Joseph Raz - Notes On Value and Objectivity

Objectivity in Law and Morals
Edited by
Brian Leiter
University of Texas, Austin
CAMBRIDGE
UNIVERSITY PRESS
CAMBRIDGE UNIVERSITY PRESS
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, Sao Paulo
Cambridge University Press

The Edinburgh Building, Cambridge CB2 8RU, UK
Published in the United States of America by Cambridge University Press, New York
www.cambridge.org
Information on this title: www.cambridge.org/9780521554305
© Cambridge University Press 2001
This publication is in copyright. Subject to statutory exception

and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 2001

This digitally printed version (with corrections) 2007
A catalogue record for this publication is available from the British Library
Library of Congress Cataloguing in Publication data

Objectivity in law and morals / edited by Brian Leiter
p. cm. — (Cambridge studies in philosophy and law)
Includes bibliographical references.
ISBN 0-521-55430-6 (hb)
1. Law and ethics. 2. Law - Philosophy. I. Leiter, Brian. II. Series.
K247.6.O25 2000
34O'.112-dc21 00-020002
ISBN 978-0-521-55430-5 hardback

ISBN 978-0-521-04149-2 paperback
Contents
Contributors page IX
Acknowledgments xi
Introduction 1
BRIAN LEITER
Legal Interpretation, Objectivity, and Morality 12
DAVID O. BRINK
Objectivity, Morality, and Adjudication 66
BRIAN LEITER
Objectivity Fit for Law 99
GERALD J. POSTEMA
Objective Values: Does Metaethics Rest on a Mistake? 144
SIGRUN SVAVARSDOTTIR
Notes on Value and Objectivity 194
JOSEPH RAZ
Embracing Objectivity in Ethics 234
PHILIP PETTIT
Pathetic Ethics 287
DAVID SOSA
Bibliography 331
Index 351
5
Notes on Value and Objectivity
JOSEPH RAZ
I. The Long Route
A. Introduction
It is natural that we should be interested in the nature of objectivity in general
and in the objectivity or otherwise of practical thought in particular. In one of
its senses objectivity is a precondition of knowledge. It also demarcates a type
of thought1 that is importantly different from others.
In this chapter our interest is in one way in which various types of thought
differ. They are subject to different disciplines. Suppose, for example, that I say
"I will be a good teacher" and you tell me: "But you have tried and failed for
years," "I may reply by saying: "So what? That does not stop me from
daydreaming," a response that is inappropriate if the thought expresses a belief.
This is but one example of one aspect of what I called the different
"disciplines" to which a thought or the holding of a thought can be subjected, a
difference that marks the distinction between classes of thoughts and of ways
of holding them. These disciplines determine whether my thought and yours,
which have the same content, belong to the same type.
Of some thoughts, for example, it is possible to say, "They were mistaken,"
whereas of others this is inappropriate. A closely connected mark is that having
or holding some thoughts can constitute knowledge. I mean that the fact I think
that a general election was held in Italy yesterday is part of what makes it true
that I know that a general election took place in Italy yesterday.
We mark this distinction between classes of thoughts that can constitute
knowledge and can be mistaken and those that cannot by calling the first
objective and the others subjective. When we ask whether practical thought is
objective or subjective we are asking whether it is subject to one type of
discipline or another. So the inquiry is natural and important. In the first part of
I am grateful to Peter Hacker and especially to Brian Leiter for comments on an earlier draft.
the chapter ("The Long Route") I will say something in general about the way
we can characterize and explain objectivity in abstract terms. Parts II and III
attempt to allay certain doubts about the objectivity of practical thought. Part II
deals with the fact that practical concepts are parochial and not universally
accessible to all. Part III deals with the social dependence of practical concepts.
B. Some Senses of the Objective/Subjective Contrast

