of the names and predicates that they contain,
and reference is a relation to things (and sets of things, and sets of or-
dered pairs of things, etc.) in the world. This is a difficult metaphysical
issue, but I don’t think it bears on whether one is a realist in Boyd’s
sense or not a realist in his sense.
Unauthenticated
D6
Hans Reichenbach
REALIST AND VERIFICATIONIST
I n Meaning and the Moral Sciences,1 I remarked that

“it is not clear that [Reichenbach’s] form of verificationism (the “proba-
bility theory of meaning”) is incompatible with realism.” (However, I
added that I thought it was wrong on other grounds.2) And after re-
hearsing the reasons why the identification of meaningfulness with
“conclusive verifiability in principle” (accepted by the Vienna Circle at
one point)3 is incompatible with realism, I went on to ask,
Why should it be impossible from a realist point of view that (1)

every meaningful sentence have some weight or other in some ob-
1. Hilary Putnam, “Reference and Understanding,” in Meaning and the Moral Sciences
(London: Routledge and Kegan Paul, 1978), 97–119. The remark I quote occurs on p. 112.
2. The grounds in question were that Reichenbach’s “probability theory of meaning,” as
he explained it, ignores two central principles of the “externalist” semantics I advocated in
“The Meaning of ‘Meaning’ ” and subsequently, namely, the linguistic division of labor and
the fact that paradigms external to the speaker often serve to fi x reference. In the case of
the word gold, for example, speakers rely on experts (whose criteria the average speaker
doesn’t know), and the experts rely on ongoing scientific investigation of the metal. I wrote
that “there is nothing in the meaning to enable us individually to assign a ‘weight’ to ‘such
and such is gold’ in many circumstances— contrary to the “probability theory of meaning”
(Putnam, “Reference and Understanding,” 113).
3. In “Testability and Meaning,” Part I, Philosophy of Science 3 (1936): 422, Carnap wrote
that this identification was “exhibited in the earliest publications of our Vienna Circle; it is
still held by the more conservative wing of this Circle.”
99
Unauthenticated
100 REALISM AND VERIFICATIONISM
servable situation, and (2) every difference in meaning be reflected

in some difference in weight in some observable situation? ( These
are, in my formulation, the two principles of Reichenbach’s “prob-
ability theory of meaning.”) At least it should be an open question
for the realist whether this is so or not, and not something that
realism rules out.4
What I want to argue here (and did not argue at that time) is that
this observation is essential to understanding the structure and
content of Reichenbach’s only work in pure epistemology, Experience
and Prediction.5 That is, not only is it the case that one might suppose
that realism and Reichenbach’s form of verificationism are compatible,
but that was, so to speak, the point of Experience and Prediction. What
Reichenbach was trying to do in the whole book was simultaneously to
be a particular sort of realist (a materialist, in particular) and preserve
a weak form of the verifiability theory of meaning—weak, but not too
weak to exclude metaphysics. This is not to say that Reichenbach did
not feel some discomfort; after all, his friend (with whom he co-founded
Erkenntnis) Rudolf Carnap (and the other members of the Vienna Circle
as well)6 had characterized the realism / idealism issue as a pseudo-
issue, and Reichenbach defers to this view to the extent of claiming
that the issue is just a question of a choice of a linguistic framework.
But I wish to claim that the arguments Reichenbach gave only make
sense if we realize that his realism was much more robust than these
disclaimers suggest.
To argue this properly would require a much longer paper than this
one. What I shall do here is point to places in Reichenbach’s argumen-
tation at which his realism is most evident, and also briefly discuss the
claim just alluded to, that the issue is one of choosing a “language.”
4. Putnam, “Reference and Understanding,” 113 (emphasis in original).

5. Hans Reichenbach, Experience and Prediction (Chicago: University of Chicago Press,
1938).
6. Reichenbach, it should be noted was not a Vienna Circler; he led his own group, the
“Berliner Gesellschaft für empirische Philosophie,” in Berlin. Its most famous members,
besides Reichenbach himself, were C. G. Hempel and Olaf Helmer. Erkenntnis, founded in
1930, was the official journal of both the Wiener Kreis and the Berliner Gesellschaft. In
1933, a few days after Hitler took power, Reichenbach fled Eu rope. Experience and Predic-
tion, which appeared in 1938, was written in Istanbul, which provided positions for dozens
of refugee professors from Nazi Germany.
Unauthenticated
Hans Reichenbach 101
I also want to consider another aspect of Reichenbach’s thought, one

not, as far as I know, mentioned in his writings. When I was Reichen-
bach’s student he repeatedly expressed the view that in “Testability and
Meaning”7 and thereafter Carnap had “converged” to a position close
to or perhaps identical with (my memory is not clear here) Reichenbach’s
own position. I want to briefly consider this idea, and argue that while
there is a superficial appearance of convergence, Carnap’s way of
thinking is really very far from what I have described as Reichenbach’s
realism, even after “Testability and Meaning,” and, in fact, to the end
of Carnap’s life.
Background: The Probability Theory of Meaning

I have already mentioned the two principles of “the probability theory
of meaning.” Here they are in Reichenbach’s words): (1) a proposition8
has meaning if it is possible to determine a weight, that is, a degree of
probability, for the proposition;9 and (2) every difference in meaning
is reflected in some difference in weight in some observable situation.10
It is necessary to add that at this time neither Reichenbach nor Carnap
considered any other possible meaning for “probability” than relative
frequency, and that Carnap questioned whether it was possible to as-
sign numerical degrees of probability (in this sense) to hypotheses.11
7. Rudolf Carnap, “Testability and Meaning,” Philosophy of Science 3 (1936): 420–471, and
4 (1937): 2–38.
8. Reichenbach’s term for a sentence. “The words ‘sentence’ and ‘statement’ are also in use.
But this distinction being of little importance and rather vague, we shall make no distinction
between ‘propositions’ and ‘sentences’ and ‘statements’ ” (Experience and Prediction, 21n).
9. Reichenbach neglects to say “on the basis of some possible observation,” but the for-
mulation of the second principle makes it clear that this is what he means.
10. Reichenbach, Experience and Prediction, 54.
11. At least we know that as late as “Testability and Meaning” (1936–1937) Carnap could
write (speaking of Reichenbach’s probability theory of meaning), “It presupposes the thesis
that the degree of confi rmation of a hypothesis can be interpreted as the degree of proba-
bility in the strict sense which this concept has in the calculus of probability, i.e., as the limit of rela-
tive frequency. Reichenbach holds this thesis” (427). Of course, Carnap was later to develop
his theory that “degree of confi rmation” is itself a quantitative notion and a possible inter-
pretation of “probability” dif ferent from “the limit of the relative frequency.” But in “Test-
ability and Meaning” he wrote (also on p. 427), “It seems to me that at present it is not yet
clear whether the concept of degree of confi rmation can be defi ned satisfactorily as a quan-
titative concept, i.e. a magnitude having numerical values” (emphasis added).
Unauthenticated
(The “chief difficulty,” Carnap wrote—citing Popper as the origin of this

criticism—“lies in how we are to determine for a given hypothesis
the series of ‘related’ hypotheses to which the concept of frequency is
to apply.”)12
Reichenbach, as we shall see, viewed scientific hypotheses in a very
dif ferent way from that in which they are most often viewed (although
his view has some affinity to “bootstrapping” views about scientific con-
firmation, such as those proposed by Clark Glymour). His belief that
we can assign a numerical probability to, say, the existence of electrons
was based on the idea that electrons are by no means mere “constructs,”
added to the language of physics for the sake of predictive efficacy. Elec-
trons are things, much as billiard balls are things, and we infer their
existence. Moreover, the very inferences by which we infer their existence
permit us to assign a probability to that existence. It is obvious that
this is (intuitively at least) a realistic conception, in stark contrast to
that of the Vienna Circle.
There are many other symptoms of Reichenbach’s attraction to re-
alism, indeed to an almost naive materialism, in the opening chapter
(titled “Meaning”) of Experience and Prediction. Not only does Reichen-
bach emphasize that linguistic symbols are “physical things” (although
this plays no role whatsoever in the probability theory of meaning), but,
oddly, he adopts what sounds like a naive correspondence account of
reference, writing:
Let us formulate our first answer [to the question, “What is . . .

meaning”] as follows: Meaning is a function which symbols acquire by
putting into a certain correspondence with facts.”13
And again,
If “north” means a certain relation of a line to the North Pole of

the earth, the symbol “north” will occur in connection with the
symbols “London” and “Edinburgh,” as for example, in the sen-
tence, “Edinburgh is north of London,” because the objects
London and Edinburgh are in the relation to the North Pole cor-
12. Carnap, “Testability and Meaning,” 427.

13. Reichenbach, Experience and Prediction, 17 (emphasis in original).
Unauthenticated
responding to the word “north.” So the carbon patch “north” be-

fore your eyes has a meaning because it occurs in relation to other
carbon patches in such a way that there is a correspondence to
physical objects such as towns and the North Pole. Meaning is just
this function of the carbon patch acquired by this connection.14
This pair of sentences in Experience and Prediction, contrary to what

one might expect, is not the beginning of a sustained discussion of how
one gets from putting symbols in correspondence with “facts” or “ob-
jects” to the definition of the concept “truth-value of a sentence.” (There
is no hint of Tarskian semantics in Experience and Prediction.) In fact,
although we have just been told that meaning depends on a “correspon-
dence,” Reichenbach moves at once to other topics (that meaning is a
property of a proposition as a whole, that it is related to the direction
of actions, and then to the discussion of the “truth theory of meaning”
[i.e., the Viennese theory that meaningfulness requires the possibility
of “verification”] and to his own “probability theory of meaning”).
“Correspondence” is never mentioned again. So what are these assertions
doing here?
I submit that they represent declarations of faith. Reichenbach is telling
us that he subscribes to an intuitive realist picture of meaning as corre-
spondence. He has already told us that epistemology deals with “rational
reconstruction”15 (Reichenbach takes the term from Carnap’s “rationale
Nachkonstruction” in the Aufbau). The subsequent discussion, which
does not mention correspondence, is I believe, precisely Reichenbach’s
rational reconstruction of the notion of meaning; but the initial declara-
tions I have quoted show that it is not offered as a repudiation of realism,
but intended to make these realist intuitions more precise.
Reichenbach’s Cubical World Argument

The most interest ing but also the most difficult chapter of Experience
and Prediction is titled “Impressions and the External World.” The dif-
ficulties begin with the title. Although Reichenbach adopts the strategy
of speaking of “impressions” in this chapter, and makes the traditional

Unauthenticated
assumption that impressions are what we “directly observe,” he opens

the very next chapter, “An Inquiry Concerning Impressions,” by at-
tacking that whole way of thinking.
The foregoing chapter was based on the presupposition that im-

pressions are observable facts. We introduced them because we
found that physical observations, even of the most concrete type,
can never be maintained with certainty; so we tried to reduce them
to more elementary facts. It may be doubtful, we said, that there is
a table before me; but I cannot doubt that at least I have the im-
pression of a table. Thus impressions came to be the very arche-
type of observable facts.
This train of thought is of convincing power and there are not
many philosophers who have been able to resist it. As for myself, I
believed it for a long time until I discovered at last some of its weak
points.16 Although there is something correct in these reflections,
it seems to me now that there is something in them which is es-
sentially false.17
Reichenbach immediately goes on to declare, “I cannot admit that

impressions have the character of observable facts. What I observe are
things, not impressions.”18 To be sure, Reichenbach does not doubt
the existence of impressions. “I believe that there are impressions; but I
16. One of those “weak points,” in Reichenbach’s view, is the assumption that no propo-
sition could be known with probability unless some basic propositions were known with
certainty. In a remarkable paper titled “Are Phenomenal Reports Absolutely Certain?”
Philosophical Review 61 (1952): 147–159, delivered at a symposium with C. I. Lewis and
Nelson Goodman, Reichenbach returns to this question late in his life and argues devastat-
ingly against Lewis’s defense of this assumption. Compare my “Reichenbach and the Myth
of the Given,” in my Words and Life, ed. James Conant (Cambridge, Mass.: Harvard Uni-
versity Press, 1994), 115–130.
17. Reichenbach, Experience and Prediction, 163. The only one of the “exceptional” phi-
losophers who have been able to resist the idea that impressions are observable facts cited
by Reichenbach is Richard Avenarius. This is interest ing because we know that it was
reading Avenarius that fi rst inspired James to turn to his own version of direct realism.
D. C. Lamberth, Metaphysics, Experience and Religion in William James’ Thought (Cam-
bridge: Cambridge University Press, 1997) provides a detailed account of the development
of James’s metaphysics.
Unauthenticated
have never sensed them. When I consider this question in an unpreju-

diced manner, I find that I infer the existence of my impressions.”
What is happening in chapter 2 of Experience and Prediction, then, is
not really an account of how we infer the existence of things from im-
pressions; if Reichenbach’s view (in chapter 3) is correct, there is no such
problem. Rather, it is an account of how we infer the existence (and be-
hav ior) of unobserved things from the existence and behav ior of ob-
served things. What Reichenbach assumes for the purposes of chapter 2
is that if “things” only existed when observed by me and had only the
properties that they appear (to me) to possess, then these “things” would
be, to all intents and purposes, just as “phenomenal” as the empiricist
impressions; this justifies, he thinks, (temporarily) identifying the tra-
ditional problem of justifying “the inference from impressions to the
external world” with the problem of justifying the inference from ob-
served things to unobserved things.
He argues vigorously against what he calls the “positivist” view
(Reichenbach never accepted the label as applied to himself) that talk
about unobserved objects is just highly derived talk about one’s own im-
pressions and / or about observed objects. Instead, he likens it to infer-
ring the existence of birds we cannot see, but whose shadows are visible
to us on the ceiling of the cage. To make the inference more compli-
cated, he imagines a setup in which the human race is confined to an
enormous cubical room with translucent walls. There are birds outside
the cube and a mirror that causes the images of the birds to be reflected
onto a sidewall of the cube so that each bird produces two “shadows,”
one on the ceiling and one on the side wall. To make sure that it is phys-
ically impossible for human beings to directly verify the existence of the
birds, Reichenbach also stipulates that there is a “system of repulsive
forces” that makes any near approach to the walls “impossible for men.”19
And he asks, “Will these men discover that there are things outside
their cube different from the shadow-figures?”20
The answer he gives is that
after some time, however, I think there will come a Copernicus.

He will direct telescopes to the walls and discover that the dark

Unauthenticated
spots have the shape of animals, and what is more important still,
that there are corresponding pairs of black dots, consisting of one
dot on the ceiling and one dot on the side wall, which show a very
similar shape.21
Reichenbach goes on to describe how Copernicus will
surprise mankind by the exposition of a very suggestive theory. He

will maintain that the strange correspondence between the two
shades of one pair cannot be a matter of chance but that these two
shades are nothing but effects caused by one individual thing sit-
uated outside the cube within free space. He calls these things
“birds” and he says that there are animals flying outside the cube,
dif ferent from the shadow figures, having an existence of their
own, and that the black spots are nothing but shadows.22
Reichenbach now imagines a positivist, who insists that the “birds”

are just logical constructions; in fact they can be identified with pairs
of black spots.23 And he insists that a physicist would reject this posi-
tivist interpretation.
The physicist, however, would not accept this . . . theory . . . It is

not because he wants to combine with the term “causal connec-
tion” some metaphysical feelings, such as “influence from one thing
to another” or “transubstantiation of the cause into the effect” . . .
Freed from all associated representations, his inference has this
form: Whenever there were corresponding shadow-figures like the
spots on the screen [in the case of similar phenomena observable
within the cubical world], there was, in addition, a third body with
an independent existence; it is therefore highly probable that there
is also such a third body in the case in question. It is this proba-
bility inference which furnishes a dif ferent weight [probability] for
the projective complex and the reducible complex.24
21. Reichenbach, Experience and Prediction, 116–117.

Unauthenticated
In this last quotation, two technical terms appear: “projective com-

plex” and “reducible complex.” A “projective complex” is an item to
whose existence we infer with probability on the basis of similar
cases in the past; a “reducible complex” is something whose existence is
guaranteed by the existence of the evidence because it is a mere logical
construct out of that evidence. Reichenbach is arguing that, on the
(early Vienna Circle) view he is attacking, the existence of the “birds”
[unobserved objects] is certain, because they are nothing more than log-
ical constructs out of the observables (pairs of shadow-figures); on the
view he regards as correct, their existence has a probability less than
one, and is thus not logically equivalent to the existence of the logical
constructs. On Reichenbach’s “probability theory of meaning,” the two
propositions which the “positivists” regarded as equivalent in meaning
are quite distinct assertions.
“Illata” versus “Abstracta”

Two quite opposed reactions are possible to this argument. On the one
hand, a traditional epistemologist might say that the argument simply
begs the question. Reichenbach’s inference presupposes that there are
places outside the cubical world at which unobserved objects can be situated. Is
this not precisely begging the question of “the existence of unobserved
objects”? [This is Reichenbach’s “stand-in” for the question of the ex-
istence of objects other than our sense impressions.]
Reichenbach’s answer to this objection is stated in what he himself
acknowledges to be Carnapian language, which leaves him open to the
suspicion (which he himself shared) that the difference between
Reichenbach and “the positivist” is just the difference between the
Carnap of, say, “Testability and Meaning,” and the earlier positions
of the Vienna Circle. (This is the second possible reaction.) What
Reichenbach says in answer to the first reaction is that the decision to
postulate places we do not observe and to allow inductive speculation
about the behav ior of (possible) objects at such places is a matter of “the
difference of two languages.” I shall discuss this claim in a later sec-
tion, as well as the question of whether, indeed, there is an identity of
view between the Reichenbach of Experience and Prediction and the
Carnap of “Testability and Meaning.” But certainly, as far as the status
of statements about objects we do not observe but could (physically
Unauthenticated
possibly) have observed, what Carnap writes in “Testabililty and

Meaning” is in full agreement with Reichenbach’s position. Discussing
“the following sentence discussed by [C. I.] Lewis and Schlick: ‘If all
minds (or: all living beings)25 should disappear from the universe, the
stars would still go on in their courses’,” Carnap writes:
Both Lewis and Schlick assert that this sentence is not verifiable.
This is true if “verifiable” is interpreted as “completely confirm-
able.” But the sentence is confi rmable and even testable, though
incompletely . . . The sentence in question is meaningful from the
point of view of empiricism, i.e., it has to be admitted in an empir-
icist language, provided generalized sentences are admitted at all
and complete confirmability is not required.26
The “partial confirmation” Carnap speaks of here is just the confirma-

tion (during our lifetimes) of “the laws C of celestial mechanics.”27
In his lectures at the University of California–Los Angeles, Reichen-
bach used to employ the following example: suppose I see a tree-shaped
shadow in front of me on the ground. Whether I turn around and look
at the tree or not, I infer that there is a tree behind me, because I have
confirmed—in situations in which I was in a position to see the place in
question—that whenever there is a tree-shaped shadow on the ground,
then (almost always, if we allow for a small number of cases where the
shadow is cast by something else), in the place at which the base of the
tree-shaped shadow is located, there is also a real tree. The statement
that there is an unobserved tree behind me is simply a deduction (or an
induction, if the generalization is statistical) from a generalization
which is itself inductively confirmed. Concerning this view, I wrote in
Meaning and the Moral Sciences:
What seems right to me about this is that if we had no inductive

logic at all—if we had only pattern recognition and deductive
logic—there would be no basis for ascribing to us any concept of an
“unobserved object.” Our linguistic behavior would fit the account
25. The parenthetical remark is Carnap’s.

26. Carnap, “Testability and Meaning,” Part II, Philosophy of Science 4, (1937): 37–38.
27. Carnap, “Testability and Meaning,” Part II, 37.
Unauthenticated
“tree” means “observed tree”—and, more generally, “object” means

“observed object.” In this sense our inductive logic is part of our
concept of an unobserved object, and hence of an object at all.28
But Reichenbach does not only speak of the inference from observed
objects (birds, trees) to unobserved objects. As already remarked, he re-
gards the inference to unobservable objects (e.g., “electricity, radio
waves, atoms”) as of exactly the same nature.29 Such theoretical enti-
ties, he claims, are illata (inferred entities),30 not abstracta (logical con-
structions). With respect to this claim I want to make two remarks, one
historical and one purely philosophical.
The historical claim is that the agreement between Carnap and
Reichenbach regarding the status of sentences about unobserved objects
decidedly did not extend to agreement about the status of unobservable
objects, such as the “electricity and radio waves” Reichenbach mentions.
The predicates we need to even speak about such objects are only
“partially interpreted,” Carnap insisted to the end of his life; talk about
them is highly derived talk about observables (and, strange to say, sets
of observables, sets of sets of observables, etc.). Although these “theo-
retical constructs,” as Carnap’s followers continued to call them,31 were
no longer claimed to be individually definable in “observational vocab-
ulary,” the theory of contemporary science as a whole was held to be
expressible in a language with no primitive vocabulary except “observa-
tion terms.”32 This is precisely the attitude that Reichenbach rejected.
28. Putnam, Meaning and the Moral Sciences, 112.

29. I use “unobservable” here as the logical empiricists themselves did. In The Threefold
Cord: Mind, Body and World (New York: Columbia University Press, 1999), 43–70, I criticize
the idea that things we observe only with the aid of scientific instruments are “unobserv-
able” in any epistemologically significant sense. But the view I defend there is not compat-
ible with Reichenbach’s view that all scientific knowledge comes from (1) unaided human
perception and (2) what he was prepared to allow as “inductive logic.” (On the problems
with the latter, see “Reichenbach and the Limits of Vindication,” in my Words and Life
(Cambridge, Mass.: Harvard University Press, 1994), 131–148.
30. Reichenbach, in Experience and Prediction, writes, “We use the participle illatum of
the Latin infero, to denote this kind of thing” (212).
31. This term appears repeatedly in the volumes of Minnesota Studies in the Philosophy of
Science edited by Carnap’s ally Herbert Feigl during Carnap’s lifetime.
32. For references, and also for a criticism of this view, see Hilary Putnam, “What The-
ories Are Not,” in Mathematics, Matter and Method, vol. 1 of Philosophical Papers (Cambridge:
Cambridge University Press, 1975), 215–227.
Unauthenticated
The philosophical remark (a critical one) is that Reichenbach’s view

is, however, almost certainly incompatible with his own doctrines re-
specting inductive logic. For Reichenbach, induction is simply the pro-
jection of an observed relative frequency, and induction in this sense is
the only legitimate method of nondemonstrative inference.33 Even if
we accept (as I think we should) some inferences to the existence and
behav ior of physical things too small to see or touch as of essentially
the same character as the inference to the existence of the birds out-
side the cubical world, what Reichenbach ignores is the problem of in-
ferences that are not to the existence of little things at all, but to the
existence of fields (modifications of space itself), and of novel magni-
tudes of all kinds. (It is striking that when he discusses his doctrine of
“illata” he writes that “modern physics has shown that electrons, posi-
trons, protons, neutrons, and photons are the basic elements out of
which all things are built up in the form of reducible complexes.”34 Here
he entirely ignores the predicates involved, and thinks only of the ob-
jects!) In The Philosophy of Space and Time, Reichenbach allows himself
to pretend that there are arbitrarily small “clocks” and “measur ing
rod[s]”; however useful and legitimate such a device may be in explaining
relativity theory, we cannot seriously suppose that an inference to the
existence of such a magnitude as electricity and to the differential
equations that it obeys are simply “inductive” inferences (in Reichen-
bach’s limited sense) to the behav ior of arbitrarily small measur ing
instruments!
The problem will appear strange to most philosophers since we are
used to thinking, with Mill, that something called the “hypothetico-
deductive method” is an essential part of the apparatus of “induction,”
and that it is by confirming the predictions of theories about electricity
that we confirm the existence and behav ior of electricity. But Reichen-
bach rejected the hypothetico-deductive method, or, rather, he had a
Bayesian attitude toward it: the method is legitimate only when sanc-
tioned by Bayes’s Theorem, which requires a knowledge of the prior
probabilities of the alternative hypotheses. But the difficulty is in seeing
how Reichenbach’s inductive method can assign any probability at all
to hypotheses about such “illata.”
33. See part V of Experience and Prediction, “Probability and Induction.”

34. Reichenbach, Experience and Prediction, 215 (emphasis added).
Unauthenticated
The reason that Reichenbach held such a restrictive view of induc-

tion has to do, strangely, with a rationalist strain in his thinking: he be-
lieved that he had found a deductive vindication of induction, and that
this limited sort of induction was the only method for which such a de-
ductive vindication could be given.35
Just “the Difference between Two Languages”?

In §16 of Experience and Prediction, “An Egocentric Language,”
Reichenbach formulates the difference between positivism and realism
as a matter of choice between an “egocentric language” (in which
things exist only when observed to exist by the subject, and have all
and only the properties they appear—to that same subject—to have),
and “usual language” which allows inferences to unobserved objects.
(A similar distinction occurs in “Testability and Meaning,” although
not under these labels.) At the beginning of the next section, §17,
“Positivism and Realism as a Problem of Language,” he writes, “The
difference of the positivistic and the realistic conception of the world
has taken a dif ferent turn; this difference has been formulated as the
difference of two languages.” And he immediately adds, “This form
of consideration, which has been applied particularly by Carnap, seems to
be a means appropriate to the problem in question, and we shall make
use of it for illustration of our results.”36
What a reader of Carnap would expect at this point is some such
argument as the following: the more liberal language (“usual language”)
allows the formulation of statements from which we can deduce or
induce many useful predictions which are not permitted by the re-
strictions of the older positivism ( because the statements cannot be
formulated in “egocentric language”). In short, the statements science
makes about, for example, objects too small to see, or events before there
were sentient beings, are part of a system that leads to valuable predic-
tions about what we observe here and now, predictions each scientist
could in principle test for herself. But Reichenbach offers a remarkably
dif ferent argument:
35. See part V of Experience and Prediction, “Probability and Induction.”

36. Reichenbach, Experience and Prediction, 145 (emphasis added).
Unauthenticated
The insufficiency of a positivist language in which talk of events

after my death is construed as a device for predicting my experi-
ences while I am alive is revealed as soon as we try to use it for the
rational reconstruction of the thought processes underlying actions
concerned with events after our death, such as expressed in the ex-
ample of [purchasing] life insurance policies.37
I contend that this argument makes a deep point, and one quite un-
like anything to be found in Carnap’s writing. What Reichenbach is
telling us is that if I view my whole language as just a device for pre-
dicting what experiences I myself will have—if even statements about
my family, and about what will happen to them after I die, are no more
than gears in a prediction machine, a machine whose whole purpose is
to predict what I will experience here and now—then that view will vi-
olate the deepest intuitions we have about what we are doing when we
utter sentences about others and about events after (and before) our own
lives. One might add (although Reichenbach unfortunately did not) that
even if I deny that such statements are translatable into an “egocentric
language,” if the only account I have of what it is to understand a “re-
alist” language is that it consists in being able to use it to predict one’s
own sensory stimulations, the view remains just as unsatisfactory. As I
have myself put this further moral:
Moral: to preserve our commonsense realist convictions it is not enough

to preserve some set of “realist” sentences; the interpretation you give those
sentences, or, more broadly, your account of what understanding them
consists in, is also impor tant! 38
I do not claim that Reichenbach drew this further moral. Indeed, I

have argued elsewhere that he failed to see that in consistency his ar-
37. Reichenbach, Experience and Prediction, 150. In “Logical Positivism and Intention-
ality,” in Words and Life, 90–93, I argue that Reichenbach fails to realize that this defense of
realist language is, in fact, incompatible with the defense he offers of the claim that all that
is at stake is a choice of a language. It assumes, in par tic u lar, that my understanding of the
language doesn’t consist merely in my ability to assign weights to sentences on the basis of
my own experiences, but recognizes this involves recognizing that some account of refer-
ence is needed.
38. Hilary Putnam, “Richard Rorty on Real ity and Justification,” in Rorty and His
Critics, ed. Robert Brandom (Oxford: Blackwell, 2000), 81–87 (emphasis in original).
Unauthenticated
gument required him to give a nonpositivist account of understanding.39

But the intuition behind Reichenbach’s little argument about “the life
insurance policies” is, I think, quite clear.
Conclusion
In closing, it is appropriate to ask, when all is said and done, just what did
Reichenbach understand by his contrast of “the positivistic and realistic
conceptions of the world”? I take it, that at least this much is safe to say:
(1) A realistic conception of the world did not, in Reichenbach’s eyes,
presuppose a theory of truth. His realism was, in the main, a rejection of a
picture that he saw as inadequate to our scientific lives and, in the re-
markable paragraph I just quoted, to our humanity as well, not a proposal
for a metaphysical foundation of some kind. (2) Reichenbach did not,
however, think through the question of whether what he retained from
positivism was fully adequate to his “realistic conception of the world.”
39. Putnam, “Richard Rorty on Real ity and Justification.”
Unauthenticated
D7
Between Scylla and Charybdis
DOES DUMMETT HAVE A WAY THROUGH?
L i k e Q u i n e’s e ss ays, starting with “On What There

Is” and “Two Dogmas of Empiricism,” and like Donald Davidson’s es-
says, starting with “Meaning and Truth,” and like Kripke’s lectures
published as Naming and Necessity, Michael Dummett’s essays and books
blazed entirely new paths in philosophy of language, construed not nar-
rowly, but as absolutely central to the whole of metaphysics and episte-
mology. In all three of these cases, the reverberations still continue to
be felt, and they will be going on being felt for a long time to come. In
my own case, it was hearing Dummett’s William James Lectures de-
livered in 1976 at my own university that exerted a profound influence.1
Even if my own orbit has carried me some distance away from the point
of closest approach to Michael Dummett’s philosophical planet, I con-
tinue to feel his gravitational attraction, and I feel tremendously en-
riched by our interactions, personal as well as philosophical.
The Problem Dummett Sees with Realistic Semantics

The difficulty Dummett sees with realistic semantics is summed up in
his famous “manifestation argument,” to the effect that if we did have
1. Michael Dummett, The Logical Basis of Metaphysics, The William James Lectures 1976
(Cambridge, Mass.: Harvard University Press, 1991).
114
Unauthenticated
Between Scylla and Charybdis 115
a conception of “realist” truth, truth as fully independent of all possi-

bility of verification, then there would be no way in which we could
“manifest” our conception in our behav ior, and hence no way in which
we could teach our realist conception of truth to others, determine
which others really have this concept of truth, and so on.2 But one all-
impor tant remark: although Dummett has concentrated for many
years on arguing that there are (apparently insuperable) difficulties
with realism, there is a sense in which the “antirealism” or “justifica-
tionism” that Dummett regularly treats as the alternative to realism is
not precisely advocated by Dummett. Instead, Dummett (admirably, in
my view) sees himself as exploring a philosophical issue, or exploring a
philosophical dialectic (with, to be sure, a predilection to one side of
the debate). As he himself puts it in the preface to Truth and the Past,3
“I do not think anyone should interpret everything that a philosopher
writes as if it was just one chapter in a book he is writing throughout
his life. On the contrary, for me every article and essay is a separate
attempt to arrive at the truth, to be judged on its own.”
If the “antirealist” position Dummett proposes as the alternative to
realist semantics remains in this way “work in progress,” the manifes-
tation argument against realism has remained a constant in the many
years Dummett has been making his “separate attempts to arrive at
the truth” in this area.4 A pithy statement of the manifestation argu-
ment occurs, in fact, in an essay titled “Truth: Deniers and Defenders”
that he added to the lectures that make up the body of Truth and the
Past. Dummett writes,
2. Michael Dummett, “What Is a Theory of Meaning, I,” collected in his Seas of Lan-
guage (Oxford: Clarendon Press, 1993), 97–138.
3. Michael Dummett, Truth and the Past (New York: Columbia University Press, 2004).
4. However, I don’t wish to give the impression that it was the manifestation argument
that was responsible for what I described as a close approach of our philosophical orbits
around the time of Dummett’s William James Lectures (and for a number of years subse-
quently). If anything it was because of a dif ferent argument, my model theoretic argument
against metaphysical realism, that I was prepared to accept the claim that “we have no
means” for acquiring the realist notion of truth. The model theoretic argument is set out in
a lecture titled “Realism and Reason” to the American Philosophical Association reprinted
as the last chapter of my Meaning and the Moral Sciences (London: Routledge and Kegan
Paul, 1978), and at more length in my Reason, Truth and History (Cambridge: Cambridge
University Press, 1981). I explain what I think is wrong with it in The Threefold Cord: Mind,
Body and World (New York: Columbia University Press, 1999).
Unauthenticated
What prompts [justificationist] theories of truth is a thought

similar to that of Rorty in concluding that justification is the
goal of inquiry: namely that when we acquire the practice of
using language, what we learn is what is taken to justify asser-
tions of dif ferent types. We learn what is accepted as entitling us
to make those assertions; we learn also whether what justifies us
in doing so is conclusive or whether it is defeasible, that is, capable
of being overthrown by subsequent counterevidence. We do not
learn what it is for those assertions to be true independently of
any means we have for establishing their truth. How could we? If
we are not in a position either to assert or to deny a given propo-
sition, we cannot be shown what nevertheless makes it true or
false. So, according to a theory of this kind, to grasp the meaning
of a statement is to know what would justify asserting it or de-
nying it.5
Redefining “Realism”
What I hope to do in this essay is to make clear how it was that I came
to see what I regard as insuperable difficulties with antirealism. In this
way, I hope to open yet another page in the dialogue that Michael Dum-
mett and I have been having for forty years.
The term “realism” occurs quite often in my essays prior to 1975.6
In those essays, it means one thing and one thing only: the rejection
of logical positivism, operationalism, and related positions. As I ex-
plained in one of those essays, “A Philosopher Looks at Quantum
Mechanics”:
All attempts to literally “translate” statements about, say, electrical

charge into statements about so-called observables (meter readings)
have been dismal failures, and from Berkeley on, all a priori argu-
5. Michael Dummett, “Truth: Deniers and Defenders,” in Truth and the Past, 97–116;
quotation from p. 114.
6. I am thinking in par tic u lar of the essays I collected in the fi rst two volumes of my
Philosophical Papers, which were published in 1975 by Cambridge University Press. (A third
volume, Realism and Reason, published in 1983, represents my subsequent “internal realist”
position— a position I now regard as a false start to dealing with the very real problem of
the normativity of language use.)
Unauthenticated
ments designed to show that all statements about unobservables

must ultimately reduce to statements about observables have con-
tained gaping holes and outrageously false assumptions. It is quite
true that we “verify” statements about unobservable things by
making suitable observations, but I maintain that without imposing
a wholly untenable theory of meaning, one cannot even begin to
go from this fact to the wildly erroneous conclusion that talk about
unobservable things and theoretical magnitudes means the same as
talk about observations and observables.7
Although I pointed out in that essay that Carnap had given up the
attempt to reduce statements about unobservables to “observation lan-
guage,” I criticized him in “Explanation and Reference,” for an “ide-
alist” tendency manifested in the fact that according to the theory of
the meaning (or “partial interpretation”) of theoretical terms in science
he defended in his late writings, every change in the total scientific
theory amounted to a change in the reference of every one of those
terms, so that theoretical terms were not treated as names of, say, unob-
servable things and forces concerning which scientists change their
minds, but as merely parts of a machine for predicting “observations,”
parts which have no meaning in themselves apart from their role in the
particular theory.8
The term “idealism” in those essays was virtually synonymous with
“phenomenalism.” Prior to my reading Dummett’s William James Lec-
tures, the only “idealism” I knew was Berkeley’s, and the only “antire-
alism” I knew was antirealism about unobservables (and, in the case of
phenomenalism, about “middle sized dry goods,” which were treated as
unobservables by phenomenalists). Thus it was an eye-opener that “re-
alism” and “antirealism” could be understood as positions about the
nature of truth itself, and not simply as positions about the reducibility
or nonreducibility of “theoretical terms” to “observation terms” or of
“thing-language” to “sense-datum language.”
7. Hilary Putnam, “A Philosopher Looks at Quantum Mechanics,” in Mathematics,

Matter and Method, vol. 1 of Philosophical Papers (Cambridge: Cambridge University Press,
1975), 130–158, quotation from p. 131.
8. Hilary Putnam, “Explanation and Reference,” in Mind, Language and Reality, vol. 2 of
Philosophical Papers (Cambridge: Cambridge University Press, 1975), 196–214.
Unauthenticated
What I then called “idealism” is better called reductionism, Dummett

taught me.9 Reductionism, with respect to a class of statements, is the
philosophical theory that statements in that class are “made true” by
facts described by statements in what is claimed to be some epistemolog-
ically or metaphysically more “basic” class. For example, the phenome-
nalist view that statements about tables and chairs and other ordinary
“material objects” are made true by facts describable in a sense-datum
language is a reductionist view of the kind I called “idealist.”
If a view is reductionist with respect to assertions of one kind, only to
insist on a “correspondence” notion of truth for statements in the re-
ducing class, then that view is, according to Dummett, metaphysically
realist at base. A truly nonrealist view is nonrealist all the way down.
I say that this redefinition of realism (and antirealism) was an “eye-
opener” because it seemed to open a way out of the difficulties I had
been having in thinking about the model theoretic argument against
metaphysical realism—an argument that had occurred to me before
Dummett’s William James Lectures, but that I could not see my way
clear to either accepting or rejecting at that time, which is why I did
not present it publicly until 1976.10
Not surprisingly, Dummett’s redefinition of “realism” and “antire-
alism” was contested: most vehemently, perhaps, by Michael Devitt.11
According to Devitt, the realism issue is simply, “Is there a mind-
independent reality or not?” (thump) and there that question has
nothing to do with semantics. This short way with the issue reminds
one of Lenin’s (disastrously incompetent) polemical book against Ma-
chian positivism.12 Lenin simply claimed that positivists, since they
took human sensations as the class of truth-makers for all propositions
(I am using present-day terminology, not Lenin’s, of course), could not
9. The views I am referring to were those in the William James Lectures (see note 1,
above). They were set out more briefly in “What Is a Theory of Meaning? (I),” and “What Is
a Theory of Meaning? (II),” respectively published in Mind and Language, ed. S. Guttenplan
(Oxford: University Press, 1975), 97–138, and in Truth and Meanings: Essays in Semantics, ed.
Gareth Evans and John McDowell (Oxford: Oxford University Press, 1976), 67–137.
10. See note 4 for details.
11. Michael Devitt, Realism and Truth (Oxford: Blackwell, 1984); see Drew Khlentzos,
Naturalistic Realism and the Antirealist Challenge (Cambridge, Mass.: MIT Press, 2004) for a
convincing criticism of Devitt’s response to Dummett.
12. V. I. Lenin, Materialism and Empirio- Criticism (Moscow: Foreign Languages Pub-
lishing House, 1952).
Unauthenticated
accept the statement that the solar system existed before there were
human beings.13 This argument simply assumes—what positivists of
course deny14 —that the positivists cannot interpret “the solar system
existed before there were human beings” in their rationally recon-
structed “language of science.”
But what of the word “independent” in “the behav ior of the stars is
independent of human sensations and thoughts and beliefs”? (This is
what Devitt portrayed antirealists as denying.)
Well, there are many kinds of independence. Presumably causal in-
dependence is what Devitt was talking about, since logical independence
is a property of statements (or, perhaps, of events under a description), and
whether statements are or are not logically independent is certainly a
question about their semantics, which Devitt claimed to be irrelevant
to the realism issue. But then Devitt’s argument once again simply as-
sumes—what antirealists of course deny—that the antirealist cannot
interpret the sentence “the behavior of the stars is independent of human
sensations and thoughts and beliefs” in a “justificationist” way, inter-
pret it so that it is “true” (in the antirealist sense). Devitt cannot, after
all, say, “But that’s not what the sentence means!” without engaging in
a discussion of—guess what?— semantics.
I remain convinced that Dummett has made a truly lasting contri-
bution to our appreciation of the depth of the realism/antirealism issue,
just as Berkeley did at an earlier time. Devitt’s dismissive attitude is as
unphilosophical as Samuel Johnson’s stone-kicking.
The Antirealist Tries to Steer between Scylla and Charybdis

But can one really develop a defensible antirealist account of the seman-
tics of our language? When I reread what I wrote in Reason, Truth and
History (especially in chapters 3 and 5), I see my former self as trying
13. Indeed, Devitt seems to have read Lenin closely—at least his fi rst published attack
on my “internal realism” was titled “Realism and the Renegade Putnam,” a clear play on the
title of Lenin’s famous article “Marxism and the Renegade Kautsky.”
14. For example, Carnap discusses this sort of statement (his example is “If all minds (or
all living beings) should disappear from the universe, the stars would still go on in their
courses”), without mentioning Lenin, however, in “Testability and Meaning,” Part II, Phi-
losophy of Science 4 (1937): 37–38, and concludes that it is both cognitively meaningful and
well confi rmed.
Unauthenticated
to steer between the Scylla of solipsism and the Charybdis of metaphys-

ical realism, and when I read Dummett’s Truth and the Past, I find that
I have the same feeling. But this takes some explanation, I know.
In §16 of Experience and Prediction,15 Reichenbach considered and re-
jected a form of verificationism that he described as “the choice of an
egocentric language” (a language in which things can meaningfully be
said to exist only when observed to exist by the subject, and can mean-
ingfully be said to have only the properties they appear—to that same
subject—to have). According to this form of verificationism (Reichen-
bach is clearly thinking of a position held by members of the Vienna
Circle at a certain point), all cognitively meaningful assertions that
seem to be about times or places the speaker never observes are to be
reinterpreted as asserting that if the speaker were to do so-and-so, then
she would experience such-and-such. As we recalled in Chapter 6
[“Hans Reichenbach: Realist and Verificationist”], Reichenbach argued
that such a reinterpretation is unable to express the rationale for certain
“actions concerned with events after our death, such as . . . [purchasing]
life insurance.”16
One obviously doesn’t buy life insurance so that one will oneself have
certain experiences if one does so (or if one has the experience of doing
so)! Let us see now how a similar difficulty might confront a “justifica-
tionist” philosopher.
Suppose that this philosopher maintains that what it is for a state-
ment (any statement) to be true is for the statement to be justified by
experiences she has or will have in her lifetime. If what makes the state-
ment true is also supposed to be what the statement “means”—what it
asserts to be the case— she is a solipsist. If not, an unacceptable gap
seems to open between what a statement says and what makes it true. In
either case, Reichenbach’s observation has force: she cannot intelligibly
rationalize buying a life insurance policy.
A verificationist who denies that there is such a thing as logically con-
clusive verification of an empirical proposition may nonetheless face a
similar problem. If my understanding of my own sentences is alleged
to consist in my ability to assign them a degree of verification on the
15. Hans Reichenbach, Experience and Prediction (Chicago: University of Chicago Press,
1938).
Unauthenticated
basis of my own experiences, then, in particular, my understanding of

the prediction that my family will be better off after my death if I buy
life insurance now consists in the knowledge that certain experiences I
have or could have in my lifetime justify asserting that sentence to this,
that, or the other degree. Again, if the justificationist says, “Oh, but
I didn’t say claim that the sentence is about those experiences,” she
creates a gap between the semantics of the sentence— since what
makes her a justificationist philosopher, in Dummett’s sense, is precisely
that her program is to give the semantics of the sentence in terms of
justification—and what she claims sentences mean. But how can seman-
tics not be about what sentences mean?
Avoiding Scylla: Step One

The first step in avoiding the Scylla of solipsism is one that Dummett and
I both took at different times. That step was simply to insist that the jus-
tificationism or verificationism that we advocated was social and not indi-
vidualistic in nature.17 Dummett puts it very clearly in Truth and the Past:
What is the concept of truth appropriate to a justificationist theory

of meaning? Plainly it must turn on the notion of our being justi-
fied in asserting a statement. It is evident at the outset that the word
“our” must be taken in a collective, not a distributive, sense.18
Avoiding Scylla: A Second Step

In chapter 3 of my Reason, Truth and History, this first step was com-
bined with a further one: replacing talk of verification with idealized
verification:
Truth is an idealization of rational acceptability. We speak as if

there were such things as epistemically ideal conditions, and we
call a statement “true” if it would be justified under such conditions.
17. Dummett uses the term “justificationism” in Truth and the Past; I spoke of “verifica-
tionist semantics” in my fi rst internal realist writing, “Realism and Reason,” my American
Philosophical Association (Eastern Division) Presidential Address of 1976, collected in my
Meaning and the Moral Sciences (London: Routledge and Kegan Paul, 1978), 123–138.
18. Dummett, Truth and the Past, 41.
Unauthenticated
“Epistemically ideal conditions,” of course, are like “frictionless

planes”; we cannot really attain epistemically ideal conditions, or
even be absolutely certain that we have come sufficiently close to
them. But frictionless planes cannot really be attained either, and
yet talk of “frictionless planes” has “cash value” because we can
approximate them to a very high degree of approximation.19
Perhaps it will seem that explaining truth in terms of justification

under ideal conditions is explaining a clear notion in terms of a vague
one. But “true” is not so clear when we move away from such stock ex-
amples as “Snow is white.”
The simile of frictionless planes aside, the two key ideas of the ideal-
ization theory of truth are (1) that truth is independent of justification
here and now, but not independent of all justification. To claim a state-
ment is true is to claim it could be justified. (2) Truth is expected to be
stable or “convergent”; if both a statement and its negation could be “jus-
tified,” even if conditions were as ideal as one could hope to make them,
there is no sense to thinking of the statement as having a truth-value.20
The Roar of Charybdis Is Heard in the Near Distance

The foregoing account of truth made use of the counterfactual condi-
tional as well as of the notion of epistemically ideal (or “close to ideal”)
conditions. S is true, according to this view, just in case the following
counterfactual is true:
S would be justified if epistemic conditions were good enough.21
But how is this counterfactual to be understood? As I put the diffi-

culty in my Dewey Lectures,22 explaining my reasons for giving up
the whole approach:
19. Putnam, Reason, Truth and History, 55.

20. Putnam, Reason, Truth and History, 55–56.
21. I employed a similar counterfactual in Representation and Reality (Cambridge, Mass.:
MIT Press 1988), 115.
22. “The Dewey Lectures 1994: Sense, Nonsense, and the Senses: An Inquiry into the
Powers of the Human Mind” (special issue of Journal of Philosophy 91, no. 9, 1994); these are
collected as part I of The Threefold Cord (pp. 3–70).
Unauthenticated
Unlike Dummett’s “global antirealist,” I did not suppose that em-

pirical propositions could be unalterably verified or falsified.23 And
I was bothered from the start by the excessively “idealist” thrust
of Dummett’s position, as represented, for example, by Dummett’s
fl irtation with strong antirealism with respect to the past, and I
avoided that strong antirealism by identifying a speaker’s grasp of
the meaning of a statement not with an ability to tell if the state-
ment is true now, or to tell whether it is true under circumstances
the speaker can actually bring about . . . but with the speaker’s pos-
session of abilities which would enable [any] sufficiently rational
speaker to decide if the statement is true in sufficiently good epi-
stemic circumstances.
To the objection that this is still an “idealist” position, I re-
plied that it certainly is not, on the ground that while the degree
of confirmation speakers actually assign to a sentence may simply
be a function of their sensory experiences . . . the notion of suffi-
ciently good epistemic circumstances [was] a “world involving” no-
tion. That is why the totality of actual human sense experiences
does not, on this position, determine the totality of truths, even
in the long run.
On my alternative picture (as opposed to Dummett’s), the world
was allowed to determine whether I actually am in a sufficiently
good epistemic situation or whether I only seem to myself to be
in one—thus retaining an impor tant idea from commonsense
realism—[but] the conception of an epistemic situation was, at
bottom, just the traditional epistemological one. My picture still
retained the basic premise of an interface between the knower and
every thing “outside.” But while the need for a “third way” besides
early modern realism and Dummettian idealism is something I feel
as strongly as ever, such a third way must . . . undercut the idea that
23. Although Dummett is not unaware that the verification of an empirical statement is
typically corrigible, as a rule, he tends to prescind from this fact. This tendency may spring
from his expressed desire to carry Brouwer’s intuitionist logic, a logic designed by Brouwer
in connection with an antirealist philosophy of mathematics, over to empirical language.
The simplest possible way to make such a carryover is to extend the notion of “proof,”
which is the basic notion in the intuitionist semantics for mathematical language, to a bivalent
predicate ‘verified’ ” applicable to mathematical and non-mathematical language alike, and
this is what Dummett does. In this quotation, only the fi rst italic (“unalterably”) was in the
original text.
Unauthenticated
there is an “antinomy,” and not simply paste together elements of

early modern realism and elements of the idealist picture.24
As mentioned earlier, the “idealization theory of truth” was presented

in chapter 3 of Reason, Truth and History. In chapter 5, the problem of
the understanding of counterfactuals like “S would be justified if epi-
stemic conditions were good enough” was addressed, however, by
adopting a verificationist account of how we understand counterfac-
tuals. I said simply that a “nonrealist” or “internal realist” regards con-
ditional statements as statements that we understand (like all other
statements) in large part by grasping their justification conditions. This
does not mean that the “internal” realist abandons the distinction be-
tween truth and justification, but that truth (idealized justification) is
something we grasp as we grasp any other concept, via a (largely) im-
plicit understanding of the factors that make it rationally acceptable to
say that something is true” (122–123). The dilemma I faced (but was
not aware that I faced at that time) was this: let us suppose, as seems
reasonable, that whatever makes it rational to believe that S makes it
rational to believe that S would be justified were conditions good
enough. If my understanding of the counterfactual “S would be justi-
fied if conditions were good enough” is exhausted by my capacity to tell
to what degree it is justified to assert it, and that is always the same as
the degree to which it is justified to assert S itself, why did I bother to
mention the counterfactual at all? Why did I not just say that my
understanding of S is just my capacity to tell what confirms S to what
degree, full stop? It seems that the whole appeal to “idealized” verifica-
tion, to counterfactual verification, was an unnecessary shuffle. But then
the jaws of the Scylla of solipsism close on me! On the other hand, if I
repudiate the justificationist account of our understanding of counter-
factuals, the Charybdis of the metaphysical realism I was trying to avoid
sweeps me into its whirlpool. It was the impossibility, as I now think it
to be, of steering an antirealist course between the Scylla of solipsism
and the Charybdis of metaphysical realism that led me to develop and
defend what I believe to be an unmetaphysical version of realism in The
Threefold Cord: Mind, Body and World. It is time now for us to see if
24. Putnam, “The Dewey Lectures,” 14–15.
Unauthenticated
Michael Dummett has a found a way where I failed, to steer between

Scylla and Charybdis.
Dummett’s Way
As was already said, the first step in avoiding the Scylla of solipsism was
simply to insist that the justificationism that we both advocated was
social and not individualistic in nature, or as Dummett put it, that the
word “our” in “our being justified in asserting a statement” must be
taken “in a collective, not a distributive, sense.” Whether this is some-
thing that Dummett can say without doing violence to his own justifica-
tionist principles, and what it means when understood in a justificationist
way, is what we must now ask.
The question is impor tant because of its connection with another
question, the question to which, indeed, Truth and the Past is devoted:
how to avoid antirealism about the past. I referred earlier to Dummett’s
“flirtation with antirealism about the past”; he himself says that this is
a topic that troubled him for years, and that this new book is an attempt
to see if a justificationist can repudiate “the view that statements about
the past, if true at all, must be true in virtue of the traces past events
have left in the present.” He still not does find that view incoherent,
but he does say it is “repugnant.”25 The way in which Dummett pro-
poses to reject antirealism about the past involves simply counting past
observers as fully on a par with present observers. The details of this
proposal are complex, and it would take a much longer essay than this
to examine them in detail. But the essential ideas are (1) that the se-
mantically crucial verification of a statement about the past is not the
present “indirect” verification, via a memory, or, in the case of the dis-
tant past, via a historical trace, that the statement is true, but the veri-
fication by a witness at the time (“Dying does not deprive anyone of
the status either of an observer or of an informant”);26 and (2) that while
that “direct” verification may be transmitted to us via a trace, it counts
as a verification whether it is transmitted or not (“For all the messages
25. Dummett, Truth and the Past, ix. I believe that Yuval Dolev has successfully argued
that, contrary to Dummett, antirealism about the past is incoherent. See Dolev’s “Dum-
mett’s Antirealism and Time,” European Journal of Philosophy 8 (2000): 253–276.
Unauthenticated
that have been lost, it remains that statements about the past must count
as having been directly established, and therefore as true, if someone
observed them to be true at the, or an, appropriate time”).27
The problem, however, is that while language is indeed social, com-
petence is individual. Earlier I quoted Dummett as asking (rhetorically)
“How could we learn what it is for assertions to be true independently
of any means we have for establishing their truth?” but one can also
ask “How could” an individual language learner learn to understand
Dummett’s “we.” Dummett, indeed, tells us that:
A child who had learned only when he was right to come out with
simple assertoric utterances, such as “Doggie” when a dog was in
sight, would serve as an extension of adults’ range of observation,
but could not yet be credited with saying that anything was so; he
can be credited with that only when he has learned to treat the
utterances of others as extending his range of observations. It is
intrinsic to the use of language that we accept the testimony of
others: to believe what we are told is the default response. Language
binds us into society.28
I agree that it does, of course. But Dummett moves too rapidly from
the child’s acquisition of the practice of accepting the testimony of
others—an idea the justificationist seems clearly entitled to—to the
child’s “treating the utterances of others as extending his range of ob-
servations,” where this “treating” must amount to something quite dif-
ferent from a mere disposition to accept sentences that the child hears
if it is to account for the child’s grasp of the thought that a statement
about the past may be true even though no testimony is available (to the
child). (Dummett also speaks of the child’s forming a mental “grid” that
shows the relations of other places and times to one another and to the
child’s present location.)
The Alternative to Antirealism

In The Threefold Cord, discussing the antirealism about the past that
Dummett once defended (or experimented with), I wrote:

Unauthenticated
If we accept it that understanding the sentence “Lizzie Borden

killed her parents with an axe” is not simply a matter of being able
to recognize a verification in our own experience— accept it, that
is, that we are able to conceive of how things that we cannot verify
were—then it will not appear as “magical” or “mysterious” that we
can understand the claim that that sentence is true. What makes it
true, if it is, is simply that Lizzie Borden killed her parents with
an axe. The recognition transcendence of truth comes, in this case,
to no more than the “recognition transcendence” of some killings.
And did we ever think that all killers can be recognized as such? Or
that the belief that there are certain determinate individuals who
are or were killers and who cannot be detected as such by us is a
belief in magical powers of the mind?29
I believe that Michael Dummett would agree with this. But I think
that he would still say that we cannot rest with the commonsense thought
that being able to think about past events that we cannot verify (being
able to locate them in our mental “grid”) is an ability that we acquire as
we acquire language and all the skills that language brings with it. He
believes that we need to show that that ability can be accounted for in a
“justificationist” way. (Perhaps he thinks this because he thinks that anti-
realism about the past is only “repugnant” and not actually “incoherent”).
But, first, I don’t see that he has accounted for that ability in a justifica-
tionist way, for the reasons I have just briefly laid out, and, second, I do
think that antirealism about the past is incoherent.30
29. Putnam, Threefold Cord, xxx.
30. See Yuval Dolev’s “Dummett’s Antirealism,” cited previously. Dummett himself
pointed out that if one holds the antirealist view that statements about the past, if true at all,
“must be true in virtue of the traces past events have left in the present,” then it must be that
the meaning of every empirical statement about the past changes every time the reference of
“the present” changes—that is to say, must change at every moment. But, as Dolev points
out, any attempt to say what a statement about the past meant a moment ago must then like-
wise be semantically unstable—indeed, the change in question will be indescribable.
Unauthenticated
D8
When “Evidence Transcendence”
Is Not Malign
A s a lways, I learned from Wright’s “Truth as Sort of

Epistemic: Putnam’s Peregrinations,”1 and I am sure that the points
on which we still disagree will provide fruitful material for future dis-
cussions. I can only express my heartfelt gratitude for the careful reading
and the enormous amount of thought that obviously went into the
writing of this essay. In many ways, Wright is the ideal reader I had in
mind when I wrote the papers and books he discusses.
In what follows, I have found it easiest to discuss Wright’s points in
roughly the reverse of the order they occur in his paper. I begin with a
quote from the final paragraph of his paper:
It is essentially the interface picture that we need to make sense of

the idea that “There are no intelligent extraterrestrials” might be
true beyond all evidence available in principle, beyond the verdict
of even the very best possible empirical theory.2
To explain why I don’t agree with this, let me go back a few paragraphs
in Wright’s paper to some remarks in which Wright lays out in a clear
1. Crispin Wright, “Truth as Sort of Epistemic: Putnam’s Peregrinations,” Journal of

Philosophy 97 (2000): 335–364.
2. Wright, “Truth as Sort of Epistemic,” 364.
128
Unauthenticated
When “Evidence Transcendence” Is Not Malign 129
and very general way the position for which he is arguing. Wright makes
the point that there are benign forms of evidence-transcendence:
This benign form of evidence-transcendence is conceived as going

with—in the broadest sense— contingencies of epistemic opportunity:
in all cases where we have a conception of this kind of how the
truth-value of a particular statement could be unverifiable, a de-
veloped specific account of that conception will consist in detailing
limitations of opportunity, or spatio-temporal situation, or percep-
tual or intellectual capacity, which stop us from getting at the
relevant facts but to which we, or others, might easily not have been
subject—or at least, to which we can readily conceive that an in-
telligible form of investigating intelligence need not be subject.3
In the paragraphs that follow, Wright contrasts this benign form with
the “malign” kind of evidence-transcendence posited by interface con-
ceptions, writing:
On this type of view, evidence-transcendence is a product not

of contingencies of opportunity, or of contingent limitations
of our powers, but of metaphysical shortcomings and divides:
consciousness—unless it is the consciousness of God, of which we
have no satisfactory conception—is necessarily and essentially
insular.
The essence of metaphysical realism, we might say, is thus
interface- realism. And the evidence-transcendence of truth which
metaphysical realism brings in train is a malignant kind, the kind
that goes with an interface-conception of mind’s interaction with
the world. About these ideas there is nothing “natural” or “com-
monsensical.” Rather, it is supposed to take philosophical sophis-
tication to appreciate them.4
I agree with Wright that the sorts of evidence-transcendence that we

can explain as resulting from “contingencies of epistemic opportunity”
in the broadest sense should be admitted to exist by any philosophy that

Unauthenticated
aims to do justice to the natural realism of the common man (as Dum-
mett’s antirealism about the past evidently did not).5 But unlike
Wright, I believe that “There are no intelligent extraterrestrials” might
conceivably be true; that this too should be admitted by any philosophy
that aims to do justice to commonsense realism; and that the evidence-
transcendence of the truth of this proposition should be classed as “be-
nign” and not as “malignant.”
To see why, let us compare it with a “Fitch-example” (Wright calls it
F) whose truth is also evidence-transcendent, but in a way Wright con-
siders “benign”:
Q and no and one will ever rationally believe that Q.
Here let Q be any empirical proposition that it might happen that no

ever rationally believes, say, There is a gold mountain one mile high.6
According to Wright, we can give a truth-condition for F, in the
fashion of a Carnapian “reduction sentence” (Wright calls it a “Provi-
sional Biconditional”). The condition is:
(1) If F was to be appraised under sufficiently good circumstances,

F would be true just in case it was believed to be true.7
But there is a catch! As Wright notes, there are only two sorts of
circumstances in which it is possible to “appraise” the truth-value of
F: these are (1) circumstances in which it is possible to know that Q is
false, and (2) circumstances in which it is possible to know that someone
rationally believes that Q is true, and in both sorts of circumstances F
is false (because either its first or its second conjunct is false). In sum,
what (1) captures is the fact that F is falsifiable. But it isn’t verifiable. If we
say that F might be true, we must presuppose our hearer’s possession
of an idea of what constitutes F’s being true which (1) does not cap-
ture. This would become crystal clear if the “sufficiently good circum-
stances” in (1) were spelled out “constructively,” in Wright’s sense, for
5. My internal realism was never committed to this position of Dummett’s.

6. I used this very example, not knowing of Fitch’s earlier argument, in Hilary Putnam,
“Logical Positivism and the Philosophy of Mind,” in Mind, Language and Reality, vol. 2 of
Unauthenticated
then it would be obvious that the “circumstances” in question are all

circumstances in which F is false. A “reduction sentence” in Carnap’s
sense leaves the extension of the concept it “reduces” undetermined
when the test condition mentioned in the antecedent is not fulfilled—
something Carnap and Hempel both emphasized.8 Wright’s Provi-
sional Biconditional is meant as a reduction sentence for the predicate
“true,” one which partially specifies its extension as applied to F, and it
leaves the extension of that predicate undetermined in every case in which
the “sufficiently good circumstances” do not obtain—including all the
circumstances in which F is true.
Thus, one of the following two remarks by Wright must be wrong:
(a) A moderate internalist conception of truth [in Wright’s sense]

is friendly to . . . classes of potentially recognition-transcendent
truths, typified by . . . Fitch examples.9
(b) In all cases where we have a conception of this kind of how
the truth-value of a particular statement could be unverifiable, a
developed specific account of that conception will consist in de-
tailing limitations of opportunity, or spatio-temporal situation, or
perceptual or intellectual capacity, which stop us getting at the rel-
evant facts but to which we, or others, might easily not have been
subject—or at least, to which we can readily conceive that an in-
telligible form of investigating intelligence need not be subject.10
To spell this out: if Wright’s “moderate internalist conception of

truth” does indeed allow that some Fitch propositions can be true
(I take this to be what is meant by the reference to the “friendliness” of
the conception to this class of potentially recognition-transcendent
truths), it had better not assert (b)! For it is not the case that our
conception of how a Fitch proposition might be true (as opposed to
our conception of how it might be false) enables us to say how “we, or
others” might have been able to recognize its truth.
8. C. G. Hempel, “The Empiricist Criterion of Meaning,” Revue Internationale de Philos-

ophie 4 (1950); collected in A. J. Ayer, Logical Positivism (New York: Free Press, 1959). See,
for example, Logical Positivism, 120.
Unauthenticated
To make matters even stranger, the situation is exactly the same if

instead of F, which Wright sees as exemplifying “benign” recognition-
transcendence of truth, we take N: “There are no intelligent extrater-
restrials,” whose (possible) recognition-transcendent truth Wright sees
as “malign.” For that proposition is falsifiable-though-not-verifiable just
as F is. Its truth-value, like F’s, can be “appraised under sufficiently good
circumstances”— circumstances in which it is false! For if N is false,
then there are intelligent extraterrestrials, and we may easily imagine
there being observers who know this. We can even formulate a “Pro-
visional Biconditional” for N— simply replace F by N in (1)!
This brings us to the question: given that he doesn’t think that “Fitch
examples” involve any sort of an interface conception, why does Wright
say that “It is essentially the interface picture that we need to make sense
of the idea that ‘There are no intelligent extraterrestrials’ might be true
beyond all evidence available in principle, beyond the verdict of even
the very best possible empirical theory?” Earlier in the paper, he writes:
If moderate internalism can accommodate the example [our N ] as

a truth-apt claim, then the view has to be that under such infor-
mationally fortunate circumstances [knowledge of the conditions
for the evolution of intelligent life, and “enough understanding of
the distribution and diversity of matter in the cosmos to allow the
assignment of a well-grounded probability to the supposition that
such conditions are replicated elsewhere besides on this earth”],
thinkers would take the correct view of the question.11
It is the idea that thinkers would not take the correct view of the ques-
tion under such circumstances that Wright analogizes to belief in a
metaphysical “interface.” But this is obviously a confusion.
To see the confusion, let us assume, as is reasonable on present
knowledge, that the “very best possible empirical theory” that Wright
speaks of tells us that the cosmos, or at least the part of space-time
that might contain intelligent life, is fi nite, and that there is no one
place within that fi nite region from which one can survey (or even
receive causal signals from) the entire region. Let us suppose, with
Wright, that the theory, call it P, provides “enough understanding of
Unauthenticated
the distribution and diversity of matter in the cosmos to allow the

assignment of a well-grounded probability to the supposition that
such conditions are replicated elsewhere besides on this earth,” and
let us even suppose that the probability in question is quite high, say
.999 or .9999. Now, it is a conceptual truth (what Wright calls an “a
priori” truth)12 that what is highly probable does not always happen. Thus
it is internal to physical theory itself (to P)— and not the product of some
metaphysical conception of “the kind that goes with an interface-
conception of mind’s interaction with the world”—that N may be true.
If “moderate internal realism” requires us to believe that all those
propositions that our best theory regards as highly probable (but no
more than that) are in fact true, then “moderate internal realism”
must certainly be rejected!
What then is the source of the recognition-transcendence of true
Fitch propositions and true negative existentials similar to N? The an-
swer I offer both in the Dewey Lectures and in a paper I believe Wright
may have missed13 is that while it is indeed the case that our ability to
perceptually verify them under favorable circumstances is part of what
is comprised in our understanding of such atomic descriptive proposi-
tions as “Here is a chair” (and thus I do not need to be persuaded by
Wright that we should not “recoil” so far in the direction of metaphys-
ical realism as to think of truth as recognition-transcendent in the most
ordinary cases), the use of logical connectives and quantifiers does fre-
quently lead to recognition-transcendence for reasons that are clear
from the logical structure itself. (Note that even the Fitch examples in-
volve negative existentials!). Nor is the principle of bivalence what is
involved here. Just as the logical structure of a Fitch example makes it
clear that what is supposed in such an example— and I see no reason to
regard it as inconceivable, or “metaphysical,” or “confused”—is some-
thing that, if it happens, we cannot verify, so the logical structure of N
makes it clear that what is supposed in this example— say, that some-
thing improbable (if it is improbable) simply happens—is something
12. My qualms about “a priori” stem from the long-standing association of the notion
with the idea of unrevisability, not with any rejection of the idea of conceptual analysis.
Compare with Hilary Putnam, “Pragmatism,” Proceedings of the Aristotelian Society 95,
part 3 (1995): 291–306, especially pp. 299–302.
13. See Putnam, “Pragmatism,” 291–306.
Unauthenticated
that if it happens we cannot verify.14 Both sorts of example (and, of

course, there are others) spell disaster for any elucidation of truth in
terms of rational acceptability, even by means of reduction sentences
(Provisional Biconditionals).
Much of Wright’s paper is concerned with interpretation of and im-
provement on the theory presented in my Reason, Truth and History
(which I no longer accept for the reasons given in my Dewey Lectures).15
I will only venture three brief remarks on what I believe I meant in the
insufficiently clear passage that Wright wrestles with so insightfully.
(1) What I had in mind was the following: for a statement P to be

true (on the “internalist account”), there must be a set of
conditions C that are “good” (say, if we indulge in the fiction
that there are numerical degrees of goodness, to some degree g)
for appraising P, such that P is assertible under conditions C,
and such that it is the case that if conditions C′ were still better
(say, the degree of goodness for appraising P were g′ instead of g)
then P would still be assertible under C′. That we can treat
conditions that are close to (but not the same as) some ideal
conditions I as if they were the ideal conditions (as if the incline
were frictionless) if improving the approximation wouldn’t
affect the result is a principle that governs all use of idealiza-
tions in solving practical problems (hence the example of
“frictionless planes”). I believe, in fact, that my internalist
notion of truth coincided with the notion that Wright calls
“superassertibility” in his Truth and Objectivity (although that is
not how he interprets me!), if we prescind from possible differ-
ences over the notion of “assertibility.”
14. More precisely, the conjunction P&N&S, where S is the statement that an observer
cannot verify whether there is intelligent life in a region of space-time that that observer is
unable to receive causal signals from, P is the “possible empirical theory” that tells us that
(1) causal signals do not travel faster than light, (2) it is physically possible (and highly prob-
able) that there are intelligent extraterrestrials, but (3) it is also physically possible that
there aren’t, and (4) there are large regions of space-time that any par tic u lar physically
possible observer is unable to receive causal signals from, is a statement that it is logically
impossible to verify if true, and yet it is a statement that P tells us could be true (and even
assigns a probability to).
15. These are collected in Hilary Putnam, The Threefold Cord: Mind, Body and World
(New York: Columbia University Press, 1999), 3–70.
Unauthenticated
(2) I did not and do not, however, think that one can survey in
advance what one might reasonably come to understand by
“sufficiently good epistemic conditions,” even for a particular
topic. Certainly there is no fi xed set of “standards” associated
with language that could enable one to do this. (This is the
possible difference over the notion of “assertibility” to which I
just referred.)
(3) Wright is quite correct in supposing that I did not consider
the problems posed by the Conditional Fallacy in any of my
formulations. I agree that they are serious—indeed, I do not
think that Provisional Biconditionals can meet them either
(precisely because, like all reduction sentences, they leave the
extension of the predicate that they “elucidate”— Carnap’s
term was “reduce”—undetermined with respect to cases in
which the test condition is not fulfilled, and this is precisely
what we don’t want in the case of recognition-transcendent
truths). I at least want to say that, for example, N can be true in
such cases, not that it is indeterminate. Even in the case of
truths that are not recognition transcendent but simply such
that (contingently) the test condition is unfulfilled, it is not
enough to specify via a provisional biconditional what the truth
value would have been if the recognition condition had been
fulfilled, for this leads right back to the problems of the Condi-
tional Fallacy. My conclusion is that the problem Wright poses
for my former view is also an (unsolved) problem for Wright’s
own “moderate internalism.”
Unauthenticated
Unauthenticated
D9
Sensation and Apperception
T h e r e f l e c t ions about the topic of perception and its

relations to conception that follow are the product of more than seven-
teen years of thinking about John McDowell’s great book Mind and
World. If I fi nd myself forced to disagree with him at certain points
(Amicus Plato, sed magis amica veritas), that does not alter the fact that
his book was a pathbreaking one in the highest sense of the term. For
that reason I shall begin the present chapter by reviewing some its ar-
guments, and explain my own view as a correction of his.
John McDowell’s Reasons for Thinking