The objective/subjective distinction originates in philosophy. With time its
uses multiplied. People, beliefs, sometimes even facts are said to be objective
or subjective, not only thoughts. And the suspicion is strong that more than one
distinction is drawn by the use of these words. Here, to illustrate the point, are a
couple of senses of the "objective/subjective" contrast: In science it is some-
times said that an experiment is subjective if its results cannot be mechanically
replicated. "Objective" here has something to do with intersubjectivity, when
the intersubjective standard assumes the very strict form of a more or less
mechanical process of experimentation whose results can be established by
relying on little more than simple perception (assuming, of course, a rich
background knowledge). A much more common use of the contrast, and a
much less interesting one, is that by which any mental states are subjective,
meaning belonging to a subject, and statements about mental states are subjec-
tive simply in virtue of being about mental states.
More relevant for our purpose is a third common sense in which the distinc-
tion is used. In what I will call the epistemic sense people are objective about
certain matters if they are, in forming or holding opinions, judgments, and the
like, about these matters, properly sensitive to factors that are epistemically
relevant to the truth or correctness of their opinions or judgments, that is, if
they respond to these factors as they should. Their views or beliefs may be
wrong or mistaken, but there are no emotionally induced distortions in the way
they were reached, or the conditions under which they are held. That means, for
example, that people are objective in this sense if they form their opinions and
judgments on the basis of the relevant evidence available to them, mindful of
whether or not they are in circumstances that might affect the reliability of their
perceptions or thought processes, and when their selection of information as
relevant and their evaluation of it are sensible and not affected by such emo-
tional or other psychological distortions. Opinions and beliefs are objective if
they are reached or held in an objective manner. Sometimes people are said to
be objective if they are capable of being objective in the sense just explained.
A related sense is that of objectivity as impartiality. People are impartial if
and only if, in matters affecting others as well as possibly themselves, they act
on relevant reasons and shun irrelevant ones, in particular if they shun irrele-
vant considerations that favor themselves or people or causes dear to their
hearts, and if their evaluation of the situation is not distorted by the fact that
196 JOSEPH RAZ
such people or causes are dear to them.2 Epistemic objectivity as a capacity to

avoid bias, or other emotional distortions, requires impartiality as one of its
constituent conditions, and this condition by itself is sometimes taken to make
a person an objective judge of a certain matter. But the impartiality sense of
objectivity reaches beyond these cases. It is the most important sense of objec-
tivity in practical matters, objectivity in action (e.g., when the objectivity of
judges or of officials is in question, usually impartiality is meant). Not infre-
quently official action is based in part not on the beliefs of the officials con-
cerned, but on the basis of propositions that are accepted by them as a reason-
able basis for action in the circumstances of the case. Objectivity as
impartiality can characterize deliberation and decisionmaking based on
accepted — rather than believed — premises as well. This requirement of objec-
tivity is a requirement of impartiality in all the stages of the decision-making
process.
Our interest is in a different distinction drawn by the objective/subjective
contrast. I identified it at the outset as the sense of objectivity in which it is a
condition of the possibility or perhaps of the conceivability of knowledge, and
a condition for the applicability of the notions of mistaken or correct (true)
thoughts. Inter subjectivity, albeit in a weaker sense than that required by the
scientific usage that I mentioned, and the possibility of holding views that are
not influenced by bias, or by similar cognitive distortions, are among the
general conditions for the possibility of knowledge. But equally clearly we
have not yet identified the core sense that is relevant for our purpose. For lack
of a better word, I will sometimes refer to domain-objectivity when wishing to
distinguish this sense of objectivity from others. Primarily it is domains of
thoughts, propositions, or statements that are objective in that sense.3 By
natural extension, single propositions, statements, and thoughts are objective if
they belong to an objective domain, and so is whatever they are about.
It is not easy to describe what makes domains of thought objective. The
difficulty is that in trying to do so one enters disputes about the nature of what
there is, the nature of reality, and about the structure of thoughts or of proposi-
tions about how things are. A proposition is objective if it is about an object of
a nonempty kind, that is, such that there are objects belonging to it, some
writers suggest.4 But what counts as an object and how do we determine what
kind it belongs to? Worse still, how do we determine what propositions are
about? Consider, for example, propositions like "The African lion lives to be
23 years old and fathers 3.7 cubs"; "There are two types of energy, static and
kinetic." Both are objective propositions. That is not in dispute. But is it
because they are about objects such as the African lion and energy? Or are they
about lions, all the lions bred in Africa, and all the objects in the world? To
resolve these questions we need to come to a view regarding the status of
universals, of dummy objects, and of the conditions for the existence of objects
in general. We may also have to enter into the morass of determining what
propositions are about. And there are further difficulties in store.
David Wiggins seems to sidestep these problems when he says that "A
subject matter is objective (or relates to an objective reality) if and only if there
are questions about it (and enough questions about it) that admit of answers
that are substantially true — simply and plainly true."5 The qualification "sub-
stantially true" puts us on our guard, however. It would seem that in order to
gain an understanding of objectivity we need to resolve issues concerning
various kinds of truths.6
Trying to sidestep these issues as well I suggested at the beginning of this
article that a domain is objective if it is capable of being the object of knowl-
edge, if propositions that belong to it can state what one knows, thoughts that
belong to it can manifest what one knows. If only true propositions can be
known, and only true thoughts can express one's knowledge, domains are
objective in this sense only if thoughts, propositions, and statements that be-
long to them can be true or false. But problems are to be found here too. If there
can be aspects of reality that it is logically impossible for us (but who are "us"?
The human race? Or all agents capable of knowing something? Or each one of
us?) to know anything about, then while the objectivity of a domain is a
precondition of knowledge, the reverse is not the case. We cannot say that a
domain is objective only if it can express knowledge, that is, express what is
knowable. There may be (objective domains of) propositions that state how
things are regarding matters of which it is impossible to have any knowledge.
This is one reason why I will not try to define domain-objectivity or to give
necessary and sufficient conditions for the objectivity of domains. Rather I will
attempt to elucidate the notion by considering a number of conditions that
partially characterize it.
While the epistemic and domain senses of objectivity are obviously interre-
lated, they differ in important respects. When judging someone or some opin-
ions to be epistemically objective (or lacking objectivity), one presupposes that
the matter which is so judged is domain-objective. In other words, domain-
objectivity marks the fact that we are dealing with a domain about which one
can be objective in the epistemic sense, and statements that are domain-
objective are those that can also be epistemically objective.
One way in which the two clusters of senses differ is in the responsibility7 of
people who are or are not, or whose opinions are or are not, objective in the
various senses. People should be epistemically objective and their opinions
should not be epistemically subjective. They are responsible for failing to be as
they should, for such failures are due to biases and other distortions of their
cognitive functioning for which people are held responsible. Their responsibil-
ity for their failures of epistemic objectivity does not assume that knowledge is
a good that they should pursue. It only means that they should not form
198 JOSEPH RAZ
judgments that are tainted by bias. Even though in some of its variants noted
above epistemic responsibility is a matter of ability rather than performance, it
is an ability the presence or lack of which can only be established by perfor-
mance. Hence those who lack it are those who are lacking in performance.8
The only responsibility that can arise if an area of inquiry or judgments
about it are domain-subjective is responsibility for failure to realize that the
domain is subjective. There can be, of course, no responsibility for these
matters or judgments being subjective. This is just how things are. Nor for that
matter can people be responsible for treating matters subjectively in these
cases, for there is nothing else they can properly do.
It is crucial to notice that in no sense is objectivity identified with truth,
though in all of them it is related more or less indirectly to truth. Thoughts that
are objective can be the subject of knowledge, and they are expressed in
propositions that can be true or false (or correct or incorrect). People who are
epistemically objective judges of certain matters are not free from error in
those matters. They are merely free from errors arising out of biases and an
emotionally blinkered inability to respond to the evidence and evaluate it.
Statements and opinions that are epistemically objective are those that are
neither reached nor sustained through biases of this kind. But obviously they
can be false or mistaken for other reasons. Epistemic objectivity is closely
associated with rationality. It is a condition of being a rational judge of certain
matters that one is epistemically objective regarding them. Being objective is
not, however, sufficient to make one a rational judge. Various inabilities may
undermine one's rationality that do not arise out of bias but are a result of
cognitive incompetence, hastiness, and like defects that do not undermine
one's objectivity.
C. Some Conditions of Domain Objectivity