That Experience Is Conceptualized
I once wrote,
What McDowell means by saying that our conceptual powers

are “drawn on” in experience, albeit “passively,” is not anything
mysterious, nor is this to be construed as psychological specula-
tion of some kind; it is articulated by the work that this idea has
to do: to show how experience involves “openness to how things
anyway are.”
If we put aside temporarily McDowell’s difficult (and fasci-
nating) discussion of experiences with “inner accusatives” (e.g.,
139
Unauthenticated
140 NAIVE REALISM, SENSATION, AND APPERCEPTION
experiences of pain, or of after-images) and confine our attention to

experiences of how it is “out there,” what McDowell is saying is that
such experiences, when they are experiences in what McDowell
calls “the demanding sense” (when they function in the justifica-
tion of belief), are intrinsically about the outer world, and the pos-
sibility of having them depends on the possession of the relevant
world-involving concepts.
They are not inner signs with a magical connection to the outer
world, but takings in of how it is (in the best case), or how it seems
to be (in more problematic cases), with the outer world.1
I confess that I no longer understand how I could have thought that

McDowell’s claim that the possibility of having perceptual experiences
“depends on the possession of the relevant . . . concepts” can be any-
thing but “mysterious.”
However, McDowell means exactly what he says, as is shown by other
things he says. For example, in Mind and World, he writes, “No subject
could be understood as having experiences of color except against a back-
ground of understanding that makes it possible for judgments endorsing
such experiences to fit into her view of the world.”2
And he makes clear that this is supposed to apply to inner experi-
ences such as the experience of “seeing red” produced by a blow on the
head, or even the judgment that I have a pain. According to this view,
experiences must be conceptually articulated.3 What I want to explore
1. “McDowell’s Mind and McDowell’s World,” in Reading McDowell on Mind and World,
ed. N. H. Smith (London: Routledge, 2002), 174–190; quotation from pp. 176–177.
2. John McDowell, Mind and World (Cambridge, Mass.: Harvard University Press,
1994), 30 (emphasis added).
3. In “Avoiding the Myth of the Given,” in his Having the World in View (Cambridge,
Mass.: Harvard University Press, 2009), 256–272, McDowell gives up the claim that the
experiences of language-using human beings (which he refers to by the Kantian term “in-
tuitions” in this chapter) presuppose that the subject already possesses the concepts (including
demonstrative concepts, like “this shade”) needed to describe their content, but he takes
the ability to form such concepts to be “in play” when one has those experiences, and he
explicitly describes such abilities as “discursive” abilities. Thus, McDowell still takes the
learning of a language to be a prerequisite for having experiences in the sense in which
rational beings— beings that can understand and respond to requests for justifications for
their judgments— have experiences. Hilla Jacobson and I discuss this revised version of
the thesis that experiences are “conceptualized” in “Against Perceptual Conceptualism,”
International Journal of Philosophical Studies (forthcoming).
Unauthenticated
Sensation and Apperception 141
is how McDowell arrived at this metaphysical position! What follows

is a very brief account.
McDowell Believes the Space of Reasons Cannot Be

Reduced to Facts about the Causation of Our Beliefs
McDowell is not satisfied with a merely reliabilist account of justifica-
tion and other epistemological notions. And I agree with him. Such an
account, if offered in a reductionist spirit, presupposes many notions
that the reductionist naturalist is not entitled to, for instance, inten-
tional notions such as reference and truth, not to mention the use of
counterfactuals, often accompanied by talk of “possible worlds” a là
David Lewis, as well as being open to a number of counterexamples.4
McDowell must be understood as seeking to produce an account of
experience compatible with the idea that perceptual experience justi-
fies beliefs about the layout of the world around one and doesn’t merely
causally “trigger” true beliefs a high percentage of the time. This is the
first thing one needs to know about Mind and World.
This idea, that perceptual experiences justify accepting and rejecting
beliefs about the world and do not merely trigger noises and subvocal-
izations is what McDowell calls “minimal empiricism.” McDowell
clearly identifies this claim with the claim that “impressions” can do
this, and I will come back to this identification later.
James Conant has distinguished between two varieties of skepti-
cism, which he calls “Cartesian skepticism” and “Kantian skepticism.”5
Cartesian skepticism, in Conant’s sense, is skepticism about the possi-
bility of knowledge of things and events “outside” the mind. Kantian
skepticism is a puzzle about the very possibility that one’s thoughts,
whether seemingly about an external world or even about one’s own
sense impressions, can have content at all. Cartesian Skepticism as-
sumes our thoughts about the world are genuine thoughts, that is, that
4. See my criticism of Fodor’s attempted causal theory of reference in Hilary Putnam,

Renewing Philosophy (Cambridge, Mass.: Harvard University Press, 1992), 35–59.
5. James Conant, “Varieties of Skepticism,” in Wittgenstein and Skepticism, ed. Denis
Mc Manus (Abingdon, United Kingdom: Routledge Press, 2004): 97–136. Conant claims
that in my work The Threefold Cord: Mind, Body and World (New York: Columbia University
Press, 1999), I failed to see that Kantian skepticism was McDowell’s target, and not Carte-
sian skepticism. I think that McDowell, in fact, attacks both targets in that book.
Unauthenticated
they are true or false, and only worries about whether we can ever
really know that any of them are true. Kantian skepticism threatens
to undercut even Cartesian skepticism. This poses a problem for em-
piricism, in that it is not clear how merely saying that our judgments
come from experience explains how any of them have content, as op-
posed to being merely what Rorty called “marks and noises.” “Come
from” needs to mean something more than “are caused by,” McDowell
tells us, if experience is to be a “tribunal” before which our judgments
are to stand.6 The second thing we need to know to understand Mc-
Dowell is that he is concerned with Kantian, not Humean, skepticism
in Mind and World.
McDowell’s Way into the Problem

McDowell’s way into the problem is via an interest ing interpretation
of Wittgenstein’s Private Language Argument, which he reads in the
light of Wittgenstein’s remarks on the supposition that words acquire
meaning by “ostensive defi nition.”
Wittgenstein’s point can be illustrated thus: Russell (as of 1912)
thought that we acquire our basic concepts by “acquaintance by intro-
spection” (in effect, by a sort of inner pointing to our private impres-
sions).7 But concepts are general, as Russell well knew. So to get the
concept red (as applied to sense data),8 because I obviously can’t direct
my attention to all my red sense data, including the future ones, I have
to “abstract” the appropriate quality. In other words, I need a private
ostensive definition. But all ostensive definition (e.g., holding up a glass
and saying “glass” to teach someone the concept) presupposes that the
pupil possess at least implicit knowledge of the relevant sortal concept
(e.g., “implement” as opposed to “material”). So ostensive defi nition
can’t be the way we acquire all our concepts. That is, pointing to “bare
presences” can’t explain how language and minds “hook on to the
6. “We can coherently credit experiences with rational relations to judgment and belief,
but only if we take it that spontaneity is already implicated in receptivity; that is, only if we
take it that experiences have conceptual content.” McDowell, Mind and World, 162.
7. Bertrand Russell, The Problems of Philosophy (Oxford: Oxford University Press,
1912), 47.
8. How we can possibly acquire concepts that apply to anything besides sense data was a
puzzle for classical empiricism.
Unauthenticated
world.” (In addition to Wittgenstein, McDowell attributes this insight

to Sellars and Davidson.)
A way out that McDowell doesn’t discuss would be to say that some
concepts are innate. (Think of Quine’s “similarity spaces.”) But innate
similarity spaces in the behaviorist sense are just innate patterns of
response, dispositions to be caused to respond (say, with the noise “red”),
and if that’s all we have, the corresponding judgments are merely con-
ditioned responses to make certain noises, and we lose the fact that our
concepts are concepts. And if this behaviorist story isn’t what the nativist
has in mind, then it isn’t clear what an “innate concept” is. (Chomsky,
today’s best-known nativist, rejects the question of how words refer as
too unscientific to discuss.)9
I agree with McDowell, Davidson, and Sellars that appealing to “bare
presences” can’t provide an answer to the question as to how concepts
and experiences are connected, or, in McDowell’s terms, how experi-
ences can rationally constrain beliefs. That confrontation with bare
presences can do this is what Sellars meant by “The Myth of the Given,”
and what Davidson meant by the idea of “content” in the phrase “du-
alism of scheme and content,” according to McDowell. (Davidson
should have said that what he rejected was the dualism of conceptual
scheme and Given, on McDowell’s interpretations.) These are fasci-
nating and plausible interpretations, needless to say.
Davidson
Davidson—who evidently despaired of fi nding a rational linkage be-
tween concepts and experiences (or impressions—we need to keep an
eye on McDowell’s identification of experiences with impressions!) held
(reviving an idea of Neurath’s) that justification begins with beliefs
(which, for Neurath, were just sentences) and not with experiences. But
how sentences about experiences can be justified by other sentences is
just as problematic! (The coherence of the whole system does the work?
How do we know it isn’t “spinning in the void?”) And if observation
sentences (Neurath called them “protocols”) aren’t justified by any-
thing, how can they in turn justify anything? Davidson responds to
9. Noam Chomsky, “Language from an Internalist Perspective,” in New Horizons in the

Study of Language and Mind (Cambridge: Cambridge University Press, 2000), 134–163.
Unauthenticated
the problem that bothers McDowell by giving up minimal empiricism.

But that can’t be right, McDowell feels. Nor can the answer be to just
recoil back to the Given. We saw that that didn’t work. So what way
out is left?
McDowell’s Way Out

McDowell’s way out is to say that perceptual experiences aren’t just the
products of our sense organs. According to McDowell our conceptual
capacities are “in play” in experience. Impressions “already possess con-
ceptual content.” Since both beliefs and “impressions” are conceptually
articulated, there is supposedly no obstacle to the idea that impressions
can rationally justify, and not merely “trigger,” beliefs. Thus, Mc-
Dowell claims that we can “dismount from the seesaw” that threatens
to keep us oscillating between the Myth of the Given and an equally
untenable “coherentism.”
It seems to me that McDowell, far from giving a philosophical worry
peace by “exorcizing the question,” has worked himself into an unbe-
lievable metaphysical position! What I find unbelievable is not the claim
that some of our experiences are conceptualized (in some sense of “con-
ceptualized”), nor the claim that conceptualized experiences are epis-
temologically fundamental, but the claim that all experiences, indeed
all sensations, involve and presuppose our conceptual powers. Surely,
the reader is going to want to know, how exactly is it supposed to be
the case that my conceptual capacities are “in operation?”10 A prima
facie difficulty is that any given experience has an enormous number
of different aspects.11 Moreover, as McDowell recognizes, “demonstra-
tive expressions” have to be used to describe many of those aspects;12
it is not the case that one always has the appropriate concept before
having an experience that falls under that concept. McDowell (as of
10. “Experiences in general are states or occurrences in which conceptual capacities are
passively drawn into operation” (McDowell, Mind and World, 30). “Conceptual capacities
must be in play in experience if we are to avoid the Myth” (McDowell, Having the World in
View, 262).
11. The Stanford Encyclopedia of Philosophy refers to this as the problem of “the richness of
experience.” Susanna Siegel, “The Contents of Perception,” Stanford Encyclopedia of Philos-
ophy, March 18, 2005, www.illc.uva.nl /~seop/archives/spr2009/entries/perception-contents/.
12. McDowell, Mind and World, 58.
Unauthenticated
Mind and World) was clearly wedded to the idea that one could not have
a par tic u lar sensation (“impression”) if one did not have the concepts
under which that sensation falls. But how could one’s conceptual abili-
ties be “in play” when one has a sensation (say, a par tic u lar color-
sensation) if one didn’t previously have that concept?
McDowell’s view is clearly in flux here. In Mind and World, the an-
swer to this question was that one forms the demonstrative concept and
one has the sensation at the same time. In Having the World in View, how-
ever, the view McDowell defends is that to have a sensation it suffices
that one could have formed the appropriate demonstrative concept; it is
conceded that in many cases one does not, in fact, form it. Yet the claim
that “conceptual capacities are in play” in all experiences is not given
up, and that includes the claim that one could not have any par ticular sen-
sation if one did not have the ability to form a concept under which it falls.
But why not? Let us grant that it is a conceptual truth that one could
not have the experience of recognizing that one is having a sensation that
falls under a particular concept if one did not have that concept (or form
it on that occasion, in the case of a demonstrative concept). But it does
not seem to be a conceptual truth that one cannot have the sensation of
seeing a particular color (or of seeing red flashes because one has been
hit on the head—a McDowell example),13 or feel a particular sort of
pain unless one has, or has the ability to form, a demonstrative con-
cept of precisely that color (or precisely that sort of “red flash”), or pre-
cisely that sort of pain.
Why should our human conceptual abilities (which McDowell iden-
tifies with “discursive abilities”—that is, the abilities that enable us to
use language and speak about “reasons”) be required to feel pains or
see red flashes when one is struck on the head? And why are brain sci-
ence and cognitive psychology supposed to be irrelevant to these ques-
tions? McDowell’s supposed “exorcism” of the questions only leaves this
reader more puzzled.
A natural place to look for an account of the extent to which human
discursive abilities affect sensation would be a sophisticated and nonre-
ductionist cognitive science. But for McDowell, there is one thing that
is the natural-scientifi c kind of intelligibility, and that he identifies with
“bald naturalism,” that is, the sort of naturalism that seeks a reductionist
13. See McDowell, Mind and World, 30.
Unauthenticated
account of normative epistemic notions, and there is a dif ferent thing that
is philosophical understanding, and never the twain shall meet.14 But
what I have elsewhere called a “liberal functionalism,”15 a nonreductive
investigation of the functional roles of biological and psychological
processes, has no reason to eschew normative notions (e.g., the notion
of the “normal function” of an organ or a capacity itself involves nor-
mative judgment) or intentional notions. Moreover, the whole idea of a
sharp line between philosophical speculation and empirical hypothesis
is one I find extremely uncongenial.
In any case, I propose that if we are willing to think about the phil-
osophical problem of how experience can have rational bearing on
belief and the cognitive psychological question as to what a reasonable
picture of the extent to which our conceptual powers are be “in play”
in experience together, we are more likely to make progress than if we
assume from the beginning that philosophy is philosophy and natural
scientific intelligibility is natural scientific intelligibility, and the two
have nothing to do with each other.
A Brief Sketch of an Approach

The approach I propose depends on drawing a distinction that Wil-
liam James employed, but that McDowell’s claim that all experiences
and / or impressions (he sometimes uses the terms interchangeably) are
conceptualized in effect denies, the distinction between “sensations”
and apperceptions. James did not, indeed, invent this distinction; at least
since Kant’s Critique of Pure Reason, the distinction has been a standard
one in the psychological literature, but Kant also thought that a sensa-
tion that was not accompanied by apperception would be “nothing to
us,”16 and this is perhaps the basis for McDowell’s denial that there can
14. The possibility stressed by Tyler Burge (discussed in Chapter 1 of this book) that
cognitive science might use and be entitled to use a concept like representation goes
missing in McDowell’s picture, of course. [Note added July 7, 2015.]
15. See “Corresponding with Real ity,” chapter 3 of Philosophy in the Age of Science: Physics,
Mathematics, and Skepticism, ed. Mario De Caro and David Macarthur (Cambridge,
Mass.: Harvard University Press, 2012) and “How to Be a Sophisticated ‘Naïve Realist’,”
chapter 37 in the same volume.
16. “The I think must be capable of accompanying all my presentations; other wise some-
thing would be presented to me which could not be thought at all, which means no less
than: the presentation would be either impossible, or at least nothing to me . . . Conse-
Unauthenticated
be such a thing. James did not deny this, but he recognized that the
line between sensation and apperception is sometimes fuzzy; in fact, he
asserted that in the case of a
presented and recognized material object . . . Sensations and ap-

perceptive ideas fuse here so intimately that you can no more tell
where one begins and the other ends, than you can tell, in those
cunning circular panoramas that have lately been exhibited, where
the real foreground and the painted canvas join together.17
In The Threefold Cord, I attempted to use James’ notion of “fusion” as

a way of taking the metaphysical strangeness out of McDowell’s notion
that perceptual experiences are conceptualized (while also restricting
that claim to the case of a “presented and recognized material object
[or event]”).18 But this was a misinterpretation of McDowell on my
part; in “Avoiding the Myth of the Given,” McDowell makes it clear
that while he is willing to abandon the claim that sensations are con-
ceptually articulated he is not willing to give up the claim that in all
(mature human) experience, our discursive abilities are “in play.” In
some sense, thus, McDowell still thinks that sensations (or “impres-
sions”) must be conceptualized.
However, pace McDowell, and also pace Kant, there is no paradox in-
volved in the idea that we are aware of some unconceptualized sensa-
tions. In short-term memory, at least, we can access sensations we had a
moment ago that we did not conceptualize at the time. Our present
memory of those sensations is conceptualized, to be sure, but they weren’t.
Indeed, I have one or two vivid visual memories of sensations I had when
I was just learning to talk and had virtually no concepts available for
most of the things in the remembered “scene.”
That “apperception”—recognition of what one is perceiving—
involves concepts was already stressed by Kant. This is also something
that the phenomenological school, beginning with Husserl, likewise
quently every manifold of perception has a necessary relation to the I think, in the same
subject in which the manifold is found.” Immanuel Kant, Critique of Pure Reason, §15 and
§16, B 131.
17. William James, Essays in Radical Empiricism, ed. R. B. Perry (Cambridge, Mass.: Har-
vard University Press, 1976), 16.
18. Putnam, Threefold Cord, 158.
Unauthenticated
emphasized: I see a building as something that has a back, Husserl

pointed out, even when I don’t see the back.19 (This is something also
emphasized today by the phi losopher and cognitive scientist Alva
Noë.)20 Such perception is fallible, to be sure, but so is the perception
that something is red or circular.
But pointing this out isn’t going to convince McDowell that he should
confine the claim that our experiences are conceptualized to appercep-
tions, because he believes that giving up “minimal empiricism” would
make skepticism unanswerable, and minimal empiricism, as he under-
stands it, requires that impressions belong to the “tribunal” before which
all our judgments have to stand.
On “Fusion”
The phenomenon that interested James is that on some occasions the
phenomenal character of a sensation is actually changed by conceptual-
ization. For example, the phenomenal character of the sensation of
seeing a word in an alphabet one doesn’t know certainly changes when
one learns to read texts in the relevant language. When literate English
speakers see the word STOP on a traffic sign, our awareness of the letters
and our awareness that they “say” stop are, as James said, “fused.” But
there are some concepts that need to be distinguished here.
First, there is a difference between attending to something and con-
ceptualizing it. If a child feels pain she certainly attends to the pain,
whether she applies concepts to it or not. The psychological literature
strongly indicates that all experiences are altered to some degree by the
particular way in which attention is focused, and Hilla Jacobson and I
have argued, on the basis of those findings,21 that counterfactuals to
the effect that a subject would have been able to apply the appropriate
demonstrative concept to a sensory experience if she had attended to it
presuppose an untenable “snapshot model” of sensory experience, that
19. According to Ned Block, “Seeing-As in the Light of Vision Science,” Philosophy and
Phenomenological Research 89 (2014): 560–572, cases of what I am calling (visual) “appercep-
tion” are cases of “secondary seeing,” combinations of both primary visual repre sentation
(in Burge’s sense of representation) and ensuing conceptualization. I fi nd this plausible.
[Note added July 7, 2015.]
20. Alva Noë, Action in Perception (Cambridge, Mass.: MIT Press, 2004).
21. We argue this in “Against Perceptual Conceptualism,” cited in note 3 above.
Unauthenticated
is, a model in which a given sensory experience would have been just
the same if one had attended to it, just as a tree would have been the
same if one had photographed it. (McDowell employs such counterfac-
tuals to explain the sense in which our experiences are allegedly “con-
ceptualized” in “Avoiding the Myth of the Given.”) The phenomenal
character of a red patch or a pain are different when one attends to them,
but I don’t believe that they are necessarily dif ferent when one recognizes
the color as red or the pain as a pain (although sometimes they may be)
and hence I don’t need to deny that an animal or a child who has not
learned to talk and I ever have “qualitatively” the same or similar feel-
ings of pain or sensations of red (even when we both attend to those
feelings or visual sensations). But I do maintain that I have conceptu-
alized awareness that the red I experience is the color of a surface (or the
result of a blow on the head, as the case may be), and that the pain is an
unpleasant sensation, that the prelinguistic child and the animal lack.22
Using Kantian terminology, we may say that the prelinguistic child
and the animal see red and feel pain, but they do not apperceive pain,
that is, they do not recognize pain as pain. The kind of awareness that
constitutes a mature human being’s apperceptive awareness of a sensa-
tion is like a belief, in that it can be accepted or rejected (whereas it makes
no sense to accept or reject the mere sensation!), and it presupposes we
have the requisite background of concepts, but it is not an explicit judg-
ment, although it can justify a judgment.
To repeat, apperceptions are conceptually shaped, and they can justify
judgments. But they are not the same as perceptual beliefs. I may have an
apperception or, better, a seeming apperception that one “shaft” is longer
than the other in the Müller-Lyer illusion, but I do not believe what I
seem to be apperceiving. One can seem to oneself to apperceive when
one knows that one does not (successfully) apperceive, but one cannot
seem to oneself to believe when one does not believe. The distinction
between sensations and apperceptions is not the same as the distinction
between sensations and perceptual beliefs.
I say this because I fear that other wise this will sound to McDowell
like Gareth Evans’s story, according to which “the informational
22. Although if Burge is right, the prelinguistic child and even the animal may enjoy
representations of the fact that the color is on a surface and the pain is unpleasant that are
more primitive than conceptualizations. [Note added July 7, 2015.]
Unauthenticated
system” supplies inputs, which causally trigger (but do not justify) the
relevant beliefs.23 This, McDowell might say, is Davidson and Neurath
all over again. But while this is right as a criticism of Evans’s account,
it is not right as a criticism of the account I just proposed (and not only
because apperceptions can be rejected; they do not automatically “trigger”
beliefs, although that is an important point about them). What Evans
was doing was speculating about the brain’s hardware; thus his account
was a reductionist account. In my view, talk about what the organism
does in intentional terms is not reducible to “hardware” talk.24 To be
sure, any intentional state, such as knowing the meaning of the word
“ephebe,” has to be realized in “hardware” somehow. But intentional
states are realizable in a practically infi nite number of dif ferent ways,
and the fact that a particular piece of hardware is a realization of a par-
ticular intentional state is only visible from the intentional level, not
from either the neurological or the computational level by themselves.
Amodal Apperception
When my apperception is not of something that produces a sensation
but of something “amodal,”25 it is even clearer that an apperception is
dif ferent from a sensation. Amodal apperception includes awareness of
aspects that are “present in absence” such as my awareness of a tomato
that I am looking at as having another side. Alva Noë accounts for such
apperception in terms of sensorimotor expectations,26 but he does also
mention shaping by thought as a possibility. And it is more than a mere
possibility, in fact; my awareness of the computer in front of me as a
computer does not consist just of “sensorimotor” expectations.
23. For McDowell’s description of Gareth Evans’s view, see Mind and World, 47–63.
24. This is something I argue in many papers; among them are Hilary Putnam,
“Meaning Holism,” in The Philosophy of W. V. Quine, ed. L. E. Hahn and P. A. Schilpp (La
Salle, Ill.: Open Court, 1986), 405–431; reprinted in Realism with a Human Face (Cambridge,
Mass.: Harvard University Press, 1990), 278–302; and Hilary Putnam, “Functionalism:
Cognitive Science or Science Fiction?,” in The Future of the Cognitive Revolution, ed. D. M.
Johnson and C. E. Erneling (Oxford: Oxford University Press, 1997), 32–44; reprinted in
my Philosophy in the Age of Science, ed. Mario De Caro and David Macarthur (Cambridge,
25. On “amodal” awareness, see Alva Noë, Action in Perception (Cambridge, Mass.: MIT
Press, 2004), 77–79.
26. Alva Noë, Action in Perception, 7–9, and throughout the volume.
Unauthenticated
Still, I think we should agree when Noë writes that
the way the coin’s circularity is present to me when I view it from

a single point of view, is as available to me thanks to my implicit un-
derstanding that were I to move in these and these ways its ap-
parent shape would change correspondingly . . . The way in which
the strictly unperceived circularity of the coin is present is decid-
edly not the way inferred objects are judged to be present. This is
demonstrated by the fact the coin would look no less circular even
if one knew, for a fact, that it was elliptical. This is a familiar oc-
currence when we look at pictures. What is drawn is an ellipse, but
we experience it as the outline of a plate, say.27
In fact, all the claims McDowell makes about perceptual experiences

are true of apperceptions, including these amodal apperceptions: they
are conceptualized but noninferential, and they can be judged as wrong
even though they seem right.
Of course, there are many questions to be investigated if the picture
I suggest is right, including the contribution (overstressed by Noë, but
undoubtedly impor tant) of sensorimotor skills. But that is all to the
good; these are questions that a broad-minded and nonreductionist cog-
nitive science should investigate. It is all to the good that a philosoph-
ical account should link up with plausible research programs in “natural
science.”
What will probably make my story unacceptable to McDowell is that
it isn’t our sensations that are conceptualized on my account. And that
means that sensations aren’t the “tribunal.” But that I think is precisely
right. If “minimal empiricism” insists other wise, then “minimal empir-
icism” is too empiricist. Apperceptions are the tribunal, not sensations.
27. Alva Noë, “Real Presence,” in Varieties of Presence (Cambridge, Mass.: Harvard Uni-
versity Press, 2012), 54, 51.
Unauthenticated
D 10
Perception without Sense Data
B e c ause t h e L au e n e r P r i z e was bestowed upon

me for my entire “oeuvre” in analytic philosophy, it would be appro-
priate for me to say something about all of my past work, but a lecture
that covered all that would be much too long. Instead, because that seems
a more reasonable task, I shall mostly confine my discussion to my views
in the philosophy of mind as they developed and are still developing after
I gave the Dewey Lectures at Columbia University in 1994.1
Those lectures, I have to say, were often read in a way that I found
disappointing. Many analytic philosophers saw them as a “capitulation”
to Wittgensteinian or Austinian ordinary language philosophy. At the
same time, what I myself thought was most important in those lectures,
namely, the explanation of my reasons for giving up the “internal
realist” views I defended in Reason, Truth and History2 (and subsequently,
in several papers written between 1976 and 1990),3 went almost unno-
1. Hilary Putnam, “Sense, Nonsense, and the Senses: An Inquiry into the Powers of the
Human Mind,” Journal of Philosophy 91 (1994): 445–517; reprinted in The Threefold Cord:
Mind, Body and World (New York: Columbia University Press, 1999), 3–70.
2. Hilary Putnam, Reason, Truth and History (Cambridge: Cambridge University Press,
1981).
3. Most of those papers are collected in Hilary Putnam, Realism and Reason, vol. 3 of
Philosophical Papers (Cambridge: Cambridge University Press, 1983), and Realism with a
Human Face (Cambridge, Mass.: Harvard University Press, 1990).
152
Unauthenticated
Perception without Sense Data 153
ticed, or, when noticed, those reasons were often misunderstood. Obvi-
ously I was not as clear in the Dewey Lectures as I should have been.
Let me begin by saying something about these points, before I turn to
philosophy of mind.
To take the second point fi rst, it appears that a reason that many
people continued to describe me as an “internal realist” is that they con-
fused two dif ferent doctrines: conceptual relativity, a doctrine I con-
tinue to defend, and internal realism, a doctrine that now I regard as
mistaken. Conceptual relativity is the claim that sometimes two scien-
tific theories have dif ferent “ontologies” in the familiar Quinian sense,
that is, “taken at face value” their quantifiers range over dif ferent sorts
of objects, and yet there is a systematic way of interpreting each theory
in the language of the other that renders them (not just empirically
equivalent, but) explanatorily equivalent—that is, the translation of
any explanation of a physical phenomenon provided by one theory is
a perfectly good explanation of the same phenomenon in the language
of the other theory (but one that is very different, at least at “face value”).
Internal realism, however, was the claim that what truth amounts to is
justification under epistemically ideal conditions. At one time I thought
that the existence of cases of conceptual relativity was an argument
against metaphysical realism and hence for “internal realism” (which I
took to be the only reasonable alternative to metaphysical realism). But
that was a mistake.
It was a mistake because one can perfectly well be a metaphysical re-
alist (as, perhaps to your surprise, I now admit to being) in the simple
sense of believing that there is a real world that is largely independent
of our mental workings, beliefs, or concepts, and many truths about that
world that outrun what we can possibly verify, without denying the ex-
istence of the phenomenon I call “conceptual relativity.” If, for ex-
ample, it turns out, as according to some quantum physicists it already
has,4 that a theory in which certain particles are fermions and a theory
in which the same particles are bosons are simply two “representations”
of the same facts, then the moral will be that what is fundamental, at
least in the ontology of physics, is not particles but states of affairs. If
what are dif ferent ontologies in the familiar (Quinian) sense I referred
4. For example, C. F. Burgess and F. Quevedo, “Bosonization as Duality,” Nuclear Physics

B 421, no. 2 (1994): 373–390.
Unauthenticated
to above are just dif ferent representations of the same states of affairs,
then that may be a problem for the Quinian ontologist, but I no
longer think it is fatal for metaphysical realism. I repeat, one can accept
the possibility of conceptual relativity and simultaneously reject internal
realism. A fortiori, they are not the same thing.
On the other hand, the charge that my Dewey Lectures represented
a “capitulation” to Wittgenstein and / or Austin is partly but not wholly
right.
In my “Wittgenstein: A Reappraisal,”5 I criticized a view that Witt-
genstein undoubtedly held, the view that metaphysics is, in some very
literal sense of the term, nonsense—that is, completely devoid of lin-
guistic meaning, as claimed by the “New Wittgensteinians,” or in vio-
lation of the “grammatical conditions for making sense” as claimed by
more traditional Wittgensteinians. But I do not find, when I read Plato,
or Aristotle, or Maimonides, or Kant, or even Hegel, that I am liter-
ally reading nonsense, although there is much about their thought that
I cannot imagine believing. Some of my talk about “unintelligibility”
in the Royce Lectures now seems wrong to me. Although those lec-
tures were published together with the Dewey Lectures, Wittgenstein
figures in the Dewey Lectures themselves in a much more modest role,
although I do engage in some metaphysics bashing.
Austin, however, was clearly front and center in the Dewey Lectures.
But so was William James. They both figured, not as metaphysics
bashers, but because the main point of the Dewey Lectures was that
philosophy of perception needs to return to the view of the man and
woman on the street, that is, to a view in which we actually perceive
shoes and ships and sealing wax and cabbages and kings and many of
the other items in the passing show—the view that Russell dismissed
as “naive realism.” And I saw Austin and James as allies because (albeit
in dif ferent ways) they both criticized the traditional epistemological
paradigm, under which what we “directly perceive” are just our own
sense data, and external objects are known by “inference,” if they in-
deed exist as something more than logical constructs out of sense data
or “permanent possibilities of sensation.” [Note: In this essay, I shall use
5. Chapter 28 in Hilary Putnam, Philosophy in an Age of Science: Physics, Mathematics, and

Skepticism, ed. Mario De Caro and David Macarthur (Cambridge, Mass.: Harvard Univer-
sity Press, 2012), 482–492.
Unauthenticated
“sense data” to refer to experiential phenomena that are supposed to be

mental but nonconceptual, and to occur in our “private spaces” (fol-
lowing Russell, not Moore).]
That we need to reject that paradigm is something I still believe.
However, I went further and opted for “disjunctivism,” which is the
view that, in veridical perception— say, seeing objects in one’s vicinity
as they actually are—there are no such objects as sense data at all. Ac-
cording to disjunctivists, even a “perfect hallucination”—for instance,
a hallucination produced by an “experience machine” that causes one’s
visual cortex and all the other relevant parts of the brain to be in ex-
actly the state that they would be in if one were perceiving, say, a white
cat on a blue sofa— and the veridical experience one would have in that
state have no common elements. It is true that the two experiences
cannot be distinguished on the basis of how things seem to one visually,
but, according to disjunctivists, one cannot say that they are indistin-
guishable because they have the same phenomenal quality or qualities. One
can only say that they are indistinguishable, Punkt.
It is no accident that reading Austin’s Sense and Sensibilia influenced
me in the direction of disjunctivism. Although a close friend of Austin’s,
J. M. Hinton, is usually credited with being the father of disjunctivism,
Austin surely deserves to be called at least its grandfather. For, although
most of Austin’s Sense and Sensibilia consists of arguments against Ayer
and other traditional epistemologists, if one tried to extract a positive
claim about perception from that book, it would surely be in the gen-
eral ballpark of disjunctivism. And it now seems to me that the disjunc-
tivist view cannot be right. But neither can Russell’s view according
to which the colors we see when we look at a table (and the smoothness
we feel when we touch it, etc.) are in our own “private spaces,” that is,
colors and textures and shapes are just our own sense data, and objec-
tive, physical colors, textures, shapes are inferred properties of, in
effect, noumenal objects.6
6. This is argued in the fi rst three chapters of Bertrand Russell’s Problems of Philos-
ophy (Oxford: Oxford University Press, 1912). In chapter 1, Russell proposes that “we
give the name ‘sense data’ to the things that are immediately known in sensation: such
things as colors, sounds, smells, hardness, roughness, and so on” (12) [I have American-
ized the spelling]. By chapter 2, we have learned that “the sense data are private to each
separate person” (20), and in chapter 3 we are told, “Now our sense- data are situated in
Unauthenticated
In my Dewey Lectures, I defended disjunctivism with a skeptical ar-

gument: I claimed (following Wittgenstein) that there could no be
such things as sense data (aka “qualia”) because there is no criterion for
intersubjective sameness here, or even sameness of sense data at dif-
ferent times for one subject.
The par tic u lar form of disjunctivism I sympathized with in the
Dewey Lectures is due to John McDowell (although, in an afterword
to the volume in which I collected those lectures,7 I refused to en-
dorse the claim that all experiences are conceptualized, which figured
large in Mind and World).8 I now believe, for example, that the best
“mesh” of phenomenological considerations and neuroscientific ones
will involve recognizing that many of our experiences are not concep-
tualized in any of the senses that McDowell has proposed, and will
also involve rejecting the idea (advocated by Wittgenstein and later by
Hinton) that there is no possibility of a scientific criterion for same-
ness and difference of, say, color experience of dif ferent subjects in
which intersubjective spectrum inversion is an intelligible possibility.
In sum, I believe that there is a qualitative and nonconceptual dimen-
sion of experience that can be scientifically investigated. To use a term
revived by Ned Block, there are qualia, and progress can be made and
has been made in saying something about their neurophysiological
basis.9 But if so, the problem becomes how does one criticize the Rus-
sellian conception I described in which what we perceive are just our
own qualia?10
The ramifications of this question are large, precisely because one
dogma of empiricism is still very much with us, and that is the dogma
that perception begins with sense data.11 If disjunctivism were right,
our private spaces” (30) and not in “one all- embracing physical space in which physical
objects are” (31).
7. In The Threefold Cord, I wrote, “My own ‘Girondist’ position is that while it may be
that some experiential content is nonconceptual in the sense of not being conceptualized,
such epistemologically impor tant content as recognizing something as an object of a cer-
tain kind is irreducibly conceptual” (156–157). In the note to that sentence I added that, “I
fi nd McDowell’s talk of our conceptual powers being in play uncomfortably reminiscent of
faculty psychology” (218n24).
8. John McDowell, Mind and World (Cambridge, Mass.: Harvard University Press, 1994).
9. I fi rst saw it in the writings of C. I. Lewis.
10. I shall use Russell’s term at times to bring out the traditional roots of the current
debates about perception, but I will not distinguish between “qualia” and “sense data.”
11. Note that following David Chalmers, The Conscious Mind (Oxford: Oxford University
Press, 1996), but not Block, I do not take talk of qualia to rule out “adverbial” views of phe-
Unauthenticated
the problem of avoiding the errors of empiricism would be solved—just

deny that there are any such things as “sense data.” Voilà, perception
without sense data! But for the sorts of reasons I just referred to, I think
this dismissive strategy is misguided.
Indeed, it may seem obvious that, if there are sense data, then they
are what knowledge of the world is “based on.” What else do we have,
after all? Even before we try to see more clearly what the alternatives
before us really are, here is a point to keep in mind: the fact (if you agree
with me that it is a fact) that our experience has a qualitative and noncon-
ceptual dimension does not entail that we perceive qualia. Indeed, Moore
himself thought that it is difficult to perceive qualia, and Shoemaker has
argued that we cannot perceive them. As Block writes,
Shoemaker’s view is shared by Fred Dretske, Gilbert Harman, Mi-

chael Tye, and many others who advocate what G. E. Moore
termed the diaphanousness (or sometimes the transparency) of ex-
perience. Harman puts the point by saying that the more one
tries to attend to one’s experience of the tree, the more one attends
to the real tree instead. Although Moore is sometimes cited as the
originator of this point, he did not actually accept it. I have heard
him quoted saying “the moment we try to fi x our attention upon
consciousness and to see what, distinctly, it is, it seems to vanish:
it seems as if we had before us a mere emptiness. When we try to
introspect the sensation of blue, all we can see is the [objective]
blue; the other element [the quale] is as if it were diaphanous.” But
these words are followed by what I regard as a more significant
truth: “Yet it can be distinguished if we look attentively enough,
and know that there is something to look for.”12
nomenal qualities, that is, views under which they are modifications of experiences rather
than “mental paint,” and ditto for “sense data.” A variant— one I prefer, in fact—of adver-
bialism, due to Wilfred Sellars, “The Adverbial Theory of Objects of Sensations,”
Metaphilosophy 6 (1975): 144–160, treats “I have a red sense datum” as better analyzed as “I
sense redly.” In this view, it is the case, as urged by Reichenbach in Experience and Predic-
tion, that what correspond to sense data / qualia are states of the subject, not properties of
mental particulars (see p. 100n5).
12. G. E. Moore, “A Reply to My Critics,” in The Philosophy of G. E. Moore, ed. P. A.
Schlipp (New York: Tudor Publishing, 1952), 546. The words in square brackets are my
interpretations of Moore’s expressions.
Unauthenticated
Hilla Jacobson and I are currently working on a book on the quest

for naive realism in contemporary philosophy of mind. We hold nei-
ther Shoemaker’s view that there is no such thing as attending to one’s
qualia nor Block and Moore’s view that we have to perform a special
act of “looking attentively enough” (unless what they mean to point out
is just that we have to conceptualize differently when we attend to the
qualitative aspect of what is presented in an experience as opposed to
its representational content). We can attend both to the objective color
(for example) of something (that shirt is blue) and to the “look” of that
color, and, indeed, to various “looks” that it (potentially) has—how it
looks from here, for example, or how it looks in the daylight, and also
to more subjective looks. In the case of veridical perception, all of those
looks are genuine properties of the object seen. I have been aware since
I was a very young man that the look of colors is dif ferent if I close my
right eye and if I close my left eye (a very common phenomenon, ac-
counted for by differences in the two macular areas). So one of the many
looks I can attend to is the look of the shirt with my right eye closed,
and another one is the look of the shirt with my left eye closed. Those
looks are properties, albeit relational ones, of the shirt in question. Ex-
periencing “phenomenal qualities” is inseparable from perceiving
properties of things “out there,” but it isn’t simply receiving informa-
tion about how things out there are independently of being perceived
(contra “representationalist” philosophers of perception). The qualia are
different. And contra the disjunctivists, if we produce those looks without
the aid of an object that has the appropriate objective colors (e.g.,
via the sort of experiments famously performed by Edwin Land), then
we do not perceive a color property of anything external to ourselves,
but we can still attend to the qualitative dimension of the experience,
and we can perceive its similarity to the qualitative dimension of a
corresponding veridical experience. But Block is right to concede, as
indeed Moore already does, that in most cases we perceive one or more
of the objective properties (that shirt is blue, that shirt has such and
such a look in artificial light, etc.) and do not conceptualize or “apper-
ceive” the subjective aspect.
With the above in mind, let us now consider three dif ferent models
of perception. The empiricist model, in which perception is based on
“sense data” (aka “impressions and ideas”) is still with us, but there are
many other models that simultaneously attract both laypersons and phi-
Unauthenticated
losophers. Two of the most important are Quine’s “surface-irritation”

model, and the Kantian model favored by McDowell (although I dis-
cern the influence of the empiricist model at times in his writing, such
as the times he uses “experiences” and “impressions” interchangeably,
and emphatically insists on the importance of preserving “minimal
empiricism”).
I. The Empiricist Model

Bertrand Russell vacillated over the years, but in his great little book,
The Problems of Philosophy, he held that colors are in our own minds, just
as Berkeley and Hume thought, and that we infer the existence of phys-
ical colors, physical shapes, and so on, from the properties of these
mental entities, and this is a standard empiricist view (sometimes dolled
up with the idea that the inferences are “unconscious,” or that we don’t
really make them but it is enough that we could make them, if we were
sufficiently “rational”). Many have criticized this view on metaphysical
and epistemological grounds, and with good reason, but today I want
to look at this and the other two models I mentioned from the stand-
point of psychology. From that standpoint, the immediate problem is
that we obviously don’t make the sorts of inference that Russell
described in The Problems of Philosophy—not unless we are Bertrand
Russell! (Nor did he think we made them; in his view most people are
utterly ignorant of their epistemological predicament.) Indeed, the
idea that we make such inferences is incompatible with the traditional
empiricist claim that (unless we have been educated by philosophers
like Russell) we simply project our qualia onto the external world
without realizing that there is a difference between those objective
properties and our qualia.
What about the dolled-up version, according to which we make such
inferences “unconsciously”? At best, this is a hypothesis about subper-
sonal mechanisms, and I believe that we also need an acceptable
personal-level description of perception, and not only an account of the
subpersonal mechanisms that subserve perceiving, thinking, deciding,
and so forth. But even at the subpersonal level, we now know of many
more kinds of computation that the brain is capable of using than prop-
ositional inference, so even as a subpersonal account, “unconscious in-
ference from our sense data” is out of date.
Unauthenticated
Moreover, the inputs to the brain are not qualia, but not because
qualia aren’t features of brain events (I believe that they are, and, in any
case, they certainly supervene on brain events). The idea of qualia as
inputs might have been at least partially vindicated had it turned out
that Fodor’s “modularity” view, on which phenomenal appearances are
simply outputs of perception-modules, turned out to be right. In that
case qualia might be identified with those outputs, and those outputs
in turn identified with the first stage of cognitive processing. But the
consensus in neural science today favors a much more complicated
story. On the modularity hypothesis, perception was supposed to be
an entirely “bottom up” affair, but today it is becoming clear that that
perception is both “bottom up” and “top down.” When I see a tomato,
for example, many parts of the brain are involved, and interactions
(reverberating circuits) take place between cortical and precortical
functions. If there are “inputs” to the brain, they begin with the eyes—
much earlier than the formation of “qualia.”
One might try to bypass this objection by adopting a dualistic view
of the unconscious mind, and posit that while the inputs to the brain
are neural stimulations, beginning in the eyes, the inputs to the uncon-
scious mind are “mental” qualia, but at this point all contact with both
science and phenomenology would have been lost. The “unconscious
inference” story was cutting edge psychology at one time,13 but it is not
tenable as either personal level or subpersonal level psychology today.
What about the version according to which the story about inferences
from qualia to material objects is supposed to be a “rational reconstruc-
tion”? When we make explicit the steps in a proof that are left implicit
in a mathematics journal, for example, we are engaged in rational
reconstruction, but in such a case the steps could have been made explic-
itly to begin with (the article would simply have been boring to read).
But to say why I believe that there is a computer in front of me as I type
these words by (1) describing all the relevant “sense data” and (2) pro-
viding an inference from all that information (assuming it exists, and
assuming it could all be stated) to “there is a computer in front of me”
is not something I could really do. At best (and assuming we ignore all
13. H. L. Helmholtz, Handbuch der physiologischen Optik, vol. 3 (Leipzig: L. Voss, 1866);
repr. 3rd ed., with extensive commentary by Allvar Gullstrand, Johannes von Kries, and
Wilibald Nagel (Hamburg and Leipzig: L. Voss, 1896), 28–29.
Unauthenticated
the objections that critics of traditional epistemology have raised), all

that the “rational reconstruction story” could show is that a person with
literally superhuman capacities for phenomenological description,
retention of information, and rapid reasoning would be justified in
believing that there is a computer in front of him or her, and not that
persons like ourselves are justified in our perceptual beliefs. Perhaps in
such a case justification is not needed—justification comes to an end,
as Wittgenstein said in On Certainty14 —but providing such a justifica-
tion was precisely the purpose of the “rational reconstruction”!
II. The Surface-Irritation Model15

Although in some of his writing in the 1950s Quine occasionally spoke
of “sense data,” his considered position was that (1) “sense data” are
“posits,” just as much as say electrons or genes, and not the subject of a
kind of knowledge that is independent of theory, as the sense datum
philosopher claims and (2) unlike electrons and genes they are bad pos-
its.16 What, then, is Quine’s picture of perception?
The answer is a straightforward causal story. Light from the window
and the desk lamp strikes the computer and is reflected to my eyes, neu-
rons in my eyes are stimulated, my optic nerves are stimulated, my
brain does various things, and as a result I acquire a set of “linguistic
dispositions” I didn’t have before, such as the disposition to assent if
Quine (or whoever) points to the computer and utters the “occasion-
sentence” computer.17 To the charge that this is not a story on which
my belief that there is a computer there is justified, Quine ought not to
respond, as he sometimes did, by speaking of the stimulations of those
neurons as “evidence” and of my saying “yes” to the query “Is that a
computer in front of you?” and the like as “warranted beliefs”; he should
simply have said that such notions as “warrant,” “belief,” and “evidence”
14. Ludwig Wittgenstein, On Certainty, ed. G. E. M. Anscombe and G. H. von Wright

(Oxford: Blackwell, 1969), §192.
15. The term “surface irritation” was famously used by Quine for the triggering of sensory
receptors in “The Scope and Language of Science,” in W. V. Quine, The Ways of Paradox and
Other Essays, rev. ed. (Cambridge, Mass.: Harvard University Press, 1976), 228–230.
16. Here I follow Peter Hylton’s Quine (Abingdon, United Kingdom: Routledge, 2007), 87.
17. The notion of an “occasion sentence” was famously used by Quine in Word and Object
(Cambridge, Mass.: MIT Press, 1960).
Unauthenticated
do not belong in science. But that even he was not willing to consis-
tently go that far reveals how unappealing the position is, as McDowell
and others have pointed out.
III. The Kantian Model

Kant’s great contribution was to call to our attention the ways in which
what we perceive is dependent on what we conceive. For example, to
perceive—in the sense of recognizing the fact that— one event hap-
pened later than another requires possessing the notion of an objective
time order, and that notion, Kant claimed (I believe rightly), is depen-
dent on (or rather interdependent with) such conceptual resources as
the notions of cause and effect. It is not enough that I experience events
in a certain subjective order; I must also know that in certain cases I
could have perceived them in a different order (I could have walked
around the house in the reverse direction, for example, and still perceived
the front, the sides, the chimney, etc.), and in other cases I could not have
perceived them in a different order (I could not have perceived that boat
sailing away from the bridge before I perceived it approaching the bridge).
Conceptualization is a precondition for full-blown perception.
In Mind and World, McDowell, however, takes Kant’s point to be that
all experiences are apperceptions (recognitions that something is so and
so). This may have been Kant’s view too; at least, he writes that an “in-
tuition” to which we do not prefi x the “I think” would “be nothing to
us,” and this could be interpreted as meaning that, as far as we can know,
our whole mental life is apperceived by the ego. If so, I believe that Kant
made a mistake, and that McDowell makes the same mistake.18 I think
we are aware that there is a great deal of detail in the scene presented
to our eyes at any given moment that we do not apperceive, and some
of that detail can be recalled by us a moment or two later. But one does
18. I also have problems with McDowell’s “minimal empiricism.” I quote from Mind and
World (Cambridge, Mass.: Harvard University Press, 1994): “Empirical judgments in
general—whether or not they are justified at all, perhaps less substantially than knowledge
requires—had better have content of a sort that admits of empirical justification, even if
there is none in the present case (say in a quite unsupported guess)” (6). As it stands, this
reads like a “verifiability” requirement, and if that is not how McDowell understands it, it
is noteworthy that he does not feel it necessary to explain why it does not mean that every
“empirical judgment” must be confi rmable or disconfi rmable by experience.
Unauthenticated
not have to speculate about this; there is good experimental work to

support the view that what we attend to in experience is far from being
the whole of experience. The idea that what is presented in experience
outruns what is attended to, conceptualized, and so on, is not simply
an illusion, as some have tried to claim. (Of course the point applies
also to senses other than vision.)
The point is not trivial, because claiming that only creatures capable
of fully conceptualized apperception—that is, of apperception whose
content they can formulate in language— can have experiences of
the sort we have posits an enormous, and I believe untenable, gulf be-
tween our experiences and those of animals and even those of children
who are not yet masters of a language.
Seeing apperceptions and not bare qualia as the basis of empirical
judgment involves a fundamental shift in perspective.
The fact is that when we have an apperception (or a seeming-
apperception—think of the Müller-Lyer illusion!), no particular “im-
pression” is essential. Block and others have pointed out that there is
considerable empirical evidence that dif ferent subjects experience dif-
ferent qualia when they look at a particular color; but, unless they are
color blind, they can all be truly said to see (apperceive) that the traffic
light is red, when it is. The apperception or seeming-apperception that
the traffic light is red is not the same as an “impression” in the sense of
a sensation. Nor is it the same as a perceptual belief. (There is no such
thing as it seeming to me that I believe something when I know that I
don’t believe any such thing, but there is such a thing as seeming to
apperceive something when I know that I don’t.) When we look for a
justification for a particular empirical judgment, apart from the very
special case in which the judgment is about qualia, what we cite as ob-
servational support are apperceptions and not qualia (or “sense data”).
Of course, McDowell also does not think we do cite sense data (in
the sense of qualia) as empirical support for our judgments; he thinks
there are no such things as qualia, and that visual impressions, for ex-
ample, are simply takings-in of properties of the objects we see. If I see
a red book, the “impression” that the book is red is just a taking-in of
the fact that the book is red. This is not idealism, as some critics of Mc-
Dowell have charged; however, it is simply wrong. What it is is an at-
tempt at an account of perception without any reference to sense data
at all. I agree with McDowell that qualia cannot do the epistemic work
Unauthenticated
of supporting our empirical judgments. In my view, that work is done

by our apperceptions and seeming-apperceptions. In fact, perceiving
something is sometimes not accompanied by any special qualia. Per-
ceiving that I raised my arm intentionally (as opposed to its just “going
up”) is something I often do, but there is no “quale” of “voluntariness.”
Moreover, there are—as Alva Noë has stressed—many forms of “amodal”
awareness. The awareness that I am seeing a tomato includes both the
awareness that the side toward me has a certain color and the aware-
ness that it has a soft interior and a side I do not see.19 But the fact that
knowledge is grounded in apperceptions and seeming-apperceptions
and not in unconceptualized sense data does not imply that the latter do
not exist. For our experience is rich in qualia, most of which we do not
conceptualize, although many of them can be conceptualized to some
extent both when they occur and after they occur. Qualia can even play
an epistemic role, when they are what we apperceive. But so can tables
and chairs and cats when they are what we apperceive. I am arguing that
qualia, as opposed to apperceptions, play no special epistemic role, but
disjunctivists are wrong to deny their existence.
To summarize, the Kantian picture of perception that McDowell
defends in Mind and World is simply this: our impressions are already
conceptualized.20 Also, disjunctivism guarantees that our “impres-
19. See Alva Noë, Action in Awareness (Cambridge, Mass.: MIT Press, 2004). While Noë
perhaps overemphasizes the role of sensorimotor expectations in such “amodal” awareness,
they are without doubt a significant part of the story.
20. In John McDowell, “Avoiding the Myth of the Given,” in Having the World in View:
Essays on Kant, Hegel, and Sellars (Cambridge, Mass.: Harvard University Press, 2009), he
gives up the requirement that experiences must be “articulated” or at least “articulable” like
propositions, and sums up some key elements of his current view as follows:
Even though the unity-providing function is a faculty for discursive activity, it is not
in discursive activity that these capacities are operative in intuitions. With much of
the content of an ordinary visual intuition, the capacities that are in play in one’s
having it as part of the content of one’s intuition are not even susceptible of discur-
sive exercise. One can make use of content’s being given in an intuition to acquire a
new discursive capacity, but with much of the content of an ordinary intuition, one
never does that. (Think of the fi nely discriminable shapes and shades of color that
visual experience presents to one.) Nevertheless an intuition’s content is all concep-
tual, in this sense: it is in the intuition in a form in which one could make it, that very
content, figure in discursive activity. That would be to exploit a potential for discur-
sive activity that is already there in the capacities actualized in having an intuition
with that content. (265)
Unauthenticated
sions” are not mental entities (“qualia”) common to veridical perceptual

experiences and certain illusions and hallucinations. Instead, they are
simply takings-in of how it is in the environment (in the case of vision,
for example). If I see a white rabbit, that there is a white rabbit there is
a fact about the world, and it is the conceptual content of some of my
visual impressions. Naive realism is the personal level account of per-
ception we need. The only task for philosophy is to explain that this is
the case. Neural science is fine in its place, but what it studies is merely
subpersonal mechanisms, and philosophers don’t need to know any-
thing about them. (A view that understandably infuriates some philos-
ophers, who think that philosophy of mind needs to be scientifically
informed.)
If you are like me, it will seem to you that none of these models can
suffice. The empiricist model was deeply infected with Berkeleyan ide-
alism from the start; the surface-irritation model does express our need
to understand the mechanisms of perception, but it ignores the envi-
ronment outside the skin (except for those “gavagai,” whatever they are)
and says nothing about the nature of the subpersonal processing as well
as the specific nature of the environment– organism transactions. The
Kantian model, as “naturalized” by McDowell, purports to defend com-
monsense realism, but it embraces a “disjunctivism” that is anything
but commonsensical. And all three models ignore a point (stressed by
both Dewey and Wittgenstein—an unlikely couple!) that what we per-
ceive depends on what we do. Action and perception are interdependent.
(Not that any of these models deny that, to be fair. But their authors
evidently didn’t think it important.)
Liberal Functionalism
In closing, what I want to do isn’t to present a fourth model of per-
ception, but to describe a way of thinking about the topic that can
help us see how studies of perception at many levels and with many
tools— perceptual psychology (including, importantly, the kind of
psychology pioneered by Gibson), evolutionary biology, neural sci-
ence, behavioral science, and conceptual analy sis— can fit together,
and how personal-level psychology, subpersonal neural science, and
studies of organism– environment transactions at more than one level
can all fit together. The aim is metaphysical in the best sense of the
Unauthenticated
term, trying to understand “how things, in the widest possible sense of

the term, hang together,” and is also scientific without being reductive.
Over the years, I have used the term “functionalism” in both a nar-
rower and a wider sense. In the narrower sense, functionalism held that
the mind is simply a computer, one with “inputs from the sense organs”
(like Quine’s “surface irritation”) and outputs to “suitable motor or-
gans,” and that to understand our mental functions it would suffice to
find the right description of the “program” of the computer. This com-
puter model of the mind led me to the “verificationist semantics”21
that I mentioned at the beginning of this essay, and that semantic theory
was the axle around which “internal realism” rotated.
That form of functionalism, computational functionalism, I gave up
even before I gave up internal realism,22 and in the Dewey Lectures
(written after giving up internal realism) I argued that it is hopelessly
solipsistic. In fact, my finding solipsism unavoidable in any picture that
limits mental life to what happens inside our heads was the impetus for
seeking a way to come closer to naive realism without coming so close
as to deny the phenomenal characters of experience. And in my Pro-
metheus Prize Lecture in 2010,23 I proposed a naturalist but antire-
ductionist view I called “liberal functionalism.”24 That view agrees with
my former computational functionalism that the question psychology
needs to address is the description of the various functions (or rather
functionings) of an organism, rather than the question of its physical
or biological makeup, but it rejects completely both the restriction to
computer programs as the sole admissible way of describing those func-
tions (which is not to say that computer programs can never be rele-
vant); it also rejects the idea that what goes on between the “sensory
21. This connection is explicit in Hilary Putnam, “Computational Psychology and In-
terpretation Theory,” in Artifi cial Intelligence: The Case Against, ed. Rainer Born (London:
Routledge, 1987), 1–17; reprinted in my Realism and Reason, vol. 3 of Philosophical Papers
(Cambridge: Cambridge University Press, 1983), 139–154.
22. Compare with Hilary Putnam, Representation and Reality (Cambridge, Mass.: Har-
vard University Press, 1987).
23. My Prometheus Prize Lecture, “Corresponding with Real ity,” is collected in De
Caro and Macarthur, Philosophy in an Age of Science, 72–91.
24. The term “liberal functionalism” was suggested by Mario De Caro and David
Macarthur’s use of “liberal naturalism” in the introduction to their Naturalism and Norma-
tivity (New York: Columbia University Press, 2010).
Unauthenticated
receptors” and the “signals from the brain to the motor organs”25 is all
that psychology is concerned with, as my computational functionalism
assumed.
Instead, liberal functionalism holds that we have naturally evolved
functions for dealing with specific environmental contingencies, 26
as well as functions that develop in the maturational history of each
individual—in the case of humans, functions that depend on and in
turn condition the sort of language we understand and use. Rather than
view what happens between the object we see and eye as simply a
matter for physics, for example, a liberal functionalist will follow Gib-
son’s pioneering lead in seeking for a description of the complex relations
between the properties of the object in question, the organism’s reper-
toire of actual and possible sensory-motor transactions with the object,
and the ways in which the object is perceived. And the vocabulary to
be used in describing those functions will not be circumscribed in ad-
vance. As I wrote in my Prometheus Prize Lecture,
An up-to-date liberal functionalist shouldn’t think she has to re-

duce all the notions she uses to non-intentional notions. It is true
that we have no idea how to reduce the predicate ‘refers to’ to non-
intentional predicates, but that does not mean that talk of certain
organisms using signs to refer to certain sorts of things and cer-
tain sorts of events should be considered talk of ‘occult’ entities or
properties. Psychologists and anthropologists and sociologists and
linguists have long been investigating sign behav ior, and most of
them do not eschew talk of what signs refer to.27
25. This is approximately the language I used in “The Mental Life of Some Machines,”
in Intentionality, Minds and Perception, ed. Hector-Neri Castañeda (Detroit: Wayne State
University Press, 1967), 177–200; reprinted in Hilary Putnam, Mind, Language and Reality,
vol. 2 of Philosophical Papers (Cambridge: Cambridge, University Press, 1975), 408–428.
26. These are what Ruth Millikan, Language, Thought, and Other Biological Categories
(Cambridge, Mass.: MIT Press, 1984), has called “normal functions” or “proper functions.”
I agree with Millikan on the importance of the notion, but not on the claim that we can
reduce intentional notions such as reference with its aid.
27. Noam Chomsky is an exception. He thinks talk of reference has no place in serious
study of language, and he adds that we can have no intuitions about reference because “the
terms extension, reference, true of, denote, and others related to them are technical inno-
vations which mean exactly what their inventors tell us they mean.” See his New Horizons in
the Study of Language and Mind (Cambridge: Cambridge University Press, 2000), 148.
Unauthenticated
Thus, unlike the functions I spoke of in “The Mental Life of Some

Machines,” which were restricted to ways of processing inputs (and
digitalized ones at that!) from our sensory receptors and sending (dig-
italized) outputs to our “motor organs,” the functions the liberal func-
tionalist will be interested in include ways of perceiving and ways of
dealing with elements of the environment beyond our “surfaces,”
elements as far away as human thought and action can reach. These
are, verily, functions with long arms.
As organisms in a world and a culture, we are able not only to per-
ceive the “quality space” of which Quine speaks in Word and Object, but
also to perceive that, for example, something is a “smart phone” and
that something “feels like a trackpad” and that someone “said that there
was an earthquake in Thailand,” and many other things that require
more and less sophisticated conceptual resources to describe. McDow-
ell’s claim that our senses provide us with forms of “openness to the
world” is right (apart from his confusion of apperceptions with “impres-
sions”) but insufficient; what it needs is supplementation with details
as to how they do that. But this is not a program to be carried out by
empirical investigation alone; as my Doktorvater Hans Reichenbach
taught me, scientific investigation and philosophical analysis need to go
hand in hand. That is as much or more the case in cognitive science as
it is in mathematical physics.
Unauthenticated
D 11
“Naive Realism” and Qualia
I n “ 12 P h i l osoph e r s — and Their Influence on Me,”1 I

described how certain teachers and, later, certain things that I read
and certain philosophical friendships that I formed, had an influence
on the development of my philosophical views. Apart from Rorty, whom
I was very fond of, but whose impact I described as “inspiring me to
refute his account of pragmatism,” all of those influences are described
in positive terms. If that lecture had to be given now, rather than in
2007, a thirteenth philosopher would be added:2 Ned Block. And al-
though John McDowell would still be on the list, I would not write
today, as I did then, that it is “the ‘disjunctivist’ school in the philos-
ophy of perception” that shows us how it is possible to defend what Wil-
liam James called “the natural realism of the common man.” The
reason for both of these (counterfactual) changes is the same: the im-
pact, over a period of a couple of years, of two papers by Ned Block— a
lecture titled “Wittgenstein and Qualia,” which I heard Block deliver
1. “12 Philosophers— and Their Influence on Me,” Dewey Lecture to the Eastern Divi-
sion of the American Philosophical Association, December 2007, Proceedings and Addresses
of the American Philosophical Association 82, no. 2 (2008): 102–115.
2. Lest the reader worry that “13 is an unlucky number,” let me point out that this is a
specifically Christian superstition (arising from the fact that there were 13 at the table at
the Last Supper, according to the New Testament), and neither Block nor I are Christians.
In the Jewish tradition, 13 is the number of the attributes of the deity!
169
Unauthenticated
at the “Putnam Fest” conference in my honor in Dublin in March 2007,3

and a paper of his titled “Consciousness, Accessibility, and the Mesh
between Psychology and Neuroscience,”4 which appeared later in the
same year. In the course of the last few years, these papers have had an
impact on my thinking about the phenomenology of perception com-
parable to the impact on my later philosophy of mathematics that
reading Quine’s “On What There Is” (1948) and “Two Dogmas of Em-
piricism” (1951) in my twenties turned out to have. The present essay
is an attempt to describe that impact.
Is “Sameness” Well-Defined in the Case of Qualia?

Block’s “Wittgenstein and Qualia” did not immediately convince me,
however. One of its theses I had already accepted: namely, that the view
(which I called “externalism” in The Threefold Cord: Mind, Body and
World5) according to which the phenomenological character of a “ve-
ridical” perceptual experience (say, the experience someone has when
seeing a white cat on a blue sofa) can be exhaustively described by de-
scribing the relevant observable properties of the presented scene in the
usual public language— such as the shade of blue of the sofa, the kind of
“white” in the case of the cat—is untenable.6 Indeed, as I mentioned in
the discussion that followed Block’s lecture,7 I have long maintained that
the way the color of something appears to a subject depends on the
properties of the perceiving subject as well as on the properties of
the something in question. As a young man, I noticed that when I would
lie on a beach with one eye shut (as would occasionally happen when
3. The papers read at this conference, including “Wittgenstein and Qualia,” are col-
lected in Reading Putnam, ed. Maria Baghramian (Abingdon, United Kingdom: Routledge,
2011).
4. Ned Block, “Consciousness, Accessibility, and the Mesh between Psychology and
Neuroscience,” Behavioral and Brain Sciences 30 (2007): 481–548.
5. Hilary Putnam, The Threefold Cord: Mind, Body and World (New York: Columbia Uni-
versity Press, 1999), 221n50.
6. In fact, I already argued that “externalism” was untenable in The Threefold Cord: “The
fact that even if color properties . . . are conceived of as properties of ‘external’ objects they
must be admitted to be perspectival, to have dif ferent looks, including looks that depend on
the condition of the perceiving subject, should not be denied (as it has been, unfortunately,
by certain ‘externalist’ phi losophers of mind” (169).
7. Block mentions my remarks in that discussion in the ( later) version of “Wittgenstein
and Qualia” that appears in Baghramian, Reading Putnam, 275–318.
Unauthenticated
“Naive Realism” and Qualia 171
one side of my face was lying on the sand) that the beach looked a slightly
less intense shade of yellow if the left eye was the open eye and if the
right eye was the open eye. But I wouldn’t have said that it seemed to
be yellow when viewed through my left eye alone, and gray when viewed
through my right eye alone. The difference was not that extreme. And
it wouldn’t have affected my “matching” performance on a color chart.
No matter which eye is shut, if the beach matches “yellow32” with the
left eye shut, it will match “yellow32” with the right eye shut. The dif-
ference is “ineffable” in Block’s sense.8 But it is not inexplicable; as
Block mentions, it is easily accounted for by differences in the maculae
of the two eyes. This is not even an “abnormality,” for most people re-
port the same phenomenon if they are asked to make the test.9 And
to ask, “Is the sand really the shade it looks to your left eye or the shade
it looks to your right eye?” is meaningless.
However, at the Dublin conference, I also rejected the idea (which is
central to Block’s defense of “qualia” talk against Wittgensteinians) that
the relation of “sameness of qualitative character” could be fi xed by
finding out which brain-states qualia are (i.e., by showing that a “mind–
brain identity theory” is correct for “qualia”). I said in the discussion,
Now, [Block’s paper] certainly suggests that some straightforward

identification of qualia with certain brain states is going to be dis-
covered. But if there were such an identification, knowing which
one it is is a probably unsolvable epistemological problem.10
And I referred to an argument I gave for this claim in Reason, Truth

and History,11 and I could also have referred to The Threefold Cord: Mind,
Body and World.12 Although that book rejected the “internal realism”
8. By “ineffable” Block does not mean “indescribable in principle” but indescribable in

ordinary language with the terms used to describe objective appearances, for instance,
“such a shade of red.”
9. You do not have to go to the beach or wait for summer; just stand 50 feet or so away
from a light-colored wall (far enough away so that parallax is not significant) and then close
fi rst one eye and then the other. Everyone I know who has tried this has reported a subtle
but noticeable difference in the apparent color of the wall.
10. I also remarked that this is a point I argued for in chapter 4 of Reason, Truth and His-
tory (Cambridge: Cambridge University Press, 1981).
11. Putnam, Reason, Truth and History, chap. 4, 75–102.
12. Putnam, Threefold Cord, 162–169.
Unauthenticated
that I defended in Reason, Truth and History, it also rejected “qualia” talk,
on the grounds that sameness and difference of subjective experiences
is something for which there are only ordinary language criteria (for
example, if a fully competent speaker of the language says her experi-
ence was the same on two occasions, then, other things being equal,
that counts as its being the same). The question as to whether the
experiences of dif ferent people, or even one person at dif ferent times,
are “really” the same, when there is no basis for doubt in the ordinary
use of “same experience” talk, could not, I claimed, be scientifically
investigated. There is no “well-defined” relation of phenomenological
sameness there for science to investigate.
This claim of mine, that qualitative sameness, supposing there is such
a thing, is epistemologically inaccessible was directly rebutted by Block
in the second of the two papers to which I referred earlier.13 It is worth
our while to review the issues at stake. Here is part of what I said in
Dublin after Block delivered “Wittgenstein and Qualia.” (I have already
quoted the first sentence.)
But if there were such an identification [of qualia with brain-states],
knowing which one it is is a probably unsolvable epistemological problem.
I think we should give up the assumption that sameness of qualitative
character is well-defined even for one person at dif ferent times or dif-
ferent people at the same time, let alone for nonconspecifics. I don’t
think that there is a fact of the matter, albeit an ineffable one, about
whether the qualitative character is the same or dif ferent in such cases.
I think that we should give that assumption up. I think that there are
no good candidates in present day neurology for a relation of “identity
of qualitative character.”
If it seems strange that one can be wrong about there being a well-
defined relation of identity here, an example from the history of science
may help. It naturally seems to us that there is a fact of the matter about
what is happening somewhere ten light years from here right now. But
after Special Relativity we’ve had to abandon that natural belief. We have
learned that “right now” is not well defined when astronomical distances
or high relative velocities are involved. Similarly, it may be that “identical
phenomenal quality” is not well defined. Nagel’s famous question, “Is the
13. Namely, Block, “Consciousness, Accessibility, and the Mesh Between Psychology
and Neuroscience.”
Unauthenticated
bat’s phenomenal experience the same as mine, or not?,” may be as mean-

ingless as “What is happening on the sun this second?”
At this point, I gave the following argument in support of the idea
that the supposed relation of “identity of phenomenal character” (of ex-
periences of a subject at dif ferent times) is, if it exists at all, epistemo-
logically inaccessible: Gerald Edelman once invited me to spend a day
in his laboratory. In the course of that fascinating visit, he said that when
we visualize, say, a colored sheet of paper, part of what happens is that the
part of the visual cortex that was active when the original experience
occurred is reactivated. Suppose that the recognition of “patterns”
(thinking of shades of color as species of “patterns”) is modularized.
(Edelman also told me that pattern-recognition “modules” tend to
“recruit” additional neurons, if they are frequently stimulated, and to
lose neurons to adjacent modules if unused compared to the adjacent
modules.) Consider the speculation that the slightest change, say the loss
or gain of a single neuron, “changes the quale.” Could the subject tell?
In Philosophical Investigations, Wittgenstein wrote, “Always get rid of
the idea of a private object in this way: assume that it constantly changes
but that you do not notice the changes because your memory constantly
deceives you.”14 What Edelman told me suggested to me that if the
changes Edelman described did cause a change in the “qualia” that we
(supposedly) have when a given module is stimulated, we would never
know because when we tried to remember how the sheet of paper (or what-
ever) looked before the neurons were added or subtracted from the module the
memory image would seem as if there were no change in the “qualia”; mem-
ories would change as the correlated qualia changed. And this shows at
the very least that the supposed “sameness / difference” of qualia is epis-
temologically inaccessible, and perhaps that the very idea that there is
such a relation, a relation of really being the same (as distinct from the
relation of “seeming the same as far as I can tell now by relying on my
memory and my mastery of the language”) is as much of a mistake as
supposing that there is such a relation of “really happening this very
second on the sun” (as opposed to happening this second in a given ref-
erence frame). Wittgenstein’s and Edelman’s remarks seemed to fit
together like hand and glove.
14. Ludwig Wittgenstein, Philosophical Investigations, trans. R. Rhees and G. E. M. Ans-

combe (Oxford: Blackwell, 1953), II, p. 207.
Unauthenticated
Problems with My Argument against Qualia

Obviously, I took a remark of Edelman’s and ran with it, and I prob-
ably ran farther than he would approve. To conclude from the experi-
ment with a colored sheet of paper he described that we can never re-
member what our experiences were like prior to changes in the visual
cortex, as I did in that discussion, was a big inferential leap. If someone
becomes red–green color blind as a result of brain damage, is it really
the case that she will not be able to remember what red and green ob-
jects used to look like? That’s certainly not something I know! But there
were fatal problems with my argument beyond the uncertainty of the
empirical premise it assumed.
Even if its premise were correct, that would only show that if a sub-
ject’s spectrum became “shifted” as a result of changes in her visual cor-
tex,15 she would not be aware of it, and hence others could not confirm
the shift on the basis of her reports. In “Consciousness, Accessibility, and
the Mesh between Psychology and Neuroscience,” Block argues that
identification of the brain-parameters responsible for “qualia” need not
depend simply on subjects’ reports. In general, mind–brain connections
have to be confirmed by arguments from the consilience of a hypothesis
with different kinds of evidence rather than from direct “tests” (a situa-
tion familiar from present-day physics, for example), as well as from a
comparison of the explanatory power (or lack thereof) of various alter-
native hypotheses, as is common in other sciences. In “Consciousness,
Accessibility, and the Mesh between Psychology and Neuroscience,”
Block outlines how such an approach is likely to go in the case of visual
phenomenology. On studying that paper I was completely convinced.
As a by-product of reading that paper I was also enabled to see what
was wrong with a key argument in J. M. Hinton’s classic paper “Visual
Experiences”16 (the foundational paper for what is today called “dis-
15. Of course, “shifts” due to insertion of color-changing lenses, or changes in the pig-
mentation of the maculae, or changes in the optic nerves exterior to the brain itself could
still be detectable if changes due to changes in the relevant “modules” themselves were not.
But while the possibility of shifts whose causes are (literally) “in the head” but not in the
brain itself is enough to show that pure “externalism” with respect to color-qualia is unten-
able, it does not seem to me that they are fatal for Wittgenstein’s skepticism about “the
private object,” or my more empirically grounded attempt to defend it.
16. J. M. Hinton, “Visual Experiences,” Mind 76 (1967): 217–227.
Unauthenticated
junctivism”) that I had long found fishy. Hinton scoffs at the idea that
someone has “experiences” that are the same when she has a hallucina-
tion and when she has a corresponding (indistinguishable by her)
veridical experience. Call such a common factor her “psi-ing” on both
occasions. We could not find out what “psi-ing” is by empirical inves-
tigation, Hinton argues, because we would fi rst have to be clear on
what psi-ing is; his picture is clearly that an empirical identity, say, “As
are really Bs” can only be confirmed if one is totally clear in advance on
what one means by “As.” Here is the passage:
If there is no such thing as (Q) [the supposed common factor—HP]

then there is—of course—no such thing as my psi-ing for the fol-
lowing statements to be about. (i) My psi-ing is one and the same
event as some happening that is describable in the language of
physics and / or physiology including neurophysiology.17
The problem with Hinton’s argument, in a nutshell, is that when an

identity (“As are really Bs”) is a matter of theory (say, “Gravitational
attraction is really a ‘tidal force’ [a deformation of space-time]”) rather
than a simple observational report such as “that man is my son,” be-
coming clear on what A means (what we mean by “gravitational attrac-
tion”) and confirming the claimed identity as the best explanation (via
the consilience of various lines of data, the fact that it has the highest
explanatory power of all the proposed explanations, and the like) pro-
ceed simulta neously; one does not come before the other. If qualia
are successfully identified with brain-states, then we can expect to
become clearer on, not exactly what we “mean” by “qualia,” but on
what we ought to mean by the term. (What we mean today by “water” is
not, in fact, exactly what we meant before we discovered the chemistry
of water. For Locke, ice and water were two dif ferent substances, albeit
one frequently turned into the other.) In 1967, Oxford “ordinary lan-
guage philosophy” was still riding high; the argument of Hinton’s that
I just quoted represents the weak side of that proudly scientifically un-
sophisticated philosophical school.
17. Hinton, “Visual Experiences,” 219.
Unauthenticated
“Naive” Realism
The idea that the mass of mankind is quite mistaken in thinking
that the colors, textures, and other “secondary properties” of the ob-
jects they claim to perceive are “out there,” in the things themselves,
when they are really just our own “sense data” is an old one.18
For Russell it was quite clear when he wrote The Problems of Philos-
ophy in 1912 (and still clear in 1940, even after a flirtation with direct
realism in The Analysis of Mind),19 that we do not observe physical ob-
jects at all. He wrote,
When, in ordinary life, we speak of the color of the table, we only

mean the sort of color which it will seem to have to a normal spec-
tator from an ordinary point of view under usual conditions of
light. But the other colors which appear under other conditions
have just as good a right to be considered real; and therefore, to
avoid favoritism, we are compelled to deny that, in itself, the table
has any one particular color.20
About ten pages later, Russell has shown to his satisfaction that the table
we see lies in our own private visual space, and the table we feel lies in
our own private touch space, and so on.21 And furthermore, we do not
perceive the external, physical, table at all:
My knowledge of the table as a physical object, on the contrary, is

not direct knowledge. Such as it is, it is obtained through acquain-
tance with the sense data that make up the appearance of the
table . . . There is no state of mind in which we are directly aware
of the table; all our knowledge of the table is really knowledge of
truths, and the actual thing which is the table is not, strictly
speaking, known to us at all.22
18. See book I, section II, part IV of Hume’s Treatise of Human Nature, for example.
19. Bertrand Russell, The Analysis of Mind (New York: Macmillan, 1921).
20. See Bertrand Russell, The Problems of Philosophy (Oxford: Oxford University Press,
1912), 10.
21. Russell, Problems of Philosophy, 29.
Unauthenticated
Finally, our knowledge of the table must be via inference:
Thus it becomes evident that the real table, if there is one, is

not the same as what we immediately experience by sight or
touch or hearing. The real table, if there is one, is not immedi-
ately known to us at all, but must be an inference from what is
immediately known. Hence, two very difficult questions at once
arise; namely, (1) Is there a real table at all? (2) If so, what sort of
object can it be?23
Nor is the Russell view wholly dated. Although philosophers who will
say that what we perceive are our own sense data are no longer thick
on the ground, there are respected contemporary philosophers who
deny that colors are “real” (Larry Hardin, for example),24 and if colors
are not real, then much of what we take ourselves to see is not there to
be seen!
Moreover, a search of the Internet under “naive realism” will reveal
that quick “refutations” of naive realism, many of them two or three
centuries old, are all over the place. For example,
If naive realism is to be taken seriously, and colors are out there in

the world, then apples regularly change color depending on how
much light is around them. It is much more plausible, though, to
think that the apples are the same as they ever were, that all that
has changed is our experience of them.25
This refutation assumes that the naive realist cannot maintain that
dispositions to have certain appearances to human beings cannot be “out
there in the world.” In fact, even naive realists would include disposi-
tions to have certain a certain look depending on the lighting conditions
and the position of the viewer among the objective properties of the
objects viewed.

24. C. L. Hardin, Color for Philosophers: Unweaving the Rainbow (Indianapolis: Hackett,
1988).
25. See Tim Holt, “Naive Realism,” Theory of Knowledge, n.d., www.theoryofknowledge
.info/naiverealism.html.
Unauthenticated
Yet today there are philosophers who are proud to call themselves
naive realists.26 Obviously, this new use represents a reaction against
views like Russell’s. And clearly the “externalism” that I argued against
earlier represents a sort of return to naive realism, as does my own call
in The Threefold Cord for a “second naiveté” in the philosophy of per-
ception. But rather than try to survey all the contemporary views that
represent (more or less) a defense of naive realism, what I shall do for a
moment is look at Russell’s position and see at what points a defender
of any degree of return to naive realism might reject it.
Russell’s Position Again

In the previous quotations from The Problems of Philosophy, the following
propositions are obviously assumed to be true:
(1) We perceive our sense data. (But the man and woman on the
street mistakenly think that what they are perceiving are
physical things.)
(2) We don’t actually perceive physical objects.
(3) What we know about physical objects we know by inference.

The premises for those inferences come from observing our
own sense data.
Some Ways of Rejecting Russell’s Picture

It is easy to find grounds for rejecting one, two, or all three of these
Russellian propositions, but, as we shall see, some of the ways of re-
jecting them hardly do justice to what William James called the “natural
realism” of the ordinary person. Many of those ways amount to little
more than an appeal to ordinary usage—not a bad thing when it is just
one of a number of points made in a philosophical discussion but hardly
a satisfactory response by itself to a deeply thought out and complex
metaphysical cum epistemological position such as Russell’s.
For example, one might reject proposition 1 by saying, “We don’t
perceive visual ‘sense data,’ we have them.” And one might then reject
26. For instance, “disjunctivists” in the philosophy of perception. See, for example,
M. G. F. Martin, “On Being Alienated,” in Perceptual Experience, ed. Tamar Szabo Gendler
and John Hawthorne (Oxford: Clarendon Press, 2006), 354–410.
Unauthenticated
proposition 2 by saying, “When we have the appropriate visual sense

data, and we unconsciously and automatically infer that there is a table
in front of us, then that is what we ordinarily call ‘seeing a table.’ ” Ob-
viously, this is not a rejection of Russell’s metaphysical picture at all but
simply a recommendation that Russellians find a way of expressing their
view that is more charitable to ordinary ways of speaking. However,
propositions 2 and 3 would be rejected by some philosophers on more
interest ing (and controversial) grounds: a widely held view today is that
the justification of observational beliefs, at least about such fundamental
sensible properties as color, shape, texture, hardness, and the like, is
simply a “reliabilist” one; we have been programmed by evolution so
that when any of those beliefs is caused by impacts on our sense organs
(an event that reliabilists identify with perceiving the objects, when the
causal chain is of the right sort), the belief is very likely to be correct.
This “reliabilist” ground for rejecting propositions 2 and 3, unlike the
“ordinary language” grounds for rejecting propositions 1 and 2, does
involve a serious disagreement with Russell’s metaphysics; today’s nat-
uralistic reliabilists (rightly in my view) see perception as beginning
with external things, and transactions involving both those things and
the organism, and not with “sense data”; but it would theoretically be
possible to produce a version of reliabilism more friendly to at least part
of Russell’s picture. I can imagine a philosopher who simply replaces
“caused in the right way by impacts on our sense organs” with “caused
in the right way by appropriate sense data,” leaving it an open empir-
ical question whether the “right sort” of causation is inference (perhaps
unconscious inference, as Helmholtz believed).27
Another common ground for rejecting propositions 1, 2, and 3 is the
so-called transparency of perceptual experience. Here is how Block,
who accepts the view as a description of part of the phenomenology of
perception, 28 describes it in “Wittgenstein and Qualia”:
Shoemaker’s view is shared by Fred Dretske, Gilbert Harman,

Michael Tye and many others who advocate what G. E. Moore
termed the diaphanousness (or sometimes the transparency) of
27. H. L. Helmholtz, Handbuch der physiologischen Optik, vol. 3 (Leipzig: L. Voss, 1866);
3rd ed., repr. with extensive commentary by Allvar Gullstrand, Johannes von Kries, and
Wilibald Nagel (Hamburg and Leipzig: L. Voss, 1896), 28–29.
28. However, Block rejects the idea that we are unable to ever attend to our own qualia.
Unauthenticated
experience. Harman puts the point by saying that the more one
tries to attend to one’s experience of the tree, the more one attends
to the real tree instead.29 Although Moore is sometimes cited as the
originator of this point, he did not actually accept it. I have
heard him quoted saying “the moment we try to fi x our attention
upon consciousness and to see what, distinctly, it is, it seems to
vanish: it seems as if we had before us a mere emptiness. When
we try to introspect the sensation of blue, all we can see is the blue;
the other element is as if it were diaphanous.” But these words are
followed by what I regard as a more significant truth: “Yet it can
be distinguished if we look attentively enough, and know that there
is something to look for.30
Any inference from it seems to us that we attend to “the real tree” to “we
actually perceive the real tree” would, of course, be questioned by Rus-
sell, and with reason. The “diaphanousness” of (much of)31 perceptual
experience is simply a fact about the phenomenology of perceptual ex-
perience, but accepting it as such is not sufficient to overthrow or even
present a clear alternative to Russell’s picture.
Externalism as a Break with Russell’s Internalist Picture

Theories like Russell’s are internalist theories in the sense that (1) the
inputs to perception, “the sense data,” like Hume’s “ideas and impres-
sions,” are entirely inside the subject’s mind, and all the cognitive pro-
cessing takes place inside the mind, where (2) the mind is conceived of
as either inside the brain (by materialists like Hans Reichenbach)32 or
as totally immaterial (by Berkeley, and, arguably, by Hume as well; Rus-
sell’s view seemed to have wavered between materialism and immateri-
alism over the years). Such views represent what I called an “interface
conception” in The Threefold Cord, because they allow us no truly
29. G. M. Harman, “The Intrinsic Quality of Experience,” Philosophical Perspectives 4

(1990): 31–52.
30. Block, “Wittgenstein and Qualia,” 280.
31. I say “much of” because I agree with Block that we can and sometimes do attend to
our qualia as such.
32. According to Reichenbach, “impressions” are our own internal states. See Hans
Reichenbach, Experience and Prediction (Chicago: University of Chicago Press, 1938), 172.
Unauthenticated
cognitive access to the world, at any rate to the world outside the brain.
But there is no good reason for a present-day naturalist philosophy of
mind to be internalist.
In a way, this is something I already argued in “The Meaning of
‘Meaning’.”33 A central thesis of that essay was that we need an exter-
nalist and anti-individualist account of what it is to understand the
words of a natu ral language.34 To think about gold, or water, or grass,
or most of the things we think and talk about, is to engage in an activity
that presupposes complex interactions with our environment and with
other people. In “The Meaning of ‘Meaning’,” I expressed this by saying
that “Meanings aren’t in the head,” but as a number of people later
pointed out, I should have said “The mind isn’t in the head.” It is not, I
hasten to add, that “the mind” is somewhere else; the mind isn’t a thing
with a location at all (so it is not simply the brain under another name),
but a system of world-involving abilities and exercises of those activities.
This view is, I believe, now accepted by a majority of philosophers of
mind and cognitive scientists.
This is the view of the mind I have had for a long time, but in the
last few years I have realized that it also captures what was right in
“functionalism.” It is true that I originally gave the name functionalism
to an internalist view; in “The Nature of Mental States,”35 the mind
was identified with brain, which was described as a computer, and our
mental states were identified with computational states of that computer.
That was wrong,36 but the idea that mental capacities and activities are
ways of functioning was right, provided we allow that those ways of func-
tioning may involve the environment and other people, and provided
33. Hilary Putnam, “The Meaning of ‘Meaning’,” in Language, Mind and Knowledge, ed.
Keith Gunderson, Minnesota Studies in the Philosophy of Science, vol. 7 (Minneapolis: Univer-
sity of Minnesota Press, 1975), 131–193; collected in Hilary Putnam, Mind, Language and Re-
ality, vol. 2 of Philosophical Papers (Cambridge: Cambridge University Press, 1975), 215–271.
34. By an “externalist” account of understanding, I mean here one that makes essential
reference to things in the environment, as the account of meaning in “The Meaning of
‘Meaning’ ” did, not one that insists that the phenomenological character of a “veridical”
perceptual experience can be exhaustively described by describing the relevant observable
properties of the presented scene.
35. Hilary Putnam, “The Nature of Mental States,” in Mind, Language and Reality,
429–440.
36. I explain why I say it was wrong in Hilary Putnam, Representation and Reality (Cam-
bridge, Mass.: MIT Press, 1988).
Unauthenticated
we do not limit the language used for the description of those ways to
the language of computer science. For that reason, in a lecture I gave
not long ago,37 I referred to my present position as “liberal function-
alism,” and to functional “states” in this liberal sense as having “long
arms” (arms that reach out to the environment). In this terminology,
“The Meaning of ‘Meaning’ ” is a liberal functionalist view of cogni-
tion, or an “externalist” view if we do not give “externalist” the meaning
I gave it in The Threefold Cord—the meaning according to which an “ex-
ternalist” view of perception requires us to hold that the phenomeno-
logical character of a veridical perceptual experience can be exhaustively
described by describing the relevant observable properties of the pre-
sented scene in the usual public language—but simply take an “exter-
nalist” view to be one that allows at least some of our mental states to
be at least partly externalistically identified.
I said earlier that an externalist conception of thinking (in this sense
of “externalist”) is now widely accepted in cognitive science and phi-
losophy of mind. But once this has become the case, as it obviously had
not in Russell’s day, nothing stands in the way of extending externalism
to perceptual states as well.38 But now a problem arises.
The Problem
From here on, I shall discuss McDowell’s views in Mind and World as
an example of a position diametrically opposed to Block’s and to the
whole idea of qualia, that is, of nonconceptual phenomenal characters
common to both veridical and nonveridical experiences.39 To see the
problem to which I just alluded, let us recall that McDowell motivates
37. Hilary Putnam, “Corresponding with Real ity,” in Philosophy in an Age of Science:
Physics, Mathematics, and Skepticism, ed. Mario De Caro and David Macarthur (Cambridge,
38. A great deal in this direction has been done by Tyler Burge in Origins of Objectivity
(Oxford: Oxford University Press, 2010).
39. Block himself usually takes “representationalists,” particularly “intentionalists” like
Tye, as his chosen targets; however, these philosophers are not in total disagreement with
Block because they do believe veridical and hallucinatory experiences (to take the extreme
case of the nonveridical) have a common phenomenal character, although they identify that
phenomenal character with information; disjunctivists such as McDowell are in total dis-
agreement with Block’s views because they deny both that sensory impressions are noncon-
ceptual and that there is such a “highest common factor.”
Unauthenticated
his complex chain of arguments in Mind and World by laying down two
requirements for a satisfactory philosophy of perception. The fi rst,
which he calls “minimal empiricism,” is that sensory impressions must
be a “tribunal” before which our beliefs about the world can stand,40
and much of the controversy connected with that book has to do with
McDowell’s claim that this requirement can only be fulfilled if those
impressions are themselves conceptualized.41 The second (which is sup-
posed to follow from the fi rst) is that “reliabilism” must be rejected,
because in the reliabilist view, McDowell charges, experience only
“exculpates” the subject from criticism for having the beliefs her im-
pressions cause her to have, but fails to justify those beliefs. This is the
central argument of Mind and World: sense impressions (the “tribunal”
before which all our beliefs have to stand) can justify beliefs only if those
sense impressions are themselves justified.
In opposition to McDowell’s views, Hilla Jacobson and I have argued,
on both empirical and conceptual grounds, that the phenomenal char-
acters of perceptual experiences are not (in any case, not always)
conceptualized in any of the senses McDowell has proposed;42 hence,
if McDowell is right, they cannot be a “tribunal,” whether or not those
phenomenal characters are identical with qualia. So, if McDowell is
right, skepticism threatens us. Moreover, qualia, Block taught us, are
brain states, and hence internalistically identified. We might, of course,
say that having (appropriate) qualia only counts as perceiving something
when they are caused in the right way; but didn’t I criticize that idea
previously when I wrote that to say, “When we have the appropriate
40. Minimal empiricism, which McDowell endorses, “makes out that the very idea of
thought’s directedness at the empirical world is intelligible only in terms of answerability
to the tribunal of experience, conceived of in terms of the worlds impressing itself on per-
ceiving subjects.” Quote from John McDowell, Mind and World (Cambridge, Mass.: Har-
vard University Press, 1996), xvi. Note the (nonaccidental) similarity of “impressing itself”
and “impressions.”
41. McDowell generally uses “impressions” and “experiences” interchangeably (he also
sometimes uses the Kantian term “intuitions,” particularly in the essays collected as Having
the World in View; Essays on Kant, Hegel, and Sellars (Cambridge, Mass.: Harvard University
Press, 2009). In Mind and World, he identifies “intuitions” with “bits of experiential intake”
(pp. 4 and 6) and “impressions” with “impacts of the world on our senses” (139).
42. Hilla Jacobson and Hilary Putnam, “Against Perceptual Conceptualism,” paper read
(by Hilla Jacobson) at the conference “Philosophy in an Age of Science: Conference in
Honor of Hilary Putnam’s 85th Birthday,” Harvard and Brandeis Universities, June 30–
July 4, 2011; forthcoming in the International Journal of Philosophical Studies.
Unauthenticated
visual sense data, and we unconsciously and automatically infer that

there is a table in front of us, then that is what we ordinarily call ‘seeing
a table’ ” is “not a rejection of Russell’s metaphysical picture at all, but
simply a recommendation that Russellians find a way of expressing their
view that is more charitable to ordinary ways of speaking?” So it looks
as if we have both handed game, set, and match to the skeptic and fallen
back into Russell’s picture.
My present view is almost the complete opposite of McDowell’s. (1)
Where McDowell requires that “impressions” must be a tribunal, I ar-
gued in Chapter 10 that it is apperceptions (and seeming-apperceptions)
that are the tribunal, and not impressions.43 Some apperceptions and
seeming-apperceptions do, indeed, include the occurrence of qualia (or
“impressions,” to use McDowell’s term, but without assuming the meta-
physics that goes with it), but such cognitive states are not the same as
qualia, nor, as we argued in Chapter 10, are they simply beliefs “trig-
gered” by qualia.44 The idea that “impressions” are the ultimate source
of the confirmation and disconfirmation of empirical beliefs is a hang-
over from empiricism that should be jettisoned.
(2) Unlike McDowell, I don’t agree that reliabilism must be com-
pletely rejected, as I will explain in the next section. That does not
mean that we should accept the view I rejected above, according to
which perceiving (for example) a rabbit on one’s lawn is just having the
appropriate visual “sense data” and unconsciously and automatically in-
ferring that there is a rabbit on the lawn in front of one. But this issue
deserves a section to itself.
A Relevant Objection to Reliabilism

The version of reliabilism that McDowell considers in Mind and World
is one on which impressions cause beliefs in a way that McDowell
43. See my “Comments on Travis and McDowell,” in Baghramian, Reading Putnam,

347–358. McDowell would deny that his “sense impressions” are what Block and I call
“qualia,” but the difference he fi nds depends on the controversial claims that (1) impres-
sions are always conceptualized; and (2) impressions are not a common factor in both verid-
ical perceptual experiences and hallucinations (even “perfect” hallucinations, in which the
brain is in the same state as it is when a veridical perception takes place). I believe that both
these claims are untenable.
44. Although awareness of the role of apperception, and its difference from simple sense
impressions and from belief, figures largely in the thought of Leibniz and Kant, it simply
goes missing in empiricism. That is why talk of “minimal empiricism” is so misleading!
Unauthenticated
associates with “bald naturalism”—that is, a way that is just a matter of

the operation of subpersonal mechanisms, and thus wholly outside “the
space of reasons.” I agree with McDowell that we want something more
than just an account of subpersonal mechanisms. Moreover, Russell’s
claim that we know about his table (or the rabbit on my lawn) via in-
ference is simply unbelievable, if the inference is supposed to be con-
scious; if it is supposed to be “unconscious inference,” as Helmholtz
proposed, then what we have is, again, just a hypothesis about subper-
sonal mechanisms in the brain (at best). Indeed, the hypothesis is not
particularly convincing for at least two reasons. (1) Subpersonal mech-
anisms may perform syntactic operations that we (at the personal level)
interpret as inferences, such as “writing” (in “Mentalese”) both “A” and
“B” given “A&B,” but “writing” one or two formulas given a third
formula as “input” is not inference, it is only a syntactic representation
of an inference. (2) Now that we know about forms of computation
that do not consist of inferences in the traditional sense at all, there are
many more possibilities for modeling the subpersonal processes in-
volved in thought and perception than Helmholtz could have thought
of. But we do not need to speculate about this; it is enough to see that
a story about visual qualia causing beliefs via unconscious mechanisms,
however interest ing to investigate, is not an account of perception at
the level we seek. About that McDowell is right. This is a relevant ob-
jection to one form that “reliabilism” can take.
Apperception is a phenomenon at the psychological level, the level
of rational agency,45 and there is no reason why there shouldn’t be an
account of the perceptual transactions of human beings (and other or-
ganisms) with their environments at that level. In the human case, such
an account will involve also an account of language acquisition and of
the role of our linguistic abilities in apperception. A “liberal function-
alist,” in my sense, can agree with McDowell that conceptualization
plays an important role in perception, in a demanding sense of percep-
tion. McDowell’s mistake is to assimilate all experience to perception in
the demanding sense.
45. My identification here of the “psychological level” with the level of “rational agency”
is something I now [ July 2015] regard as a mistake. Representation, in Burge’s sense dis-
cussed in Chapter 1, is a psychological activity (not a subpersonal one)—representing some-
thing involves the thing represented, and is thus an externalistically identifi ed psychological
process— but much more primitive than “rational agency,” which requires conceptualiza-
tion. [Note added on July 5, 2015.]
Unauthenticated
But what of McDowell’s claim that any form of reliabilism only “ex-
culpates” the subject from criticism for having the beliefs her sense im-
pressions cause her to have, but fails to justify those beliefs? What do
we do about skepticism?
Two Unsound Objections to Reliabilism

(1) Cartesian Skepticism Unanswerable
James Conant has distinguished between two kinds of skepticism,
which he calls “Cartesian skepticism” and “Kantian skepticism.”46
“Cartesian skepticism,” in Conant’s sense, is skepticism about the pos-
sibility of knowing anything about an “external world”; “Kantian skep-
ticism” is a worry about how it is possible that our thoughts, whether
supposedly about an external world or even about our own sense impres-
sions, really have content at all. (Note that this is not supposed to be an
epistemological question.)
Through the years, both Conant and McDowell have taken me to task
because, as they see it, in The Threefold Cord I mistook the target of Mind
and World to be the “Cartesian” variety of skepticism, whereas it was actu-
ally the “Kantian” variety that McDowell was concerned to exorcise.47 I
know that I was not alone in missing this important distinction, because a
criticism of Mind and World that I have often heard from students (though
not one that I ever made) is that it doesn’t “answer” the (Cartesian) skeptic.
If this objection applied to McDowell, it would apply even more strongly
to the reliabilist, since the latter does not seek to answer the Cartesian
(or “Humean”) skeptic at all. But the criticism seems to me misguided.
It is misguided because the notion of “answering the skeptic” suffers
from a fatal ambiguity. If “answer the skeptic” means give the skeptic a
proof that we do know, for example, that there are such things as fireplaces
and chairs in front of them that we sit on (Descartes’s example, slightly
modified) which the skeptic herself must accept, that is, a proof from prem-
ises she must grant, then that is evidently impossible. We do fi nd that
46. James Conant, “Varieties of Skepticism,” in Wittgenstein and Skepticism, ed. Denis
Mc Manus (Abingdon, United Kingdom: Routledge Press, 2004), 97–136.
47. See John McDowell, “Responses,” in Reading McDowell: On Mind and World, ed.
Nicholas Smith (London: Routledge, 1994), 267–305.
Unauthenticated
disturbing, but as Quine famously wrote, “The Humean predicament

is the human predicament.”48 The question we nonskeptics should worry
about is, Is there a proof from premises we must accept that we do not
know any such things? And I have argued elsewhere that the answer is
that there is no such proof, although, to be sure, there is a work of ra-
tional reconstruction of our talk about knowledge, justification, and the
like that has to go with that answer.49
A more legitimate reason for worry about reliabilism is that as an
account of our uses of the verb “to know” in general it is open to many
difficulties. In particular, reliabilists seem driven to either deny the
conceptual connections between knowledge and justification alto-
gether, or else to give implausible accounts of justification. But that
does not show that when we come to the question of our knowledge of
such elementary perceivable facts as the fact that we see something red
in front of us or something square in front of us we really have more to
say than “I see it.” And “I see it” counts, in most cases, as justification
enough simply because we have been evolved to be highly reliable in
such cases and with respect to such simple properties at least.50
That reliabilism may be right in such basic cases does not mean that
it is right in all cases and about all knowledge / justification claims, of
course. And this much reliabilism does not commit us to any confu-
sion of the subpersonal and the personal levels of explanation because
it is not a story about how our recognition abilities work in such cases,
or about the nature and the interrelations of concepts and experiences
at work in such cases. But that remark clearly does not address the Kan-
tian skeptic, who is McDowell’s concern.
(2) Kantian Skepticism Unanswerable

The problem of what to say about Kantian skepticism is quite dif ferent,
however, as McDowell rightly emphasizes. First of all, there aren’t
48. W. V. Quine, “Epistemology Naturalized,” in Ontological Relativity and Other Essays

(New York: Columbia University Press, 1969), 69–90.
49. See the papers collected as section V, “The Problem and Pathos of Skepticism,” in
De Caro and Macarthur, Philosophy in an Age of Science.
50. Note that I did not say that “the putative reliability of the process of seeing” is the
justification / the subject’s reason for believing that p; the justification is “I see that p” Pe-
riod. But we would not count it as a justification if it were not reliable.
Unauthenticated
“Kantian” skeptics in the way there are Cartesian (or Humean, or Pyr-
rhonian) skeptics. It would be obviously self-refuting to claim that “no
thought has any content at all, including this one.” What Kant is asking
of us is not to refute this claim, since it refutes itself, but to answer a “how
is it possible?” question: How is it possible that thoughts possess content?
And that is a task for all philosophers as well as for psychologists—
indeed, for all who theorize about the mind.
That there is a real task here is not an objection to reliabilism, how-
ever, because the reliabilism I have endorsed (call it minimal reliabilism)
does not purport to address it. If it seems to McDowell to be an objec-
tion to reliabilism that it offers “exculpations” where “justifications” are
called for, that is because it does not do what his own answer to the “how
can thoughts possess content?” question allegedly does.51 But I fi nd
that answer unacceptable if understood, as I think it is intended to be
understood. Strangely, because it is a vital part of McDowell’s call for
“minimal empiricism,” I have been unable to find any real discussion
of what he says about content in the literature. The connection to
Block’s concerns (lest the reader think I have forgotten that this is an
essay about Ned Block!) is that McDowell seems to be offering a “tran-
scendental argument” that qualia, if there are such things, must be
conceptualized if our thoughts are to have content at all, a thesis that goes
against much of what Block claims.
I just said that “I have been unable to fi nd any real discussion of
[McDowell’s account of content] in the literature”; in part, this no
doubt due to the fact that McDowell presents that account in very few
pages,52 and neither in Mind and World nor in Having the World in View
does he expand on that presentation in any great detail. Indeed, it may
be that I misunderstood what McDowell intends, but what follows is
how I understand him.
(1) McDowell is a realist (as am I, and as is Ned Block). He criticizes
Sellars for denying the reality of the objects of our commonsense world-
picture,53 and, although I could only fi nd one remark to this effect in
Having the World in View (and none that I recall in Mind and World ), he
51. See McDowell, Mind and World, 11–12. McDowell does not use the term “reliabi-
lism,” but the position he ascribes to Davidson is a version of the reliabilism with respect to
perceptual judgments I describe.
52. See especially McDowell, Mind and World, 3–13.
53. McDowell, “The Logical Form of an Intuition,” in Having the World in View, 41–42.
Unauthenticated
regards the unobservables of modern science as fully real.54 However,

his account of how reference is fi xed is primarily an account of refer-
ence to things we can perceive with our human senses.
(2) Like Sellars (and, again, this is something I agree with, and I am
sure Block agrees with), he accepts the fundamental Wittgensteinian
(and pragmatist) insight that grasping concepts that refer to observ-
able things and acquiring a system of beliefs about those things, and
ways of acting on them and interacting with them, are interdepen-
dent abilities. That is why the mere ability to discriminate (respond
differentially to), say, squares and nonsquares, an ability we share with
properly conditioned mice, does not suffice to have the concept square.
Because McDowell believes impressions are conceptualized, he con-
cludes that animals without language (like the mice) cannot have the
sort of sensory impressions we have. When we see a square, our sensory
impression, as it were, “speaks” to us, and its speech employs the concept
square.55 It follows that we cannot even imagine the sensory impressions
of a mouse or, for that matter, a prelinguistic infant.56
(3) However, once we have acquired the linguistic capacities that are
required to have impressions that are not “mute,” our impressions
present to us (in favorable cases) facts about the things that we look at,
touch, listen to, as they are. Our impressions are not representations
of the goings-on around us; we are directly aware of the goings-on them-
selves. The “transparency” of perceptual experience is a real transpar-
ency, not just a point about how it seems to us. (In this re spect, the
metaphor of impressions “speaking” to us is not the best figure for
McDowell’s purposes; “intentionalists,” who identify phenomenal
character with “information,” also believe our impressions speak to us,
but they speak to us even when they are nonveridical.) For McDowell,
as a “disjunctivist,” nonveridical impressions aren’t impressions at all, at
54. In his Second Woodbridge Lecture, McDowell writes, “Not, of course, that we
cannot direct thought at objects that we are unable to bring into view, perhaps because they
are too small or too far away.” And he adds, “But thought so directed is carried to its object,
so to speak, by theory” (Having the World in View, 37). He does not discuss reference to
such objects as fi elds or quantum mechanical particles or wave functions, which are not the
sort of thing that could be “viewed” if they were larger or closer to us.
55. I recall that somewhere— I cannot now fi nd the place— McDowell writes that in-
stead of saying “percepts without concepts are blind” he should have written that “percepts
without concepts are mute.”
56. I criticize this view in Putnam, “Comments on Travis and McDowell,” 347–358.
Unauthenticated
least not in the case of complete nonveridicality (i.e., complete halluci-

nation). It is the fact that in veridical perception we directly perceive
the world itself that is supposed to make it non-mysterious that we
can refer to things and events in it, and hence non-mysterious that
thoughts about observable things and events have content.
(4) At times, however, McDowell’s view seems to amount to the sort
of liberalized verificationism one finds in logical positivism after 1939:57
statements about both observables and unobservables, including
“guesses,” have content as long as and only as long as they are confirmable
by perceptual experiences (which, in the light of the previous point, means
not by private sense data but by events a human could observe). As I
already mentioned in chapter 10, in Mind and World McDowell writes,
“Empirical judgments in general—whether or not they are justified at
all, perhaps less substantially than knowledge requires—had better have
content of a sort that admits of empirical justifi cation, even if there is none in
the present case (say in a quite unsupported guess).”58 Since this occurs in
a discussion of the temptation to succumb to “the myth of the Given,”
perhaps it is not McDowell’s own view, although it does look as if it is
something that McDowell concedes the philosophers who succumb to
the myth. At any rate, this is not a premise in the argument for the
Given that McDowell ever criticizes.
Point 3 [and point 4, if it is McDowell’s view] constitute McDowell’s
answer to the “how do thoughts possess content?” question, and as I
indicated earlier, I find this answer unacceptable. I fact, I think both
points 3 and 4 are wrong.
Point 3 is wrong because it is not the case that when we perceive, say,
that something is red, or that something “looks red” in the objective
sense, the objective color, or the objective “looking red” [or whatever
color is in question], is present as the phenomenal character of our “im-
pression,” as McDowell’s account demands. It is not the case because
the phenomenal character of the “impression” varies from person to
person, as the data cited by Block show (and as my own visual experi-
57. After 1939, the logical positivists (or “logical empiricists” as they by then preferred
to style themselves) abandoned sense-datum epistemology in favor of “thing language,”
and they also recognized that theoretical terms could not be “reduced” to observation
terms. I give a brief account of this history in Putnam, The Collapse of the Fact / Value Di-
chotomy (Cambridge, Mass.: Harvard University Press, 2002), 7–27.
58. McDowell, Mind and World, 6 (emphasis added).
Unauthenticated
ence, described earlier in the essay, also shows). The strong form of
transparency that McDowell’s account needs is just not there; if this is
“naive realism,” it is too naive.
Point 4 is wrong because it is incompatible with what I take to be a
proper scientific realism.59 If we can refer to things outside our light
cone at all, as I take it we obviously can, it is not only “guesses” that
“admit of empirical justification” that have content, but so do conjec-
tures that it is impossible to justify that have content—for example, the
conjecture that “there are no intelligent extraterrestrials anywhere in
space-time.”60
With respect to how our thought can reach to the unobservable, I
could find only the single remark (in the Second Woodbridge Lecture)
that “thought so directed is carried to its object, so to speak, by theory.”
In sum, to the extent that we can understand “how is it possible that
thoughts possess content?” as an intelligible question, McDowell has
not given an answer we should accept. And, in any case, it is not an ob-
jection to what I called “minimal reliabilism” that it does not answer
that question.
Some Observations on the “How Is It Possible?” Question

The “how is it possible that thoughts have content?” question that pre-
occupies McDowell is not a request for a reduction of intentionality
(“content”) to something else. A long time ago, I myself was tempted
by the desire for a reductive account when I wrote,
The brain’s “understanding” of its own “medium of computation

and representation” consists in its possession of a verificationist se-
mantics for the medium, i.e. of a computable predicate which can
represent acceptability, or warranted assertibility or credibility.61
59. See Hilary Putnam, “Pragmatism,” Proceedings of the Aristotelian Society 95, no. 3
(1995): 291–306; “When ‘Evidence Transcendence’ Is Not Malign” [a reply to Crispin
Wright], Chapter 8 in this volume; and “On Not Writing Off Scientific Realism,” in Philos-
ophy in an Age of Science, 91–108.
60. See Chapter 8 for a defense of this claim.
61. Hilary Putnam, “Computational Psychology and Interpretation Theory,” in Realism
and Reason, vol. 3 of Philosophical Papers (Cambridge: Cambridge University Press, 1983),
130–155. The sentences quoted in the present essay are on p. 142.
Unauthenticated
McDowell’s work has been consistently free of all signs of such a

temptation. What bothers him is something else. The obvious answer
to the “how is it possible?” question, if we prescind from the difficult
question of reference to unobservables, is, “What’s your problem? Don’t
we see (touch, handle, etc.) objects in our environment all the time?”
But the fact that our perception of such familiar objects as apples and
chairs depends on “bits of sensory intake” leads, if we identify that sen-
sory intake with “impressions,” and we identify impressions in turn with un-
conceptualized “qualia,” straight to the conclusion that the basis for all
our “knowledge of the external world” is our qualia; and it is hard to
see how they can be a basis. Moreover, even if we don’t hope for a re-
ductive account of reference and “content,” it is reasonable, especially
after the rejection of Platonist and Rationalist accounts of the mind, to
posit that what we can conceive and what we can refer to depends, at
least in its initial stages,62 on what we have cognitive contact with.63
And if all we have cognitive contact with, in the initial stages of empir-
ical knowledge, is qualia . . . ?
Surely this is a reasonable worry. And McDowell’s principle philo-
sophical claim is that the way to give that worry rest is to reject the idea
of unconceptualized qualia altogether. There are “impressions,” of
course, but they put us in direct contact with the world, and they are
conceptualized—they tell us about that world. But this combination of
“externalism” and conceptualism with respect to the phenomenal char-
acter of experience is, I have argued, untenable on both empirical and
conceptual grounds. So what do I suggest instead?
McDowell is certainly right that appealing to qualia (“bare pres-
ences”) can’t provide an answer to the question as to how concepts and
experiences are connected, or, in McDowell’s terms, how experiences
can rationally constrain beliefs. But to get from that observation to the
conclusion that “the content of experience is conceptual,”64 McDowell
needs to assimilate sensory impressions themselves to apperceptions,
and that is where we disagree. In fact, there are apperceptions that have
no accompanying qualia at all. Suppose I raise my right hand. My aware-
62. See Bertrand Russell, “On Denoting,” Mind 14, no. 56 (1905): 479–493.
63. Of course, any theory has to recognize the possibility of reference “by description.”
But not all reference can be by description.
64. McDowell, Mind and World, 45.
Unauthenticated
ness that I raised it (it didn’t simply “go up”) is a genuine awareness, a
genuine act of apperception, but there is no quale of voluntariness. (I
think I remember that Elizabeth Anscombe somewhere describes this
kind of awareness as “knowledge without observation,” but this seems
to me to be a misdescription. I would say that I did observe that I raised
my hand, but this is observation without any particular qualia, or, to
use a term employed by Alva Noë, an instance of “amodal awareness.”)
Similarly, my awareness when I see a tomato that I am seeing some-
thing that has a round other side and a soft seedy inside involves amodal
awareness and not only qualia. McDowell thinks he has to say that im-
pressions warrant beliefs, and that is the reason that he needs them to
be conceptually articulated; my view is that it is apperceptions that war-
rant beliefs. Of course, certain sorts of apperceptions are internally
related to impressions. But it is the apperceptions and not the impressions
that do the warranting.65 Babies and languageless animals do not have
apperceptive awareness in the demanding Kantian or McDowellian
sense, but I see no reason to deny them qualia.
Of course,66 there is as little hope of a reductive account of appercep-
tion as there is of a reductive account of intentionality; indeed, apper-
ception involves intentionality because it involves recognizing things
and goings-on for what they are, and recognizing involves applying
concepts. Apperceiving some thing or event in my environment is what
I have called a “functional state with long arms,” a world-involving
functional state, in a very liberal sense of “functional state.” Under
normal conditions neither our perceptual experiences nor sentences we
accept are the beginning of the process of forming a perceptual belief.
The beginning is outside our heads; the process of forming a percep-
tual judgment to the effect that there is a notepad on this table is an
exercise of a “function”—in fact, a whole system of functions, some
shaped by evolution, and some shaped by cultural processes that con-
nect me to objects and goings-on in my environment (in this case, to
the notepad and to the table). Forming beliefs in accordance with our
normal biological functions and our linguistic upbringing is not just
uttering noises that are mere responses to qualia, although those qualia
are a part of the causal chain that constitutes the normal formation of
65. I spell this out in more detail in Chapter 10.

66. Part of this paragraph is adapted from Putnam, “Corresponding with Real ity.”
Unauthenticated
a particular perceptual belief on the basis of seeing something in one’s

visual field.67 On a liberal functionalist story, for either our beliefs, or
the proto-beliefs of animals and prelinguistic children, to have content
is just for them to function as representations of external states of af-
fairs. In brief, a belief about the surrounding environment has content
by virtue of being connected to possible states of the world via the ex-
ternalistically identified functional states of the speaker. This is the
classic functionalist account, liberalized by (1) liberalizing function-
alism itself (detaching it from its narrow computationalism and reduc-
tionism) and (2) making functionalism externalist (functional states can
involve tomatoes, and not only the senses and the motor organs). Kant
would, of course, say this is a question-begging answer, and I would tell
him that the sort of a priori proof that our concepts have content,
the proof of the “object validity” of our categories, that he hoped for is
a chimera.
The Epistemic Role of Qualia

If qualia cannot play the epistemic role that traditional empiricism as-
signs to “impressions,” and if they cannot be “puffed up” to play that
role by attributing propositional content to them, then do they have any
epistemic role to play at all?68 I suggest that they do, but not in all cases
of perception (as traditional empiricism insists). And this is a good mo-
ment to refer once more to the views of Ned Block. Earlier, I quoted a
passage in “Wittgenstein and Qualia” that included the following:
I have heard [Moore] quoted saying “the moment we try to fi x our

attention upon consciousness and to see what, distinctly, it is, it
seems to vanish: it seems as if we had before us a mere emptiness.
When we try to introspect the sensation of blue, all we can see is
67. Or at least that is a reasonable hypothesis on the basis of present knowledge. But it
isn’t a priori, as many phi losophers seem to think, that the relevant qualia (partially) cause
or even temporally precede the cognitive awareness that I am seeing a tomato or a rabbit; it
is logically possible that the cognitive awareness and the quale are effects of a common
cause, or that the qualia are the effect and the cognitive awareness is the cause. But these
questions relate to the subpersonal mechanisms behind awareness; the idea that they are a
question for personal-level psychology is the central mistake of classical empiricism.
68. This question was put to me by Hilla Jacobson.
Unauthenticated
the blue; the other element is as if it were diaphanous. But these

words are followed by what I regard as a more significant truth:
“Yet it can be distinguished if we look attentively enough, and
know that there is something to look for.”69
I agree with Block that we can attend to our own qualia. Moreover,
this is not just something that eighteenth- and nineteenth-century
introspective psychologists or twentieth-century phenomenologists
or the subjects of twenty-first century experimental studies of percep-
tual experience did and do; there are many moments in life when we
are interested in “what it is like” to experience such and such, in a
sense that involves attending to our qualia rather than (or, as the occa-
sion demands, along with) attending to the worldly properties that we
perceive when we “see through” the qualia as through a transparent
window. Qualia themselves are as much capable of being the objects of
perceptual experience as, say, trees or rabbits. If it seems shocking to
“downgrade” the epistemological role of qualia to the level of the epis-
temological role of the trees and the rabbits, it is no doubt because the
view that knowledge is “based on” qualia has become so deeply
entrenched.
In his Second Woodbridge Lecture,70 McDowell takes Sellars to task
for positing “non-concept-involving episodes or states in sensory and
specifically visual consciousness.”71 As McDowell reads Sellars, Sellars
thinks we need to see the flow of “conceptual representations” in per-
ception as “guided by manifolds of sensations,” where the latter are the
non-concept-involving episodes or states in question.72 And, in line
with his “naive realist” view of experience, McDowell replies that it is
not “non-concept-involving episodes or states” that do the guiding, but
the “objects themselves” (the rabbit and the trees, to stick to our exam-
ples).73 I agree with McDowell that the guiding that is epistemologically
relevant is done by the objects themselves, but I agree with Sellars
that part of the guiding that is causally relevant— assuming our current
69. Block, “Wittgenstein and Qualia,” 280.

70. McDowell, “The Logical Form of an Intuition,” in Having the World in View, 23–43,
esp. 23–33.
71. McDowell, “Logical Form of an Intuition,” 24.
Unauthenticated
neurological picture—is partly done by qualia. What McDowell cannot

contemplate as so much as a possible position is that while qualia
(“non- concept-involving episodes or states”) are causally relevant to
perception, they do not and cannot do the epistemological heavy
lifting that the empiricist tradition asks them to do. “Naive realism” is
right in insisting that we do perceive external objects and that this is
not a matter of drawing inferences from qualia; that is why we need an
account of perception that connects perception both with the objects
perceived (an “externalist” account) and with the subject’s conceptual
activity. But “naive realism” is wrong when it denies any gap between
the properties of external objects and the phenomenal character of
experience.
Unauthenticated
D 12
The Development of
Externalist Semantics
M y i n t e r e s t i n l i ngu is t ic s goes almost as far

back as my interest in philosophy.1 The latter goes back to my high
school days, when I and two or three friends started a small philosophy
club, and I went on to major in philosophy at the University of Penn-
sylvania. I don’t know how I learned that there was a small autonomous
“section” headed by Zelig Harris in the anthropology department called
“Linguistic Analysis,” but learn that I did, and from my second year in
college I took every course that was offered by Harris. (Noam Chomsky
was one of my fellow students.) Harris was not at that time teaching
his own view of semantics. In any case, what I learned was his tech-
niques for describing the syntax of a natural language, and likewise
Chomsky’s first famous publication, Syntactic Structures, was entirely
devoted to syntax. As far as my own experience for a number of years
went, “meaning” was a topic discussed by philosophers, not by linguists.
The only course on the topic I had as a student, in fact, was a course on
1. This chapter consists of the lecture I gave on November 1, 2011, on the occasion of
being awarded the Rolf Schock Prize in Logic and Philosophy by the Royal Swedish
Academy of Science for my contribution “to the understanding of semantics for theoretical
and ‘natu ral kind’ terms, and of the implications of this semantics for philosophy of lan-
guage, theory of knowledge, philosophy of science and metaphysics.” In it I describe the
path that led to the work for which I was awarded that prize, including at least one wrong
idea that I had to overcome on the way.
199
Unauthenticated
200 LOOKING BACK
Carnap’s Meaning and Necessity taught by Morton White, who, along

with C. West Churchman, was my principal mentor in philosophy at
“Penn.”
I went to graduate school, of course (at the University of California–
Los Angeles), and became an analytic philosopher and philosopher of
science (and, mainly on my own,2 a mathematician as well). My very
first published paper (1954) was titled “Synonymity, and the Analysis
of Belief Sentences,”3 and that title might lead you to think that I was
already focusing on semantic issues, but that is not the case. That paper
grew out of a conversation with Carnap about a criticism due to Benson
Mates of a criterion of synonymy (for sentences, not for individual
words) that Carnap had proposed. I suggested that Mates’s criticism
might be met by counting the logical form of a sentence as one of its
meaning-components, and Carnap urged me to publish this proposal,
which I did. But that was all there was to it.
My interest in semantics, although at the time I would not have used
that word, really began in 1957–1958, when Herbert Feigl invited me
to spend my semester of leave from Princeton at his Minnesota Center
for the Philosophy of Science. The following spring, Paul Ziff, who was
on leave from Harvard where he was an untenured assistant professor,
delivered a seminar in Princeton on what became his book, Semantic
Analysis,4 which was attended by myself and a number of my students,
including Jerry Fodor and Jerrold Katz, and those two things, the se-
mester at Feigl’s Center and the impact of Ziff’s course, together led to
a turning point in my philosophical thinking.
At the Minnesota Center, I was the only (even partial) defender of
Quinian skepticism about the analytic-synthetic distinction, and the
only critic of Adolf Grünbaum’s “conventionalism” in the philosophy
2. The only graduate course in mathematics I had, in fact, was a course in Ideal Theory
at Harvard in 1948–1949. I learned mathematical logic by solving problems posed for me by
Georg Kreisel, when I became an assistant professor at Princeton University, starting in
1953, and, after publishing a number of papers in logical and mathematical journals, I re-
ceived tenure in both the philosophy department and the mathematics department at
Princeton in 1959.
3. Hilary Putnam, “Synonymity, and the Analysis of Belief Sentences,” Analysis 14
(1954): 114–122; repr. in Propositions and Attitudes, ed. N. Salmon and S. Soames (Oxford:
Oxford University Press, 1988), 149–158.
4. Paul Ziff, Semantic Analysis (Ithaca, N.Y.: Cornell University Press, 1960).
Unauthenticated
The Development of Externalist Semantics 201
of geometry.5 I enjoyed my discussions with all the members of the

center (including Paul Feyerabend, who was the other young visitor at
the center that semester), but my main activity was to write the first
draft of a paper titled “The Analytic and the Synthetic.”6 The conclu-
sions of “The Analytic and the Synthetic” are closely connected with
what was later called “externalism.”
In those years, the fifties and sixties of the last century, young ana-
lytic philosophers tended to be either partisans of logical positivism or
partisans of Oxford “ordinary language philosophy.” Although my Dok-
torvater, Hans Reichenbach, was a logical positivist,7 I had come to
have serious doubts about both movements. The “ordinary language”
philosophers knew little, and for the most part cared less, about science,
while the logical positivists thought that terms in scientific theories,
including geometrical terms such as “distance,” have their meanings
fi xed by “defi nitions,” which they sometimes took to be operational
definitions (“coordinating definitions”), and sometimes theoretical
sentences such as “E = ½ mv2” (as a defi nition of “kinetic energy”).
Whenever a scientific revolution forced us to revise those “defi nitions”—
something they recognized had already happened more than once—it
followed from their account that the reference of those terms changed.
For example, in nineteenth-century physics, the kinetic energy of an
object was defined as one-half the product of its mass and the square of
its velocity, ½ mv2. But according to Einstein’s special relativity, the ki-
netic energy (the total relativistic energy minus the famous mc 2) isn’t
exactly ½ mv2, but that plus a correction which is tiny when the velocity
v is tiny compared to the velocity c of light. However, I argued, “kinetic
energy” did not refer to one physical magnitude in 1899 (namely ½ mv2)
and to a dif ferent magnitude (½ mv2 plus the relativistic correction) in
5. Hilary Putnam, “Memo on ‘Conventionalism’,” Minnesota Center for the Philos-

ophy of Science, March 22, 1959; fi rst published in Mathematics, Matter and Method, vol. 1 of
6. Hilary Putnam, “The Analytic and the Synthetic,” in Scientifi c Explanation, Space, and
Time, ed. Herbert Feigl and Grover Maxwell (Minneapolis: University of Minnesota Press,
1962), 358–397; repr. in Hilary Putnam, Mind, Language and Reality, vol. 2 of Philosophical
Papers (Cambridge: Cambridge University Press, 1975), 33–69.
7. As related in Chapter 6 of this volume, Reichenbach did not like the term “positivist”
because he associated it with the view that all scientific concepts referred, however indi-
rectly, to human sense-data, which was something he did not believe. He preferred “logical
empiricist.”
Unauthenticated
202 LOOKING BACK
1905; rather, I said, we were talking about the same magnitude in those
two years, but we came to have a more accurate idea of the nature of that
magnitude. And similarly, I argued, when physicists gave up the idea
that atoms are little solar systems (the Bohr model), they did not change
the reference of the term “atom”; rather, they arrived at a better idea of
what atoms are.
Two points are essential here, if one wants to understand the con-
nection with semantic “externalism”: First, the idea that terms must have
precise and unchanging definitions, and that it is those definitions that
fix their reference, has to be given up. Second, it is a combination of the-
ories and experiments that tell us what our terms refer to, and not un-
changing “definitions.” Because the results of experiments depend on
the external environment, as does the fate of the theories we use those
experiments to test, one way of expressing this is that the terms I dis-
cussed in “The Analytic and the Synthetic” (e.g., “energy,” “atom”) have
their reference fixed by the environment itself, and not simply by what is
in our heads. But that is not the way I thought about it at the time, which
is why ten years had to pass before I arrived at “semantic externalism.”
As I mentioned, the first draft of “The Analytic and the Synthetic”
was written in 1957–1958; my fi rst explicitly “semantic externalist”
paper, “Is Semantics Possible?,” was written ten years later.8 One
reason for this substantial time gap is that I thought of “The Analytic
and the Synthetic” as a paper only about one limited issue in seman-
tics, namely, what was right and what was wrong in Quine’s famous at-
tacks on the analytic-synthetic distinction. The moral I drew was that
while Quine was wrong to deny that there are any analytically defin-
able terms in the language at all—I defended, and still defend, the idea
that there is a point to saying that “all bachelors are unmarried” is an
analytic statement— the notion of analytic truth could not bear the
weight put on it by Russell, Frege, and the logical positivists, all of whom
claimed that the truths of mathematics were analytic, and some of
whom also who thought that “E = ½ mv2” was analytic in classical
physics and that Einstein—for good reason, to be sure—“changed the
8. Hilary Putnam, “Is Semantics Possible?” Metaphilosophy 1 (1970): 187–201; revised ver-
sion in Language, Belief and Metaphysics: Contemporary Philosophic Thought: The International
Philosophy Year Conferences at Brockport, vol. 1, ed. H. E. Kiefer and M. K. Munitz (Albany:
State University of New York Press, 1970), 50–63; repr. in Mind, Language and Reality,
139–152.
Unauthenticated
definition.” However the idea of theorizing about “semantics” as such

was far from my mind. After all, the linguistics I knew did not include
any such field, while in analytic philosophy at the time “semantics”
mainly referred to the part of logic that deals with Tarski’s formal
analysis of the notion of truth. To explain how semantics came to be an
issue for me, and for some of my students (particularly for Jerrold Katz
and Jerry Fodor), I need to describe Paul Ziff’s seminar.
Ziff’s theory, which we discussed for hours every day in the spring
of 1958, has been largely bypassed by the development of linguistics and
philosophy of language, but it contained important, if eccentrically for-
mulated, insights.9 It was not a theory of understanding; thus, unlike
Quine’s Word and Object to which it was clearly responding, it did not
presuppose Skinnerian or any other psychological theory.10 Nor did
Ziff try to show how semantics could be done without the notion of
“meaning,” as Davidson tried to do in a famous essay published the same
year.11 Rather, it was concerned with the question of what sort of evi-
dence a linguist trying to write a lexicon for a language (either a familiar
language, or one previously unknown) might draw on and how that ev-
idence might be organized. (In this respect Ziff’s Semantic Analysis re-
sembles Word and Object, but with a much more realistic idea of what
linguistic evidence is.) The idea of trying to describe what an ideal dic-
tionary entry for a word might look like, and what evidence would be
relevant, was to be central to my “The Meaning of ‘Meaning’ ” and its
predecessor “Is Semantics Possible?,” although the conclusions I reached
were quite different.12 But mainly it was the sheer ambition of Ziff’s pro-
ject that impressed us. As Jerrold Katz later said of Semantic Analysis, it
was “a pioneer work, in that it is the first to propose an empirically based
9. One of those insights is that a property may be linguistically associated with a natu ral
kind term although the property is only contingently, not analytically, associated with the
relevant kind. For example, the properties striped and quadrupedal are associated with the
noun tiger, but it is not analytic that all tigers are striped, and possibly not even true that
there are no three-legged tigers. “Tigers are striped” and “Tigers are quadrupeds” are
“state regularities” in Ziff’s terminology. I did not preserve the terminology, but Ziff’s
observation led me to the notion of a stereotype in the “The Meaning of ‘Meaning’,” in Mind,
Language and Reality, 215–271.
10. W. V. Quine, Word and Object (Cambridge, Mass.: MIT Press, 1960).
11. Donald Davidson, “Truth and Meaning,” Synthese 17 (1967): 304–323.
12. I mean “ideal” from a strictly linguistic point of view, with no philosophical or be-
havioristic axe to grind, of course.
Unauthenticated
204 LOOKING BACK
theory of meaning to deal systematically with the various topics that

are part of the subject of meaning, and to attempt to fit such a theory
into the larger framework of structural linguistics.”13
Following Ziff’s seminar, in 1958–1959 and 1959–1960, I worked
mainly on a famous mathematical problem, Hilbert’s Tenth Problem,
and in 1960–1961, having finally received tenure in both mathematics
and philosophy at Princeton, I finally had a full year’s leave, which I de-
voted again mainly to mathematical logic. But in 1961 I left Princeton to
become professor of Philosophy of Science at the Massachusetts Insti-
tute of Technology (MIT), where I also had the opportunity to hire
a number of philosophers and to create what was first a “philosophy
section” of MIT’s humanities department, and later an autonomous
philosophy department (and still later one-half of MIT’s Department
of Philosophy and Linguistics). Among the philosophers MIT hired
on my recommendation were the “two Jerries,” Jerry Fodor and Jer-
rold Katz. The question Paul Ziff had raised for us was a main topic of
discussion among the three of us: What might a field called “seman-
tics” look like? How could we lay the foundations for (to use Katz’s
words that I quoted a moment ago) “an empirically based theory of
meaning to deal systematically with the various topics that are part of
the subject of meaning,” that might “fit . . . into the larger framework
of structural linguistics”?
Although in the course of time Fodor, Katz, and I all came to pro-
pose very dif ferent answers to that question, at the beginning we all
agreed, and it was my later realization that what we agreed on in those
years at MIT must be wrong that brought about my move to “external-
ism.”14 What we agreed on was that the meaning of a word could be
given by “semantical rules,”15 and what we worked on was the problem
of saying what such rules might look like. But while that was a full-time
worry for Katz and Fodor, I was at that time (1961–1965) doing math-
13. Jerrold Katz, “Review of Semantic Analysis,” Language 38, no. 1 (1962): 52–69.
14. I left MIT for Harvard in 1965.
15. “How Not to Talk about Meaning: Comments on J. J. C. Smart,” in Boston Studies in
the Philosophy of Science, vol. 2, ed. R. S. Cohen and M. R. Wartofsky (New York: Human-
ities Press, 1965), 205–222; repr. in Mind, Language and Reality, 117–131. I talk about “seman-
tical rules” in the Boston Studies publication (216–222), and in Mind, Language and Reality
(126–131). This is the text of a talk to the Boston Colloquium in the Philosophy of Science
that I gave December 12, 1963.
Unauthenticated
ematical logic a lot of my time, and also beginning to apply concepts

from computer science to the philosophy of mind.16 So I pretty much
dropped semantics for the time being, apart from the conversations I
described. But when I decided to teach a course in the philosophy of
language at Harvard in 1967, I returned to the subject, and I found that
I was dissatisfied with the idea that the two Jerries and I had shared,
that one could state semantical rules with the properties that (1) they
are in some sense “internalized” by every competent speaker and (2)
they determine the meaning of every word in his or her repertoire.
Those were the wrong ideas I had to overcome to get to externalism.
Reconsidering those ideas in 1967, I recalled that Locke had once
claimed that “gold” is synonymous with “precious, incorruptible, yellow
metal soluble in aqua regia,”17 and I reflected that I had no idea what
“aqua regia” is, so that couldn’t possibly be what I (or most other English
speakers) mean by “gold.”18
The problem I faced was this: if the meaning of the word gold is given
by a battery of semantical rules, as Katz, Fodor, and I had all been
supposing, what are those rules? I had ruled out “ ‘gold’ is the name of a
precious metal,” “ ‘gold’ is the name of a yellow metal”—indeed, every-
thing except “ ‘gold’ is the name of a metal” in “How Not to Talk about
Meaning,”19 and clearly “ ‘gold’ is the name of a metal” isn’t sufficient as
a description of the meaning of the word.
16. Although I had already published “Minds and Machines,” in Dimensions of Mind, ed.
Sidney Hook (New York: New York University Press, 1960), 148–180, while I was at MIT, I
wrote “Robots: Machines or Artificially Created Life?” Journal of Philosophy 61 (1964): 668–
691; repr. in Mind, Language and Reality, 386–407.
17. Locke says dif ferent things about “our idea of gold” in dif ferent places. In one place
he writes, “He that will examine his complex idea of gold, will fi nd several of its ideas that
make it up to be only powers; as the power of being melted, but of not spending itself in the
fi re; of being dissolved in aqua regia [emphasis added— HP], are ideas as necessary to make up
our complex idea of gold, as its colour and weight: which, if duly considered, are also
nothing but dif ferent powers” (John Locke, An Essay Concerning Human Understanding,
ed. P. H. Nidditch (Oxford: Oxford University Press, 1975), II, 23, §10). But in another
place he has it that the nominal essence of the name “gold” “is that complex Idea the word
Gold stands for, let it be, for instance, a Body yellow, of a certain weight, malleable, fusible,
and fi xed” (III, 6, §2). But the same objection applies: I have no idea what the “certain
weight” of gold is, and so, according to Locke, I do not know the “nominal essence,” that is,
the meaning of the word.
18. Putnam, “How Not to Talk about Meaning,” 217–218.
19. Putnam, “How Not to Talk about Meaning,” 217–218.
Unauthenticated
206 LOOKING BACK
I found myself driven to an idea that was wholly new to me, and ap-
parently to other philosophers as well: nothing that is in the head of the
average speaker suffices to determine what her word gold refers to.
Meanings aren’t in the head.
Well, if they aren’t in the head, where are they? Of course, the brain
is in the head, and the brain has to undergo appropriate changes (mat-
uration, and all the various effects of acculturation) before one can speak
a natu ral language. “Meanings aren’t in the head” doesn’t mean that
the brain has nothing to do with semantic competence. But what fi xes
the meanings of a speaker’s words is not just the state of her brain; the
reference of our terms is generally fixed by two things that classical phi-
losophy of language either ignores or mentions only as an afterthought:
other people and the world.
You may have noticed that I just shifted from “meaning” to “refer-
ence”; they are not the same, but they are closely connected. Consider:
typical modern English speakers have only a vague idea of what an
elm tree looks like. They depend on other people, especially experts, to
determine for them which trees are “elms,” and if there is an island
somewhere in which a dialect of English is spoken in which the word
“elm” is the name of a dif ferent species of tree— say, of beeches—
then we would say that on that island “elm” means beech. For natural
kind words, for instance, names of biological species and names of sub-
stances, difference in reference counts as difference in meaning.20
One problem with viewing the reference of my words as something
fi xed by the way my brain is “programmed” is that such an individual-
istic account ignores the linguistic division of labor. Not only am I unable
to reliably distinguish elms from other species of tree; the fact is that I
do not have to be able to do this on my own. If I need to know whether a
particular tree is an elm, I can ask someone who is more knowledge-
able than I.
In the case of names of substances—metals, for example—the lin-
guistic division of labor is even more obvious; or it is obvious once
pointed out— evidently it wasn’t obvious to Locke, with his various at-
tempts to say what the “idea” of gold is that we English speakers sup-
20. The extension of a natu ral kind term—that is, the set of things it is true of—is one
of the components of its “meaning vector,” in the theory I proposed in “The Meaning of
‘Meaning’.”
Unauthenticated
posedly possess, or to me when I thought the reference of gold must be

fi xed by semantical rules that every competent speaker tacitly knows.
There are plenty of people who can determine for us whether a piece
of metal is really gold—jewelers, chemists, atomic physicists, among
others.
At this point, it is important to see that the experts themselves do
not identify substances by unchanging “definitions,” unchanging nec-
essary and sufficient conditions. (Recall what I had argued in “The An-
alytic and the Synthetic,” the paper I had first drafted ten years earlier
in Minnesota!) Even the criteria used by experts are sometimes badly
off the mark. But in cases such as the case of gold, cases in which
there are large numbers of what are agreed to be good samples of the
pure metal, the deviant examples will be declared to be such when sci-
ence discovers that their nature is not the same as that of the great
majority of the accepted samples. And if a sample is too deviant, we
will decide it isn’t gold at all, even if it superficially resembles gold.
The samples do not literally speak, but they have a decisive vote in what
counts as a member of the kind. In “The Meaning of ‘Meaning’,” I re-
ferred to samples that play this role as our paradigms.
Of course, there are natural kinds that we are all pretty good at iden-
tifying without relying on experts, for example, water. But it would be
wrong to think that the extension of the word “water” is fi xed simply
by the superficial properties of water—that is, that “water” means
“liquid that quenches thirst, has such and such a taste, . . . and so on.”
And it would be equally a mistake to say that “water” means “H2O.” It
is true that most educated people now know (or think they know) that
H2O is the chemical formula for water, but educated speakers and un-
educated speakers do not speak dif ferent languages. (Moreover, “water
is H2O” is not precisely true!21) Water has the same meaning in the
mouth of an English-speaking scientific ignoramus and the mouth of
an English-speaking chemist; they both know the meaning of “water”
in English in the sense of possessing the sort of linguistic competence
that every person who masters the language is linguistically obliged
to have in connection with that word. They refer to the same liquid
when they use the word. Moreover, speakers who lived before Dalton
21. Normal water is actually a quantum mechanical superposition of H 2O, H4O2,

H6O3 . . . Very little (if any) water is simply H 2O.
Unauthenticated
208 LOOKING BACK
introduced the modern notion of a chemical compound referred to the

same liquid when they used the term “water.” The word does not change
its meaning every time we discover a better account of the nature of
water. The meaning and reference of “water” aren’t what has changed;
it is our knowledge about water that has changed. The meaning of
“water” isn’t fi xed by a defi nition, either in terms of the observable
properties of the substance or in terms of the properties mentioned in
our latest scientific theory; it is fi xed by the nature of our paradigms of
water. To be water is to be the same liquid as this, where this can be
(almost) any of the paradigm examples of water.
What I have described is the reasoning that led me to the conclusion
that the meaning and reference of the term “water” aren’t in the head;
they are fi xed by what goes on in the world. (Saul Kripke had already
entertained similar ideas when he was a member of Harvard’s Society
of Fellows [1963–1967]. I didn’t learn of Kripke’s views until 1970, how-
ever, when news of his lectures at Princeton spread through the philo-
sophical world.)
I presented my account at the 1968 Summer Institute in Philosophy
of Language in Seattle. (David Kaplan, who was present at those lec-
tures, once wrote me, “I remember your quickly disabusing me of the
idea that the intension of a natural kind word [that which determines
the extension in a possible world— HP] is something we ‘grasp,’ as
Carnap would have put it. It ain’t in the head, as you put it. And almost
as soon as you said it, it seemed right.”) I presented a one-lecture ver-
sion of the theory at the University of Minnesota in 1969, and the first
published version appeared in 1970, under the title “Is Semantics Pos-
sible?” But it was not until December 1972 that I wrote the first draft
of “The Meaning of ‘Meaning’.” It was then that what became the best-
known argument for semantic externalism occurred to me. I refer to
the “Twin Earth” argument.
The argument is a simple thought experiment: imagine a planet like
earth— call it “Twin Earth”—on which the liquid that fills the lakes and
rivers, that people drink, is not H 2O but a dif ferent compound, XYZ,
with similar superficial characteristics. The Twin Earthers are supposed
to be our doppelgängers, and some of them even speak English. Also,
I imagine the year to be 1750, so it is before the chemical composition
of either water or twater (Twin Earth “water”) is known. The English-
speaking Twin Earthers naturally call twater “water” (and the French-
Unauthenticated
speaking ones call it “eau,” and the German-speaking ones call it

“Wasser”). The linguistic intuition of the great majority of people who
have considered this thought experiment is that upon learning that
Twin Earth “water” doesn’t consist of H2O at all, we Earthers would
say “it isn’t really water.” The word “water” has a dif ferent meaning on
Earth and on Twin Earth. Twin Earthian Oscar’s word “water” and
Earthian Oscar’s word “water” are homonyms, but not synonyms. They
do not have the same meaning—not even if Earthian Oscar and Twin
Earthian Oscar happen to be microphysical duplicates!
It is irrelevant that there is no such (possible) compound XYZ, because
we can imagine that neither Oscar nor Twin Oscar know this esoteric
fact. (Indeed, I am not sure I myself know such a thing, unless having
been told that this is so by a former chemist-turned-philosopher-of-
science counts as “knowing.”) The question is what we would say under
perfectly imaginable circumstances, not what is chemically or physi-
cally possible. Moreover, the example can easily be modified to avoid
this (irrelevant) objection, and that is what I did in a later paper:22 just
imagine that Twin Earth water consists of 50 percent water and
50 percent a tasteless and nonpoisonous liquid that does not quench
thirst. Of course, Twin Earthers, if their bodies are like ours, would
have to drink more of their “water” per day than we drink, but it is plau-
sible to suppose that ordinary speakers do not notice this, and that
they would not regard this fact as relevant to the meaning of the vo-
cable “water” if they did notice it. (It used to be the case that Brits ate
ten times as much chocolate per year as Americans, but no one regarded
this as a difference in the meaning of the vocable “chocolate.”) And it
is not essential to the Twin Earth thought experiment that Oscar and
Twin Earth Oscar be microphysical duplicates, or that their brains be
microphysical duplicates: it is enough if their brains are in sufficiently
similar states with respect to those systems that enable the appropriate
use of words.
Because I mentioned Saul Kripke, I should also mention that, in his
great book Naming and Necessity, his path to similar conclusions was a
via negativa: if you think that what we mean by a natural kind term (say,
22. Hilary Putnam, “Meaning Holism,” in The Philosophy of W. V. Quine, ed. L. E. Hahn
and P. A. Schilpp (La Salle, Ill.: Open Court, 1986), 405–431; repr. in Realism with a Human
Face (Cambridge, Mass.: Harvard University Press, 1990), 278–302.
Unauthenticated
210 LOOKING BACK
“water” or “gold”) is fi xed by a definition, just tell me that definition. You

will quickly find that all the proposed definitions either presuppose es-
oteric knowledge—the precise weight of gold, in the case of one of
Locke’s suggestions, or the fact that gold is soluble in aqua regia [what-
ever that is]—knowledge that most speakers certainly do not possess—
or have counterexamples that are easy to find.
Externalism and Perception

Because I am also being awarded the prize for my “contribution to the
understanding of semantics for theoretical and ‘natural kind’ terms, and
of the implications of this semantics for philosophy of language, theory
of knowledge, philosophy of science and metaphysics,” I want to close
by saying something about those implications. “The Analytic and the
Synthetic” was already an “application of externalist semantics to the
philosophy of science,” even if I did not think of it that way. As for impli-
cations for metaphysics, I believe the area in which there is the most
widespread acceptance of externalism, together with fascinating dis-
agreements about what externalism implies, is the philosophy of mind,
and particularly the philosophy of perception. In “The Meaning of
‘Meaning’,” I said meanings “ain’t in the head.” Our verbalized thoughts
have meaning only in conjunction with our transactions with objects in
our environment and with other speakers. It is in the context of a net-
work of social and physical interactions, and only in such a context, that
I can do such a thing as “think that the price of gold has become very
high in recent years.” If thinking that thought is what I once called a
“functional state,” it is not (as I mistakenly believed) simply a “computa-
tional state” of my brain; the “function” in question is a world-involving
function. The thought is no more simply in my “head” than the meaning
of the word “gold” is. And if thoughts aren’t in the head, then the mind
isn’t in the head either. The mind isn’t a thing with a definite location,
but a system of world-involving abilities and exercises of those activities.
On this all externalists in the philosophy of mind agree.23
23. I should mention that some implications of externalism for theory of knowledge are
explored in my well-known “Brains in a Vat” thought experiment. See Hilary Putnam,
“Brains in a Vat,” in Reason, Truth and History (Cambridge: Cambridge University Press,
1981), 1–21.
Unauthenticated
Coming to the controversial area of perception, there is also some

agreement. The long-maligned “naive realist” view that we have direct
cognitive contact with external things and a number of their proper-
ties, and not just with our own sense data, is now close to becoming
the orthodoxy. To perceive an apple is to be aware of the apple, not of
sense data, and we are able to be aware of things like apples because we
have evolved to respond to the appropriate “affordances,” to use Gib-
son’s famous term.
However, serious disagreement sets in is when one asks, “What is the
role of the phenomenology of perception, the ‘what it’s like’ aspects,
the ‘qualia,’ in the perceptual transaction with, for example, the apple?”
There are “phenomenists” about qualia, such as Ned Block, who would
say that while perception is world-involving, qualia themselves are non-
conceptual internal states (a position I myself agree with).24 There are
thinkers who would identify our visual sensations (for example) with
properties of the surface of the apple itself (call them “ultra-naive
realists”) when the perception is “veridical.”25 There are thinkers
(Dretske, Tye, and many others) who would identify our qualia with
“information” (and who differ among themselves concerning what sort
of “information”). And there are even thinkers who would say that,
just as it is possible for microphysical duplicates to mean different things
by their words if their environments are dif ferent, so it is possible that
Oscar’s visual sensation when he looks at the blue sky on Earth is what
I would call a “visual sensation of blue” and Elmer’s visual sensation
when he looks at the blue sky on Twin Earth is what I would call a
“visual sensation of green,” even if their brains are in the same state (a
position I fi nd highly implausible).26 But this is not a mad house, as a
skeptic (or a Wittgensteinian) might suspect! Important interactions
between researchers in brain and behavioral science and philosophers
24. N. Block, “Mental Paint,” in Refl ections and Replies: Essays on the Philosophy of Tyler
Burge, ed. M. Hahn and B. Ramberg (Cambridge MA: MIT Press, 2003), 165–200.
25. For example, see Michael Martin, “On Being Alienated,” in Perceptual Experience,
ed. Tamar Szabo Gendler and John Hawthorne (Oxford: Oxford University Press, 2006),
354–410.
26. For example, see Michael Tye, “Phenomenal Externalism, Lolita and the Planet
Xenon,” in Qualia and Mental Causation in a Physical World: Themes from the Philosophy of
Jaegwon Kim, ed. Terence Horgan, Marcelo Sabatés, and David Sosa (Cambridge, Mass.:
MIT Press, 2015), 190–208; and Michael Tye and Alex Byrne, “Qualia Ain’t in the Head,”
Nous 40, no. 2 (2006): 241–255.
Unauthenticated
212 LOOKING BACK
have already resulted from these discussions. No one can be sure what
the philosophy and psychology of perception will look like in another
fifty years, but that is an indication of the vitality of the field. That my
contributions to something as dry sounding as “the understanding of
semantics for theoretical and ‘natural kind’ terms” has helped to revive
interest in the metaphysics of perception gives me great pleasure.
Unauthenticated
D 13
Sixty-Five Years of Philosophy
A PARTICIPANT’S THOUGHTS AND EXPERIENCES
W h e n i n 2012 I accepted an invitation from Dr. Rupert

Read to be part of a series of Royal Institute of Philosophy Public Lec-
tures on 50 Years of Philosophy, I wrote, “Would it be too narcissistic
if we called my lecture ‘Sixty five years of philosophy: a participant’s
thoughts and experiences?’1 (I would mention in the talk that fifteen
years ago I published a piece called ‘A Half Century of Philosophy
Viewed from Within.’2 Part of what the talk would describe is how
my view of those years has itself changed in this new century. If you
don’t like the idea, I’ll try to come up with something else.)” But
when the time to write this lecture came, I quickly realized that what
I had suggested was much too ambitious. In 1997, I had strong opin-
ions about where philosophy had been headed in the previous fifty
years, and I confidently issued verdicts on which tendencies were pos-
itive and which were misguided; today I disagree with some of those
verdicts, and I would qualify the verdicts with which I still agree. For
example, at that time I wrote:
1. This chapter was a lecture delivered via Skype at the University of East Anglia on
December 5, 2013, from my home in Massachusetts. The sixty-five years counts from my
graduation from the University of Pennsylvania.
2. Hilary Putnam, “A Half Century of Philosophy, Viewed from Within,” Daedalus:
Proceedings of the American Academy of Arts and Sciences 126, no. 1 (1997): 175–208.
213
Unauthenticated
214 LOOKING BACK
If what “scientific realism” meant to philosophers like myself at the

beginning of the 1960’s was simply the rejection of positivism, and,
more generally, of the idea that the statements of the natural sci-
ences require philosophical reinterpretation, within a few years it
was to develop into an elaborated metaphysical position, or rather
a pair of positions (each of which has many versions). The fi rst po-
sition, which I shall call “panscientism,” holds that philosophical
problems are fated, in the end, to be resolved by the progress of
the natural sciences, and that the best the philosopher can do is to
anticipate that progress, and suggest how the sciences can solve
them. The second position, for which I shall employ a term intro-
duced (for his own view) by Simon Blackburn, but which I shall
use in a wider sense, I shall call “quasi-realism.” This position does
not claim that all philosophical problems will be solved by natural
science, but it does hold that the complete description of reality as
it is “in itself” is given by natural science, and, in most versions of
the position, by physics. The idea that there is a sharp distinction
between the way things are “in themselves” and how they appear
to be, or how we speak of them as being, is characteristic of the
position. What distinguishes the second position from the first is
the idea that many of the ways we speak, and, indeed, have to
speak, do not correspond to the way things are in themselves, but
represent “local perspectives.” To the extent that philosophy
has to clarify and help us understand the status of these local
perspectives, it has tasks over and above the tasks of the natural
sciences. The “local perspectives” do not, however, have any real
metaphysical significance; only natural science has that. Paul and
Patricia Churchland, Daniel Dennett, and Jerry Fodor, notwith-
standing their substantial disagreements, are all representatives
of the first position; as representatives of the second position, in
addition to Simon Blackburn, and again notwithstanding their
substantial philosophical disagreements, I will cite the distinguished
British philosopher Bernard Williams. Of course, not all analytic
philosophers are either pan-scientists or quasi-realists; but these
two attitudes have very much come to dominate the scene in “an-
alytic metaphysics.”3
3. Putnam, “A Half Century of Philosophy,” 183.
Unauthenticated
Sixty-Five Years of Philosophy 215
Those of you who are lucky enough to own a copy of my Philosophy

in an Age of Science know that the scientific realism I presently defend
does not amount only to the claim that scientific theories don’t “require
philosophical reinterpretation.” In the philosophy of physics, and par-
ticularly of quantum mechanics, I argue, the interpretations of contem-
porary physical theories produced by scientists are neither clear nor
unambiguous, and they themselves require—not reinterpretation, but—
interpretation;4 interpretation not only by philosophers, but by phi-
losophers and physicists in dialogue with one another. Of course, I still
reject the positivist interpretation(s), but I do not scorn the enterprise
of philosophical interpretation of physics. And the scientific realism I
defend in Philosophy in an Age of Science is not a claim about whether sci-
entific theories do or do not need “reinterpretation”;5 it is the claim(s)
that most of the entities postulated by our best scientific theories do
exist, and that those theories are approximately true statements about
how they behave. Moreover, as I first said in 1975, scientific realism, in
this sense, is the only philosophy that doesn’t make the success of sci-
ence a miracle.6
I do still reject “panscientism” and “quasi realism.” But, whereas the
alternative I defended in the 1997 article was a position I ascribed to
Wittgenstein, that the real task of philosophy is to lead us to see the
“unintelligibility” of philosophical “positions,” in Philosophy in an Age
of Science, I criticize the whole idea that philosophical positions are often,
let alone as a rule, literally unintelligible.7 The characterizations of
philosophy I now prefer are very dif ferent. As I say there, two defini-
tions of philosophy appeal to me the most, and each definition requires
to be supplemented by the other.8
4. Hilary Putnam, Philosophy in an Age of Science: Physics, Mathematics, and Skepticism, ed.
Mario De Caro and David Macarthur (Cambridge, Mass.: Harvard University Press, 2012),
chap. 6, “A Philosopher Looks at Quantum Mechanics (Again),” and chap. 7, “Quantum
Mechanics and Ontology.”
5. See particularly chapter 4, “On Not Writing Off Scientific Realism.”
6. Hilary Putnam, Mathematics, Matter and Method, vol. 1 of Philosophical Papers (Cam-
bridge: Cambridge University Press, 1975), 73.
7. See Hilary Putnam, “Wittgenstein: A Reappraisal,” in De Caro and Macarthur, Phi-
losophy in an Age of Science, chap. 28.
8. Putnam, “Science and Philosophy,” in De Caro and Macarthur, Philosophy in an Age of
Science, 44.
Unauthenticated
216 LOOKING BACK
One definition comes from Wilfrid Sellars’s essay “Philosophy and

the Scientific Image of Man.”9 “The aim of philosophy,” he wrote, “is
to understand how things in the broadest possible sense of the term
hang together in the broadest possible sense of the term.” And the other
is Cavell’s famous characterization of philosophy as “The Education of
Grownups.” Neither of these characterizations implies that any part of
our language, in or out of science, is only “quasi-realistic” or that all
philosophical questions are fated to be either answered or dissolved by
future science.
Instead, then, of trying to make large pronouncements as to where
philosophy should be going, as I did in 1997, what I shall do today is
look back on one part of my own philosophical career, and thereby both
recall a part of the “50 years of philosophy” to which this series is de-
voted and connect that part with questions—particularly about percep-
tion and cognition—that occupy me at the moment.
Twin Earth and Brain in a Vat

I long ago discovered that what most undergraduates know about my
philosophy is two of my thought experiments: the Twin Earth thought
experiment from “The Meaning of ‘Meaning’,”10 and the Brain in a
Vat thought experiment from Reason, Truth and History.11 I shall talk
about the connection between the two as a way of leading into the ques-
tions that, as I mentioned, occupy me presently. Some undergraduates
also hear about my “functionalism” (a position in the philosophy of
mind), but rarely if ever, in my experience, about my modal-logical in-
terpretation of mathematics, my changing positions in the philosophy
of quantum mechanics, my writings in ethics and economics, and other
9. Wilfrid Sellars, “Philosophy and the Scientific Image of Man,” in Frontiers of Science
and Philosophy, ed. Robert Colodny (Pittsburgh: University of Pittsburgh Press, 1962), 35–
78; repr. in Science, Perception and Reality (1963; Atascadero, Calif.: Ridgeview, 1991), quota-
tion from p. 37 of the 1991 edition.
10. Hilary Putnam, “The Meaning of ‘Meaning’,” in Language, Mind and Knowledge, ed.
Keith Gunderson (Minneapolis: University of Minnesota Press, 1975), 131–193. Collected
in Hilary Putnam, Mind, Language and Reality, vol. 2 of Philosophical Papers (Cambridge:
Cambridge University Press: 1975), 215–271.
11. Hilary Putnam, “Brains in a Vat,” in Reason, Truth and History (Cambridge: Cam-
bridge University Press, 1981), chap. 1.
Unauthenticated
topics, and I shall not talk about these other interests of mine. So let
me describe those two well-known thought experiments of mine.
Twin Earth
Although many of you know about Twin Earth, I need to review the
argument in which that thought experiment figured. Imagine a planet
like earth— call it “Twin Earth”—on which the liquid that fills the lakes
and rivers, that people drink, is not H2O but a different compound,
XYZ, with similar superficial characteristics. The Twin Earthers were
supposed to be our doppelgängers; some of them even spoke English.
Also, I imagined the year to be 1750, before the chemical composition of
either water or twater (Twin Earth “water”) was known. The English-
speaking Twin Earthers naturally called twater “water” (and the French-
speaking ones called it “eau,” and the German-speaking ones called it
“Wasser”). The reaction of the majority of people who have considered
this thought experiment is that upon learning that Twin Earth “water”
doesn’t consist of H2O at all, we Earthers would and should say “it isn’t
really water.” The word “water” has a different meaning on Earth and
on Twin Earth. Twin Earthian Oscar’s word “water” and Earthian Os-
car’s word “water” are homonyms, but not synonyms. They do not have
the same meaning—not even if Earthian Oscar and Twin Earthian
Oscar happen to be microphysical duplicates!
It is irrelevant that there is no such physically possible compound
XYZ, because we can imagine that neither Oscar nor Twin Oscar know
this esoteric fact. (Indeed, I am not sure I myself know such a thing,
unless having been told that this is so by one former chemist counts as
“knowing.”) The question is what we would say under perfectly imag-
inable circumstances, not what is chemically or physically possible.
Moreover, the example can easily be modified to avoid this (irrelevant)
objection, and that is what I did in a later paper:12 just imagine that
Twin Earth water consists of 50 percent water and 50 percent a taste-
less and nonpoisonous liquid that does not quench thirst (D2O, “heavy
12. Hilary Putnam, “Meaning Holism,” in The Philosophy of W. V. Quine, ed. L. E.

Hahn and P. A. Schilpp (La Salle, Ill.: Open Court, 1986), 405–431; repr. in Realism with
a Human Face, ed. James Conant (Cambridge, Mass: Harvard University Press, 1990),
278–302.
Unauthenticated
218 LOOKING BACK
water,” is such a liquid). Of course, Twin Earthers, if their bodies are

like ours, would have to drink more of their “water” per day than we
drink, but it is plausible to suppose that ordinary speakers do not no-
tice this, and that they would not regard this fact as relevant to the
meaning of the word “water” if they did notice it. (It used to be the case
that Brits ate ten times as much chocolate per year as Americans, but
no one regarded this as a difference in the meaning of the word “choco-
late.”) And it is not essential to the Twin Earth thought experiment
that Oscar and Twin Earth Oscar be microphysical duplicates, or that
their brains be microphysical duplicates: it is enough if their brains are
in sufficiently similar states with respect to those systems that enable
the appropriate use of words.
The conclusion of this argument is that what our words refer to is,
in general, not determined merely by what goes on in our brains, or,
more broadly, inside our bodies: the difference that makes a difference
between us and the Twin Earthers lies at the “other end” of the causal
chains connecting us with the stuff we drink. Reference depends on
causal connection to the extra- bodily environment. (In “The Meaning of
‘Meaning’ ” I gave other reasons that this is the case, including the so-
cial nature of language.)
Brain in a Vat
The similarity of my Brain in a Vat thought experiment and the movies
The Matrix and The Matrix Reloaded has often been remarked. In The
New Yorker, Adam Gopnik wrote,
It was, improbably, the Harvard philosopher and mathematician

Hilary Putnam who, a couple of decades back, proposed the es-
sential Matrixian setup: a bunch of brains in a vat hooked up to a
machine that was “programmed to give [them] all a collective hal-
lucination, rather than a number of separate unrelated hallucina-
tions.” Putnam used his Matrix to make a tricky argument about
meaning: since words mean what they normally refer to within a
community, a member of the vatted-brain community might be
telling the truth if it said it was looking at a tree, or, for that matter,
at Monica Bellucci. That’s because the brains in that vat aren’t
really speaking our language. What they are speaking, he said, is
Unauthenticated
“vat-English,” because by “a tree” they don’t mean a tree; they

mean, roughly, a tree image.13
The purpose of this argument was to show that a scientifically up-

to-date and less solipsistic version of the Cartesian claim that there
might be no external world at all, because I might be a disembodied
mind deceived by an evil demon, could be refuted. And one premise in
the argument was a generalization of the “semantic externalist” conclu-
sion of the Twin Earth argument. That argument showed that, in the
case of natural kind terms such as “water,” “gold,” and “elm,” reference
depends on causal connection with instances of the natural kind, either
directly (i.e., on the part of the speaker herself) or indirectly (i.e., via
communication with speakers who have communicated with speakers
who have communicated with speakers who . . . leading back eventu-
ally to speakers who have causal connection of the right kind with
instances of the natu ral kind). Of course, one can also refer to things
no one has interacted with, and things that may not exist; for example,
we can refer to the nearest intelligent extraterrestrials. This is what is
called “reference by description,” but reference by description depends
ultimately on “direct” reference. Thus we can generalize the conclu-
sion of the Twin Earth argument as follows: reference to objects in the
external world depends on information carrying causal connection with
those objects or at least with objects that have properties in terms of which the
objects referred to can be described, and the causal connection has to involve
the properties in question. The Brain in a Vat argument depends on this
generalization of the conclusion of the Twin Earth Argument.
This is obviously a controversial premise, and it has been controversial
almost as long as there has been philosophy, and I will say something
about this shortly. But first let me just state the argument. Assuming this
premise (which I shall call “semantic externalism”), the Brain in a Vat
argument is simply this: to refer, say, to cabbages, or to vats, or to rabbits,
mountains, or whatever it is necessary to have causal connection with
them, or with objects with related properties. But according to the Brains
in a Vat hypothesis, the experiences of the Brains in a Vat are just the
experiences they would have if the world they seem to experience were a
13. Adam Gopnik, “The Unreal Thing: What’s Wrong with The Matrix,” The New
Yorker, May 19, 2003, www.newyorker.com /magazine/2003/05/19/the-unreal-thing.
Unauthenticated
220 LOOKING BACK
real world, that is, they experience just what non-envatted people would
be experiencing. Because it is part of the scenario that the Brains were
always envatted, so there was never a time when they had information
carry ing causal interaction with things in the “real world” and their
properties, it belongs to the Brain in a Vat hypothesis that people in the
vat have no information carrying causal interaction with real cabbages or
real rabbits or real mountains, or even real vats (they are causally depen-
dent on one vat, of course, as they would die if it broke, but they do not
receive information from it as long as the system functions as it is de-
signed to). Nor do they have information carry ing causal interactions
with such properties as one thing’s being inside another, or being to the
south of another, or having a certain color or weight or whatever. Thus,
if “semantic externalism” is right, then the Brains in a Vat aren’t able
to refer to cabbages or kings, or ships or shoes or sealing wax.
This means that although the Brains in a Vat use the same words we
do, or at least the same vocables go through their heads, and their
Joycean “stream of consciousness” may be the same as that of a non-
envatted person, their words do not have the same meanings as those of a
non-envatted doppelgänger. That was, remember, the conclusion of the
Twin Earth argument. Thus, they are not speaking the same language
as a non-envatted doppelgänger. Let us call the language of the Brains
in a Vat (or of the ones who call their language “English”) Vat English.
The question as to whether I am a Brain in a Vat is the question as
to whether the language I speak (I call it “English, of course) is Vat
English.
But in my language, the word “cabbage” refers to cabbages. How do
I know that?
There is a fundamental logical principle, basically stated by Tarski,
according to which in one’s own language any term T refers to, pre-
cisely, Ts.14 For example, in my language, the word “cabbage” refers to
cabbages. For me to say in my language, the one I am speaking (or
writing or thinking in) now, that the word “cabbage” does not refer
to cabbages in that very language would be simply self-contradictory.
14. Tarski’s disquotation principle (the famous convention T) was stated for truth rather
than reference, but the above disquotation principle for reference follows from Tarski’s
analysis as well; formally, for all x, x satisfi es the open sentence “Tx” if and only if Tx (“satisfac-
tion” is the reference relation, in Tarskian semantics).
Unauthenticated
But we just saw (and again, I am saying all this in my own language)
that in Vat English the word “cabbage” does not refer to cabbages. From
these two facts together, it follows at once that my language is not Vat
English. Hence, I am not a Brain in a Vat.
Instead of discussing the lecture I am going to finish giving, we could
of course spend the rest of our time discussing the pros and cons of this
argument, why it seems so counterintuitive, and why nonetheless so
many people now accept semantic externalism, including this conse-
quence. I will just say one thing: the argument is not meant to convince
a skeptic. Skeptics can always reject the premise; they can reject se-
mantic externalism. On the other hand, if the skeptic assumes that
semantic externalism is false, the skeptic assumes something I do not
believe and do not have to believe. Thus, even if the argument does
not show that Cartesian skepticism is wrong, it does show that Des-
cartes, or his imagined contemporary successor who talks about Brains
in a Vat rather than disembodied minds, does not have a proof I must
accept that skepticism is right; or at least not a proof based on the pos-
sibility that we are all Brains in a Vat, because I don’t have to accept
the possibility that I am a Brain in a Vat.
A Word about the History of Externalism

Externalism, in the form in which I presented and defended it in “The
Meaning of ‘Meaning’,” is a contemporary position, but the debate
between externalists and those who think we are innately capable of
referring to things regardless of whether we have had any causal inter-
actions with them or with other things with related properties goes
back at least to the times of Plato and Aristotle. Descartes’s skeptical
scenario (I am a disembodied mind deceived by a demon) assumes that
a mind can refer to any sort of object even if the owner of that mind
has never had any causal connection with the external world at all. (As
I like to express it, Descartes thought we are semantically omnipotent.) It
is plausible to interpret Plato as having thought the same thing. On the
other side, in Aristotle’s philosophy of mind, as I interpret it,15 the only
15. “Aristotle’s Mind and the Contemporary Mind,” in Aristotle and Contemporary Sci-
ence, ed. Demetra Sfendoni-Mentzou, Jagdish Hattiangadi, and D. M. Johnson (New York:
Peter Lang, 2000), 1:7–28.
Unauthenticated
222 LOOKING BACK
properties we can conceive of are such that either we have experienced

instances of them, or have experienced other properties in terms of
which we can describe the properties we are thinking of. As I interpret
him, Aristotle was an “externalist.”
What makes externalism attractive today, in my eyes at least, is its
consonance with a naturalistic way of thinking about human beings in
the world. The idea of us as beings with an intrinsic ability to think
about things whose properties we may neither have experienced nor
detected with the aid of scientific instruments nor have the ability to
define in terms of properties we have interacted with in one of these
ways makes reference a magical power, and I believe that referring is a
perfectly natural affair.
Externalism and Perception

The Brain in a Vat argument was an opening chapter in a book (Reason,
Truth and History) concerned with issues about metaphysical realism
(but that is another story). I no longer accept the conclusion of that
book, which was that the Brain in a Vat argument refutes metaphysical
realism.16 But today I want to apply the semantic externalism that, as I
have been explaining, figured both in that argument and in the earlier
Twin Earth argument to a very dif ferent subject, the subject of percep-
tion (the subject I am working on now). But before we think specifically
about perception, let me smooth the transition by applying the causal
connection principle we have been talking about to the philosophy of
mind. One such application has already been made by a great twentieth-
century phi losopher, Donald Davidson: the famous (or notorious)
“Swampman” thought experiment.17
Davidson’s thought experiment is very simple. (I shall modify it so
that it is compatible with contemporary physics.) Imagine a swamp
16. See Putnam, “Corresponding with Real ity,” in De Caro and Macarthur, Philosophy in
an Age of Science, 72–90.
17. Donald Davidson, “Knowing One’s Own Mind,” in Subjective, Intersubjective, Objec-
tive (Oxford: Clarendon Press: 2001), 15–38; originally published in Proceedings and Ad-
dresses of the American Philosophical Association 60 (1987): 441–458. Davidson later expressed
embarrassment at his own thought experiment in his article “Interpretation: Hard in
Theory, Easy in Practice,” in Interpretations and Causes: New Perspectives on Donald David-
son’s Philosophy, ed. Mario De Caro (Dordrecht, the Netherlands: Kluwer, 1999), 31–44. I
see need for the embarrassment.
Unauthenticated
somewhere full of warm water and vegetation and thus with atoms, or
at least atomic particles, of which in principle a human being could be
formed. Suppose that by one of those almost infi nitely improbable but
not impossible quantum-mechanical scenarios,18 the particles in the
swamp jump together and form a human being with a completely false
but reasonably consistent set of memories, as of a human life, including
some memories that make sense of the fact that he is now in this swamp.
Swampman, of course, might be on Earth, he might be on Twin Earth,
or any other planet that has the right sort of atmosphere. No matter
what the “water” on that planet actually is, Swampman, if he is a dupli-
cate of an Earthean human being, will say that water is H2O. But the
question is, “What do Swampman’s words refer to?”
If semantic externalism is correct, the answer is that Swampman’s
words do not refer to anything. And if his words do not (yet) have any ref-
erence, Swampman has no concepts; and, a fortiori, anything he says is
meaningless. Of course, after he has been in existence for a while, and
has had causal connection with the liquid in the swamp and calls it
“water,” with the grass in the swamp and calls it “grass,” perhaps his
words acquire meaning, but initially, Swampman has no concepts, and hence
no thoughts. Some philosophers of mind would go further and say he
has no consciousness; I don’t see that conclusion as following, but in
any case, it follows from our externalist (or “causalist”) premise that
Swampman has no concepts. To have concepts it is necessary to have
appropriate causal connection with an environment. Semantic exter-
nalism implies externalism about the mind; if to have a mind is to have
thoughts, then to have a mind it isn’t sufficient to have the right go-
ings-on in the brain and the rest of the body; to have a mind you have
to be hooked up to an environment in the proper way, or at least to have
a mind that can think about an external world, you have to have causal
interactions that extend into the environment. One might call this an
anti-solipsist conclusion: If externalism is right, pace Descartes, an isolated
disembodied mind would have no thoughts about the world at all, not
even false thoughts. In Kantian language, the pseudo-thoughts of Des-
cartes’s isolated mind are an empty play of representations and not
thoughts at all.
18. Davidson, “Knowing One’s Own Mind,” has this all happening as a result of light-
ning bolts, which is, physically speaking, impossible.
Unauthenticated
224 LOOKING BACK
To put it another way, one suggested many years ago by John Mc-
Dowell, if meanings aren’t in the head (which is what I said in “The
Meaning of ‘Meaning’ ”), then the mind isn’t in the head either;19 that
is to say, the mind isn’t a thing, it’s a set of abilities— abilities that in-
volve the external world from the beginning.
Especially at the University of East Anglia, it is appropriate to say
at once that such a view was arrived at by Wittgenstein long before
I thought about this question, but Wittgenstein’s arguments, while
fascinating, are problematic for me. The Twin Earth argument
was my own way of arriving at the conclusion that to think of the
mind in the traditional epistemological way, as something that is
isolated from the world initially and that has to get past a kind of
“veil” of its own sensations to get to the world, is wrongheaded from
the start.
Enter William James

So far I have argued that the possession of concepts that refer to things
in the external world depends on world-involving transactions; without
those world-involving transactions, we have only an empty play of
representations and not genuine concepts. That is a statement about con-
ceptualization, and I said I was going to talk about perception. But the
connection between conceptualizing and perceiving was already pointed
out by William James, in a remark that has influenced me for many
years. The remark was that (in the case of “a presented and recognized
material object”) “sensations and apperceptive ideas fuse here so inti-
mately that you can no more tell where one begins and the other ends,
than you can tell, in those cunning circular panoramas that have lately
been exhibited, where the real foreground and [where] the painted can-
vas.”20 As James indicated, the best example of this phenomenon of
“fusion” is the recognition, say, the seeing, of an object, when one sees
19. John McDowell, “Putnam on Mind and Meaning,” in Philosophical Topics 20 (1992):
35–48. McDowell wrote, “The moral of Putnam’s thought for the nature of the mental
might be, to put it in his terms, that the mind . . . is not in the head either” (36).
20. William James, “Does Consciousness Exist,” in The Works of William James: Essays in
Radical Empiricism, ed. Frederick Burckhardt and Frederick Bowers (Cambridge, Mass.:
Harvard University Press, 1976), 3–19, quotation from p. 16.
Unauthenticated
it not only in the minimal sense of “see,”21 in which a cat might be said
to see a television set, without having a concept or a protoconcept of a
television set,22 but sees it in the sense in which to see a television set is
to see it and know that what one is seeing is a television set, the case of a
full-blown apperception. In such a case, my perceptual experience has
a sensational component—it includes what some philosophers call
“qualia” that could also be part of the experience of a child who does
not yet have the concept of a television set—but it also includes in a way
that is real but hard to describe the fact that it looks like a television set,
the fact that the concept “television set” is in play.
But now we can apply externalism: if to see a television set, in the sense
in which a master of the relevant part of a language can see a television
set, requires concept-possession, and concept-possession requires causal
connection to external things, then seeing requires causal connection
to external things—not just in the trivial sense, that we don’t call it
“seeing” if no light rays from external objects are striking the eye, but
in a much more complicated sense: seeing, in the sense of perceiving by
means of sight, requires a history of language acquisition and language use.
(In fact, according to another part of the argument of “The Meaning
of ‘Meaning’,” it also requires a history of interaction with other
speakers, but I do not have time to go into that part of the argument
today.) In the demanding sense, Swampman doesn’t “see” anything
when he appears on the scene; he doesn’t perceive anything.
How this “externalist” picture contrasts with the traditional episte-
mological account was already indicated earlier, when I said that on that
account the mind has to get past a “veil” of its own sensations to get to
the world. I could also have said that in the traditional epistemological
picture, knowledge is entirely based on experiential qualia, in the sense
that every thing we know about the world has to be inferred by the sub-
ject from those qualia, from the phenomenal character of his or her
sensations. Perhaps the best statement of that view in the twentieth
century is Russell’s classic The Problems of Philosophy. In The Problems of
Philosophy, everything we know about the external world is inferred from
21. On the senses of “see,” see Ruth Anna Putnam, “Seeing and Observing,” Mind 78
(1969): 493–500.
22. On the difference between the concepts of humans who have mastered a language
and the protoconcepts, as I termed them, of animals, see Hilary Putnam, Renewing Philos-
ophy (Cambridge, Mass.: Harvard University Press, 1992), 28–31.
Unauthenticated
226 LOOKING BACK
our “sense data,” and those sense data are all in our own “private spaces.”23
If the externalist picture I have been defending is right, this traditional
picture of us as starting from data available in our individual “private
spaces” and inferring that there is an external world at all is hopelessly
wrong. If all we have is our own private data, then we are Swampmen.
Swampman has sensations, but he cannot infer the existence of an
external world—he cannot so much as conceive of an external world.
Swampman has no concepts, and having no concepts—if we accept what
James said about “fusion”—lacking “apperceptive ideas” to fuse with his
sensations, means having no perceptions (or, in a dif ferent terminology,
no “apperceptions”). Knowledge is not the product of interactions with
mental objects in a private space, but of transactions between a human
organism and an environment, including a history of transactions with
a linguistic environment.
23. Bertrand Russell, The Problems of Philosophy (Oxford: Oxford University Press, 1912), 29.
Unauthenticated
Acknowledgments
Many of the papers reprinted here have been lightly edited to correct
minor mistakes, avoid needless repetition, and, in cases where the papers
were originally given as talks, remove indications of occasional address.
Minor additions or deletions to several papers for purposes of clarifica-
tion have been made. However, in cases where we have added a whole
paragraph indicating a significant change in view, this has been indicated
in the text by putting it in brackets and giving the date of the remark,
when pertinent. We thank the publishers and journals that have granted
us the permission to republish some of the papers contained in this book.
1. “Naturalism, Realism, and Normativity.” From Journal of the

American Philosophical Association 1, no. 2 (2015): 312–328.
Published online 19 June 2015 (journals.cambridge.org/apa).
Copyright © 2015 American Philosophical Association.
2. “On Bernard Williams’s Philosophy as a Humanistic
Discipline.” From Philosophy 76, no. 4 (October 2001): 605–614.
(Published with the title: “Reply to Bernard Williams’ Philos-
ophy as a Humanistic Discipline.”) Published by Cambridge
University Press on behalf of Royal Institute of Philosophy.
Reproduced by permission of Cambridge University Press,
Copyright © 2001 Royal Institute of Philosophy.
3. “What Evolutionary Theory Doesn’t Tell Us about Ethics.”
From Understanding Moral Sentiments: Darwinian Perspectives?
227
Unauthenticated
228 Acknowledgments
ed. Hilary Putnam, Susan Neimann, and Jeffrey Schloss (New

Brunswick/London: Transaction Publishers, 2014), 203–211.
(Published with the title “Not Very Much.”) Copyright © 2014
Transaction Publishers, New Brunswick, New Jersey.
4. “Sosa on Internal Realism and Conceptual Relativity.” From
Ernest Sosa and His Critics, ed. John Greco (Malden, Mass.:
Blackwell, 2004), 233–248. Copyright © 2004 Blackwell
Publishing Ltd.
5. “Richard Boyd on Scientific Realism.” From Reading Putnam,
ed. M. Baghramian (London: Routledge, 2013), 95–100.
(Published with the title “Comment on Boyd.”) Reproduced by
permission of Taylor & Francis Group, LLC. Copyright ©
2013 Maria Baghramian, for selection and editorial matter.
Copyright © 2013 Hilary Putnam, for contribution.
6. “Hans Reichenbach: Realist and Verificationist.” From Future
Pasts: The Analytic Tradition in Twentieth- Century Philosophy, ed.
Juliet Floyd and Sanford Shieh (Oxford: Oxford University
Press, 2001), 277–287. Reproduced by permission of Oxford
University Press, USA. Copyright © 2001 Oxford University
Press, Inc.
7. “Between Scylla and Charybdis: Does Dummett Have a Way
Through?” From The Philosophy of Michael Dummett, ed. R. E.
Auxier and L. E. Hahn (Chicago: Open Court, 2007), 155–167.
Copyright © 2007 by The Library of Living Philosophers.
8. “When ‘Evidence Transcendence’ Is Not Malign.” From
Journal of Philosophy 98, no. 11 (November 2001): 594–600.
(Published with the title “When ‘Evidence Transcendence’ Is
Not Malign: A Reply to Crispin Wright.”) Copyright © 2001
The Journal of Philosophy, Inc.
9. “Sensation and Apperception.” From Consciousness and Subjec-
tivity, ed. Sofia Miguens and Gerhard Preyer (Berlin: Ontos
Verlag, 2012), 39–50. Copyright © 2012 Ontos Verlag.
10. “Perception without Sense Data.” Derives from a lecture read
at the 5th International Lauener Symposium held in Bern,
Switzerland, in June 2012, during which Hilary Putnam was
awarded the Lauener Prize for an Outstanding Oeuvre in
Analytical Philosophy for that year. The essay will first be
published in Themes from Putnam, edited by Michael
Unauthenticated
Acknowledgments 229
Frauchiger (series: Lauener Library of Analytical Philosophy,

vol. 5) Berlin/Boston: Walter de Gruyter.
11. “ ‘Naïve Realism’ and Qualia.” Will appear in Themes for Ned
Block, ed. Adam Pautz and Daniel Stoljar (Cambridge, Mass.:
MIT Press, forthcoming).
12. “The Development of Externalist Semantics.” From Theoria 79,
no. 3 (2013): 192–203. Published by John Wiley and Sons.
Copyright © 2013 Stiftelsen Theoria.
13. “Sixty-Five Years of Philosophy: A Participant’s Thoughts and
Experiences.” Lecture given as one of the Royal Institute of
Philosophy Public Lectures on 50 Years of Philosophy, at the
University of East Anglia on December 5, 2013. Published for
the first time in this collection.
Unauthenticated
Unauthenticated
Index
Absolute conception of the world, 44, 45, Austin, John, 152, 154, 155
47, 48–49, 50–52 “Avoiding the Myth of the Given”
Absolutism, 88, 89 (McDowell), 147, 149
Action, perception and, 165 Ayer, A. J., 155
Adequacy, Tarski’s criterion of, 28–30
Aesthetic judgments, science and, 61 Being Realistic about Reasons (Scanlon), 42
Aesthetic values, 16–17 Beliefs, perceptual experience and,
Affordances, 211 141–142, 143, 183–184; apperceptions
Altruistic behav ior, 56–57, 58 and, 149, 193–194
Amodal apperception, 150–151, 193 Berkeley, George, 3, 117, 159, 180
The Analysis of Mind (Russell), 176 Blackburn, Simon, 214
“The Analytic and the Synthetic” Block, Ned, qualia and, 156, 157–158, 163,
(Putnam), 201, 202, 210 169–170, 171, 172, 174, 179–180, 188,
Analyticity, 79 194–195, 211
Analytic metaphysics, 214 Boyd, Richard, 21, 90–96
Anscombe, Elizabeth, 193 Brain in a Vat thought experiment,
Antirealism, 1–2, 3, 115, 117, 119–121, 216–217, 218–221, 222
125–126; alternative to, 126–127 Bronowski, Jacob, 16
Apperception, 184, 225–226; amodal, Burge, Tyler, 23, 28, 39, 40
150–151, 193; beliefs and, 193–194;
conceptual shaping of, 149; Kantian Carnap, Rudolf, 38, 77, 93; conceptual
model of perception and, 162–163; scheme, 81–84; reduction sentence and,
linguistic ability and, 185, 189; qualia 131; Reichenbach and, 100, 101–102, 103,
and, 192–194; sensations vs., 146–148 111; status of unobserved objects and,
A priori– a posteriori dichotomy, 59–60 107–108, 109, 117
Aristotle, 5, 57–60, 59, 61, 154, 221–222 Cartesian skepticism, 141–142, 186–187, 221
“Aristotle after Wittgenstein” (Putnam), Carus Lecture, 69
76 Case, Jennifer, 83–84, 85
Artifacts, 88 Casper, Philip, 21
Assertibility, 134–135 Categorical imperative, 41–42
231
Unauthenticated
232 Index
Causal connection principle, 218, 219, Davidson, Donald, 9, 46, 50, 114, 203;
222–224, 225–226 conceptual relativism, 68–69; justifica-
Causation, 15; reference and, 70–74, 218, 219 tion and experience, 143–144, 150;
Cavell, Stanley, 92, 216 reference and, 52, 53; thought experi-
Charity, principle of, 46, 53 ment, 222–223; truth and, 53–54
Chomsky, Noam, 63–64, 143, 199 De Caro, Mario, 21, 22, 24
Churchland, Paul and Patricia, 214 Defi nitions, 202, 207, 210
Churchman, C. West, 200 Deflationary theory of truth, 35–36, 37–38,
Cognitive equivalence, 14 40, 94
Color, perception of, 140–141, 159, Deflationism about reference, 40
170–171, 177, 211 Democracy, 59, 63
Commonsense realism, 4–5, 7–9 Dennett, Daniel, 214
Commonsense worldview, 3, 12 Descartes, René, 221, 223
Computational functionalism, 166 Descartes: The Project of Pure Enquiry
Conant, James, 141, 186 (Williams), 45–46, 70
Conception: causal connection with Descriptive statements, 96
environment and, 218, 219, 223, 224, Devitt, Michael, 2, 71, 118, 119
225–226; change in sensation and, Dewey, John, 22, 41, 91–92
148–150; innate, 143; perceptual Dewey Lectures, 25, 91, 122, 134, 152, 153,
experience and, 139–141, 143, 144–146, 154, 156, 166
183, 192–193 Dialogue on the Two Main Systems of the
“The Concept of Truth in Formalized World (Galileo), 16–17
Languages” (Tarski), 27–28, 30, 31, 34 Disjunctivism, 155–157, 169, 174–175
Concept-possession, perception and, 225 Disquotation, 35, 37–38; truth and,
Conceptualization: of impressions, 94–95
189–191; perception and, 224–226 Dreben, Burton, 79
Conceptual pluralism, 14 Dretske, Fred, 157, 179, 211
Conceptual relativity, 68–69, 77–80, Duhem, Pierre, 7
82–83, 88, 89, 153–154 Dummett, Michael: idealism and, 123–124;
Conceptual schemes, 81–88 justificationism and, 115, 116, 121–122,
Conditional Fallacy, 135 125–126, 127; realistic semantics and, 4,
“Confessions of a Hard- Core, Unsophisti- 91, 116, 121; redefi ning “realism” and,
cated Metaphysical Realist” (Maudlin), 116–119
25, 26
“Consciousness, Accessibility, and the Eddington, Arthur, 4, 5
Mesh between Psychology and Edelman, Gerald, 173, 174
Neuroscience” (Block), 170, 174 Einstein, Albert, 202–203
Constructive empiricism, 7 Eliminationism, 45–46
Content, thoughts and, 191–194 Eliminativism, 86–88, 89
Conventionalism, 200–201 Empiricism, 7, 141, 142, 144, 156–157;
Conventions, 79 minimal, 141, 144, 183; model of
Convention T, 27–28, 29–30, 35–36 perception, 158, 159–161, 165
“Cornell Moral Realists,” 22 Enlightenment, 59, 60
Correspondence theory of truth, 34–35, Epistemic situation, 123–124, 129, 135
94–96 Epistemic values, 16–17, 61
Counterfactuals, 122, 124, 148–149 Epistomological realism, 4
The Crisis of European Sciences and Equality, ethics and, 59–60
Transcendental Phenomenology (Husserl), Ethical judgments, 52
6–7, 9–10 Ethical relativism, 60
Critique of Pure Reason (Kant), 92, 146 Ethics: cultural history of, 57–62;
Cross-category identity, 85, 86 evolutionary theory and, 56–64;
Cubical world argument, 103–107 normativity and, 15, 40–42; Putnam on,
Cultural evolution, ethics and, 62–63 92; Williams and, 54–55
Unauthenticated
Index 233
Ethics and the Limits of Philosophy (Williams), Having the World in View (McDowell),
70 188–189
Ethics (Dewey & Tufts), 41 Hegel, G. W. F., 154
Ethics without Ontology (Putnam), 92 Helmholtz, H. L., 179, 185
Evans, Gareth, 149–150 Hempel, Carl, 131
Evidence-transcendence, 128–135 Hierarchy theory, 86
Evolutionary theory, ethics and, 56–64 Hilbert, David, 204
“Exist,” as univocal, 93–94 Hinton, J. M., 155, 156, 174–175
Experience. See Perceptual experience Hornsby, Jennifer, 9
Experience and Prediction (Reichenbach), Horwich, Paul, 38
100, 102–103, 107, 111, 120 “How Not to Talk about Meaning”
“Explanation and Reference” (Putnam), (Putnam), 205
117 Hume, David, 159, 180
Externalism, 170, 178, 180–182, Husserl, Edmund, 6–7, 9–10, 147–148
192, 201–202, 205; history of, 221–222; Hyperscientific realism, 4–6, 7, 9
perception and, 210–212, 222–224, Hypothetico-deductive method, 110
225–226. See also Semantic externalism
Idealism, 117–118, 123–124
Fallibilism, 12 Idealization theory of truth, 122, 124
Farber, Marvin, 10 Ideal theory, 26
Feigl, Herbert, 200 The Idea of Justice (Sen), 57
Feyerabend, Paul, 201 Identity of phenomenal character, 172–173
Field, Hartry, 28, 30–31, 33, 39 Immaterialism, 180–181
Fitch propositions, 130–134 Impressions: beliefs and, 183–185, 193;
Fodor, Jerry, 64, 71–72, 160, 200, 203, 204, conceptualization of, 189–191; Reichen-
214 bach on, 103–105
Fraassen, Bas van, 7 Indeterminancy thesis, 39
Frege, Gottlob, 202 Inductive logic, 108–111
“From Quantum Mechanics to Ethics and Inference: empiricist model on, 159–160;
Back Again” (Putnam), 90 of existence of things, 105–107; sense
Functionalism, 181–182, 194, 216; data and, 178
computational, 166; liberal, 146, Information, qualia and, 211
165–168, 182, 185–186, 194 Innate language, 63–64
Intention: apperception and, 193; reference
Galileo, 5–6, 16–17 and truth and, 71–72
Gandhi, Mahatma, 59 Intentionalists, 189
Gibson, Etienne, 165, 167, 211 Interface conceptions, 128–133
Gilson, Etienne, 25 Interface-realism, 129
Glymour, Clark, 71, 102 Internalism, 180–181
Gödel, Kurt, 27 Internal realism, 1, 2–3, 25–26, 70, 71, 124,
Gold, meaning of, 205–207 152–153
Goodman, Nelson, 91 Intrinsic properties, 75–76
Gopnik, Adam, 218–219 “Is Semantics Possible?” (Putnam), 202,
Government, as social contract, 59, 203, 208
61–62
Grünbaum, Adolf, 200–201 Jacobson, Hilla, 148, 158, 183
James, William, 91, 146–147, 148, 154, 169,
“A Half Century of Philosophy Viewed 178, 224–226
from Within” (Putnam), 213 Judgments, 16, 50, 52, 61
Hardin, Larry, 177 Justification: Davidson and, 143–144, 150;
Harman, Gilbert, 157, 179, 180 reliabilism and, 187–188
Harris, Zelig, 199 Justificationism, 115, 116, 121–122, 124,
Hauser, Marc, 58, 63–64 125–126, 127
Unauthenticated
234 Index
Kant, Immanuel, 41, 92, 146, 147, 154, 194 and, 159, 162–164, 165; on meaning, 224;
Kantian model of perception, 159, 162–165 ostensive defi nition and, 142–143;
Kantian skepticism, 141–142, 186, 187–191 perceptual experience and beliefs and,
Kaplan, David, 208 141–142, 143; philosophy of perception
Katz, Jerrold, 200, 203–204 requirements and, 182–184; Sellars and,
Kinetic energy, reference and, 201–202 195–196; thoughts possessing content
Knowledge: qualia and, 192–193, 225–226; and, 191–194
reliabilism and, 187 Meaning: as correspondence, 103;
Kripke, Saul, 114, 208, 209–210 deflationism and, 37–38; of expression,
93; linguistics and, 199–200; probability
Land, Edwin, 158 theory of, 100, 101–103, 107; truth
Language: concept of truth in, 27–30; theory of, 103; of words, 204–209,
conceptual relativity and, 77–80; 217–221. See also Reference; Semantic
conceptual schemes and, 81–88; externalism; Semantics
difference between “egocentric” and Meaning and Necessity (Carnap), 200
“usual,” 111–113; liberal functionalism Meaning and the Moral Sciences (Putnam),
and, 167; natu ral vs. optional, 83–86; 99, 108–109
normative vs. descriptive, 43; Reichen- “Meaning and Truth” (Davidson), 114
bach and choice of, 100; social nature of, “The Meaning of ‘Meaning’ ” (Putnam),
121, 125–126 181, 182, 203, 207, 208. 210, 216, 218,
Language organ, 64 221, 224, 225
Lauener Prize, 152 Meinong, Alexius, 3
Lenin, V. I., 118–119 “The Mental Life of Some Machines”
Lewis, C. I., 108 (Putnam), 168
Lewis, David, 141 Mereological sums, 74–75, 77–78, 80, 84,
Liberal functionalism, 146, 165–168, 182, 93–94
185–186, 194 Metaphysical realism, 24–27, 71, 73–74, 80,
Liberal naturalism, 9–15, 21–24, 40, 41, 42 124, 154, 222
Life-world, 6–7 Metaphysical truth, physics and, 44–45, 47
Linguistic ability, apperception and, 185, Mill, John Stuart, 110
189 Mind: internalist vs. externalist theories of,
Local perspectives, 214 180–181; philosophy of, 180–182, 210,
Location/placement problem, 10, 13 221–224
Locke, John, 205, 206–207 Mind and World (McDowell), 139, 140,
Logical positivism, 190, 201 145, 156, 182–183, 184–185, 186, 188,
Logic (Dewey), 91 190
Minimal empiricism, 141, 144, 183
Macarthur, David, 21, 22, 24 Minimal reliabilism, 188, 191
Maimonides, 154 Minnesota Center for the Philosophy of
Manifestation argument against realism, Science, 200–201
114–116 Model-theoretic argument, 68–70, 71–72
The Many Faces of Realism (Sosa), 77, 80 Modularity hypothesis, 159–160
Materialism, 180–181 Moore, G. E., 157, 158, 179, 180
Mates, Benson, 200 Moral grammar, innate, 63–64
The Matrix/The Matrix Reloaded (films), 218 Moral Minds (Hauser), 58, 63–64
Maudlin, Tim, 25, 26 Moral relativism, 57
Mazzoni, Jacopo, 5 Myth of the Given, 143–144, 147, 190
McDowell, John, 9, 92, 168; Cartesian
skepticism and, 186–187; conceptual Nagel, Thomas, 172–173
capacities and experience, 139–141, Naive realism, 158, 165, 176–180, 191, 196,
144–146, 147, 149; conceptualization of 211
impressions, 189–191; disjunctivism and, Naming and Necessity (Kripke), 114,
156, 169; Kantian model of perception 209–210
Unauthenticated
Index 235
Naturalism, 2, 9–15, 145–146; liberal, 9–15, Perceptual representation, reference and,

21–24, 40, 41, 42; meanings of, 21–22 40
Naturalism and Normativity (De Caro & Phenomenalism, 117–118
Macarthur), 24 “A Philosopher Looks at Quantum
Naturalism in Question (De Caro & Mechanics” (Putnam), 116–117
Macarthur), 21, 22, 24 “The Philosophers of Science’s Evasion of
Natu ral languages, 83–85 Values” (Putnam), 16
Natu ral realism, 178 Philosophical Investigations (Wittgenstein),
“The Nature of Mental States” (Putnam), 173
181 Philosophical Papers (Putnam), 1, 2
Neurath, Otto, 143, 150 Philosophy, defi nitions of, 215–216
Nicomachean Ethics (Aristotle), 57 “Philosophy and the Scientific Image of
Nihilism, 57, 60, 61 Man” (Sellars), 216
Noë, Alva, 148, 150–151, 193 “Philosophy as a Humanistic Discipline”
No-miracles argument, 8 (Williams), 44–55
Nonobjectivity of reference, 70, 73 Philosophy in an Age of Science (Putnam),
Normativity, 2, 15–17, 40–42; of semantic 215
relations, 48–50 Philosophy of mind, 180–182, 210,
221–224
Objectivism, 74 The Philosophy of Science (Boyd, Casper &
Observables, statements about, 116–117 Trout), 21
On Certainty (Wittgenstein), 161 The Philosophy of Space and Time (Reichen-
Ontological commitment, 92–94 bach), 110
Ontological realism, 4 Physics: metaphysical truth and, 44–45, 47;
“On What There Is” (Quine), 114, 170 philosophy of, 215
Optional languages, 83–86, 87–88 Pinker, Steven, 64
Ordinary language philosophy, 201 Plato, 5, 58, 59–60, 61, 154, 221
Origins of Objectivity (Burge), 23 Polish logician, 81–84
Ostensive defi nition, 142–143 Popper, Karl, 16, 28, 102
Positivism, 105–106, 107, 111–113, 118–119;
Panscientism, 214, 215 logical, 190, 201
Paradigms, 207 Pragmatic realism. See Sosa, Ernest
Peirce, C. S., 61, 91–92 Pragmatism, 91
Peircean realism, 26 Predicates, in Tarski’s theory, 30–35
Perception: of color, 140–141, 159, 170–171, Price, Huw, 24
177, 211; conceptualization and, 139–141, Primitive denotation, 33–34
224–226; empiricist model of, 158, Private Language Argument, 142–143
159–161, 165; externalism and, 180–182, Probability theory of meaning, 100,
210–212, 222–224, 225–226; Kantian 101–103, 107
model of, 159, 162–165; liberal function- The Problems of Philosophy (Russell), 159,
alism and, 165–168; McDowell’s 176–177, 178, 225–226
philosophy of (See McDowell, John); Projective complex, 106–107
naive realism, 176–180; surface-irritation Prometheus Prize Lecture, 166, 167
model of, 159, 161–162, 165, 166; thought
experiments, 216–221; without sense Qualia: apperceptions and, 184, 192–194;
data, 154–158, 163–168. See also epistemic role of, 194–196; knowledge
Apperception; Impressions; Perceptual and, 192–193, 225–226; McDowelll and,
experience; Qualia; Sensations 182–184; perceptual experience and, 211,
Perceptual experience: beliefs and, 225; Putnam’s argument against,
141–142, 183–184; conception and, 174–175; “sameness” and, 170–173. See
139–141, 143, 144–146, 183, 192–193; also Sense data
transparency of, 179–180, 189; veridical, Quasi-realism, 214, 215, 216
155–158, 170, 190, 211 Quietism, 24, 25
Unauthenticated
236 Index
Quine, W. V., 10, 22, 59, 64, 114, 168, 170, 101–103, 107; verificationism and,
187, 202; indeterminancy and, 46; 99–100, 120
ontological commitment and, 92–93; Reliabilism, 179, 183, 184; sound objection
radical translation and, 39–40; reference to, 184–186; unsound objections to,
and, 52–53, 54; surface-irritation model 186–191
and, 159, 161–162, 165, 166 Representationalism, 40
Republic (Plato), 60
Radically perspectival, causation, Rorty, Richard, 47, 48, 54, 142, 169
reference, and truth as, 70, 71–72 Rosen, Gideon, 93
Radical translation, 39–40 Rosenberg, Alex, 60
Read, Rupert, 213 Royce Lectures, 154
Realism, 2–9; internal, 1, 2–3, 25–26, 71, Russell, Bertrand, 75, 142, 155, 159,
124, 152–153; language and, 111–113; 176–177, 178–180, 185, 202, 225–226
manifestation argument against,
114–116; metaphysical, 24–27, 71, 73–74, “Sameness,” qualitative, 171–173
80, 124, 154, 222; naive, 158, 165, Scanlon, Thomas, 41, 42, 62
176–180, 191, 196, 211; quasi-, 214, 215, Scanlon, Tim, 9
216; redefi ning, 116–119; Reichenbach Schlick, Moritz, 108
and, 102, 111–113; scientific, 1, 11, Science: experimental testing and, 61;
90–96, 191, 214, 215; types of, 4–5; investigation of claims about nature, 60;
verificationism and, 99–100. See also role of aesthetic factors in, 16–17;
Antirealism single-theory convergence and, 74–75
“Realism and Reason” (Putnam), 24–25, Scientific hypotheses, 102
27, 91 Scientific naturalism, 9
“Realism and the Renegade Putnam” Scientific realism, 1, 11, 90–96, 191, 214,
(Devitt), 2 215
Realistic semantics, 114–116, 121 Scientific theories, 8
Real ity: explosion of, 82–83, 85–86, 88. Scientific worldview, 3, 12
See also Internal realism Scientism, 70–71
Reason, 16, 42 Sellars, Wilfrid, 9–10, 40, 95, 143, 188,
Reason, Truth and History (Putnam), 91, 189, 195–196, 216
119–120, 121, 134, 152, 171–172, 216 “The Semantical Concept of Truth”
Reconciliation problem, 11, 13 (Tarski), 31
Recursion theory, 86 Semantical judgments, 52
Reducible complex, 106–107 Semantical rules, 205–207
Reductionism, 118 Semantic Analysis (Ziff), 200, 203–204
Reduction sentence, 130–131, 134, 135 Semantic externalism: development of,
Reference: causation and, 70–74, 218, 219; 202–212; perception and, 222–224,
Davidson and, 53; defi ned, 96; liberal 225–226; thought experiments and, 218,
functionalism and, 167; liberal natu- 219–221
ralism and, 24; McDowell and, 189; Semantic realism, 4
meaning and, 205, 207–209; perceptual Semantic relations, normativity of, 48–50
representation and, 40; Reichenbach on, Semantics: antirealist account of, 119–121;
102–103; scientific revolutions and, foundations of, 204–205; Putnam’s early
201–202; truth and, 28, 30–31, 33–34, interest in, 199–203; realistic, 114–116,
37, 38 121; verificationist, 90–91. See also
Reference by description, 219 Meaning
Reichenbach, Hans, 168, 201; cubical Sen, Amartya, 57
world argument, 103–107; “illata” vs. Sensations: apperceptions vs., 146–148;
“abstracta,” 107–111; language and conceptual capacities and, 144–146;
difference between positivism and conceptualization and change in,
realism, 111–113; on the mind, 180; 148–150
probability theory of meaning, 100, Sense and Sensibilia (Austin), 155
Unauthenticated
Index 237
Sense data: empiricist model of perception Tocqueville, Alexis de, 63

and, 159–161; knowledge and, 176, 178, Translation: disquotation and, 38–39;
226; perception without, 154–158, radical, 39–40
163–168; surface-irritation model and, Transparency of perceptual experience,
161–162. See also Qualia 179–180, 189
Sentences: determinate truth conditions Travis, Charles, 93
of, 53–54; justification of, 143–144; truth Trout, J. D., 21
of dif ferent sorts of, 95–96 Truth: correspondence theory of, 34–35,
Set theory, 86 94–96; Davidson and, 53–54; defla-
Shoemaker, Sydney, 157, 158, 179 tionary theory of, 35–36, 37–38, 40, 94;
Skepticism: Cartesian, 141–142, 186–187, evidence-transcendence, 130–133, 135;
221; Kantian, 141–142, 186, 187–191 idealization theory of, 122, 124;
Skinner, B. F., 39 justificationism and, 121–122; metaphys-
Snowball/snowdiscall example, 81–84 ical realism and, 26–27; Peirce on, 91–92;
Social contract theory, 59, 61–62 perspectival, 72; radically perspectival,
Solipsism, 120, 121, 124, 125 70; reference and, 28, 30–31, 33–34, 37,
Sosa, Ernest: conceptual relativity and, 69, 38, 70; Tarski and, 27–38
77–80, 82–83; eliminativism and, 86–88; “Truth: Deniers and Defenders”
“fi nished science” and, 74–76; internal (Dummett), 115–116
realism and, 67–68; living with the “Truth and Convention” (Putnam), 69
explosion and, 85–86; model-theoretic Truth and Objectivity (Wright), 134
argument and, 68–70; “nonlinguistic” Truth and the Past (Dummett), 115, 120,
restatement and, 81–85; reference and 121, 125
causation and, 70–74 “Truth as Sort of Epistemic” (Crispin),
Speech dispositions, 39–40 128–129
Strawson, P. F., 9 Truth-conditions, 84, 93, 95–96
Stroud, Barry, 9 Truth theory of meaning, 103
Sun Yat-sen, 59 T-schema (truth schema), 27, 30, 34–35,
Supervenience, 14 37–38
Surface-irritation model of perception, “12 Philosophers— and Their Influence on
159, 161–162, 165, 166 Me” (Putnam), 169
“Swampman” thought experiment, Twin Earth thought experiment, 208–209,
222–223, 226 217–218
Sympathy, 56–57 “Two Dogmas of Empiricism” (Quine),
“Synonymity, and the Analysis of Belief 114, 170
Sentences” (Putnam), 200 Tye, Michael, 157, 179, 211
Syntactic Structures (Chomsky), 199
Syntax, 199 Ultra-naive realism, 211
Unobserved objects, status of, 105–109, 117
Tarski, Alfred, 27–38, 203, 220; criterion of
adequacy, 28–30; Field and, 30–31 Values, 15–17
“Tarski’s Theory of Truth” (Field), 30–31 Veridical perceptual experience, 170, 190,
Tenth Problem, 204 211; sense data and, 155–158
“Testability and Meaning” (Carnap), 101, Verificationism, 38–39, 99–100, 120–122,
107–108, 111 125–126, 190
Thought experiments: Brain in a Vat, Verificationist semantics, 90–91, 166
216–217, 218–221, 222; “Swampman,” Vienna Circle, 99, 100, 102, 107
222–223, 226; Twin Earth, 208–209, Virtue ethics, 57–59, 61–62
217–218 “Visual Experiences” (Hinton), 174–175
Thoughts, content and, 191–194
The Threefold Cord (Putnam), 94, 124, Walsh, Vivian, 61
126–127, 147, 170, 171–172, 178, 180–181, Water, meaning and reference of, 207–209,
182, 186 217–218
Unauthenticated
238 Index
What We Owe to Each Other (Scanlon), 42 Wittgenstein, Ludwig, 92, 152, 154, 156,
White, Morton, 200 161, 173, 215, 224; Private Language
Whitehead, Alfred North, 75 Argument, 142–143
“Why There Isn’t a Ready-Made World” “Wittgenstein and Qualia” (Block),
(Putnam), 70, 71 169–170, 172, 179–180, 194–195
William James Lectures, 114, 117 Word and Object (Quine), 39–40, 168,
Williams, Bernard, 22, 214; absolute 203
conception of the world and, 44, 45, 47, Words: meaning of, 204–209, 217–221;
48–49, 50–52; intrinsic properties and, ostensive defi nition and, 142–143
76; metaphysical realism and, 73–74; World, absolute conception of, 44, 45, 47,
Putnam on “Philosophy as a Humanistic 48–49, 50–52
Discipline,” 44–55; scientism and, 70–71 Wright, Crispin, 128–135
“Wittgenstein: A Reappraisal” (Putnam),
154 Ziff, Paul, 200, 203
Unauthenticated

Hilary Putnam, Mario de Caro (Editor) - Naturalism, Realism, and Normativity-Harvard University Press (2016)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Hilary Putnam, Mario de Caro (Editor) - Naturalism, Realism, and Normativity-Harvard University Press (2016)

Uploaded by

Copyright:

Available Formats

Naturalism, Realism, and Normativity

Acknowledgments on pages 227–228 constitute an extension of this copyright page.

Library of Congress Cataloging-in-Publication Data

Introduction: Putnam’s Philosophy and Metaphilosophy

1. Naturalism, Realism, and Normativity . . . . . . . . . . . . . . . . . . . . . . 21

2. On Bernard Williams’s “Philosophy

3. What Evolutionary Theory Doesn’t Tell Us about Ethics . . . . . . . 56

II . RE ALISM AND ONTOLO GY

4. Sosa on Internal Realism and Conceptual Relativity . . . . . . . . . . . 67

5. Richard Boyd on Scientific Realism . . . . . . . . . . . . . . . . . . . . . . . . . 90

6. Hans Reichenbach: Realist and Verificationist. . . . . . . . . . . . . . . . . 99

7. Between Scylla and Charybdis: Does Dummett Have

8. When “Evidence Transcendence” Is Not Malign . . . . . . . . . . . . . 128

9. Sensation and Apperception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

10. Perception without Sense Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

11. “Naive Realism” and Qualia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169

12. The Development of Externalist Semantics . . . . . . . . . . . . . . . . . 199

13. Sixty-Five Years of Philosophy: A Participant’s Thoughts and

T h e t i t l e of t h is volu m e , which is taken from the

as his more recent interest in what he calls the “entanglement” of fact

1. Assuming that one accepts the arguments that Putnam has

All these questions, besides being of primary philosophical interest,

internal realist period) Putnam maintained that the theoretical terms of

1. Ontological realism. The views that belong to this family claim

Often the discussions of these three forms of realism are intercon-

As far as concerns the subject in which these tastes, smells, colours,

12. Jacopo Mazzoni, In Universam Platonis et Aristotelis Philosophiam Praeludia, sive de

the surreptitious substitution of the mathematically substruc-

According to Husserl, then, the only real world is the “life-world,”

ence,”17 and scientific concepts are mere idealizations with practical

nowadays. He, however, strongly disagrees with the negative side of

But the story of contemporary realism gets more complicated. In

25. See Chapter 1 here.

his mentor Marvin Farber, who had been a student of Husserl’s in

Speaking as a philosopher, I am quite prepared to say that the

Sellars and the other founding father of scientific naturalism, V. W.

1. Liberal naturalism should incorporate scientific realism. It

3. It has to be acknowledged that both the scientific and the

7. As to the epistemological status of philosophy, it should be noted

The world cannot be completely described in the language

The consequence of this is “conceptual pluralism,” the necessity

10. A satisfying liberal naturalist view should assume that among

causation. According to Putnam, causation is an intentional,

3. That there is a sometimes useful distinction between factual

The aspect of Putnam’s point of view that may be least familiar is

a monstrous chimera composed of mutually disproportionate

And it is worth noticing that it was for the same metaphysical-aesthetic

Nowadays philosophers tend to be specialized in very few and fre-

1. Hilary Putnam, “The Content and Appeal of ‘Naturalism,’ ” in Naturalism in Ques-

writing Julius Caesar was “subject to natural laws”? Is it enough to be-

Promoters of “naturalizing” projects are often driven, I think, by

5. Tyler Burge, Origins of Objectivity (Oxford: Clarendon Press, 2010).

explanations are, as I believe, pure hand-waving, the spirit, if not the

what I meant by it, based on the natural understanding of the phrase

But what is it to be a metaphysical realist? In an elegant paper titled

The most important consequence of metaphysical realism is that

And Maudlin’s reaction:

We now have something quite concrete to discuss. The metaphys-

current view. Truth is not the same thing as warranted assertability

Truth: Tarski’s Contribution and Its Interpretation

German word “wahr.”) Although a minority of philosophers, including

Tarski’s Criterion of Adequacy

ordinary language there is such a predicate as “true in Bob,” with no