Given our interest in the objectivity of practical thought, we can leave the
epistemic sense on one side and concentrate on the question of the objectivity
of the domain of practical or evaluative thought. What exactly is domain-
objectivity? How is it to be explained?
Possibly there is no short answer to the question. A domain of thought is
objective if and only if thoughts belonging to it satisfy a whole range of
criteria. They define the discipline to which objective thoughts are subject. We
have already encountered a few of them:
1. THE POSSIBILITY OF KNOWLEDGE CONDITION. Only if a domain is objective

can it express knowledge or be said to be about something that one can have
knowledge of.
2. THE POSSIBILITY OF ERROR CONDITION. A domain is objective only if

thoughts that belong to it obey the distinction between seeming and being, that
is, only if the fact that I think that things are so and so does not constitute their
being so and so (some self-referential thoughts are the exception).9
3. THE POSSIBILITY OF EPISTEMIC OBJECTIVITY. Only if a domain is objective

can one be an objective or a subjective judge of matters in this domain.
Similarly, only then can one's opinions about it, or the way one forms them, be
objective, et cetera.
This last condition is in some sense a subsidiary of the first two. But another
condition central to our understanding of the objective is:
4. THE RELEVANCE CONDITION. Thoughts are objective only if there can be

facts or other considerations that are reasons for believing that they are or are
likely to be true or correct.
The second condition allows us
i. to talk of thoughts as being correct, true, or as being errors and mistakes.10

ii. to admit that two views are contradictory and therefore that there are
logical relations among thoughts
iii. to talk of changing one's mind (with the implication that one was in error
then, not only that one now wants something different), et cetera.
The fourth condition allows us to apply the distinction between what is

relevant to the truth of a thought and what is not, and the whole terminology of
support and confirmation. The reasons for holding a thought to be true are of a
large variety of kinds. They may be evidential (there are heavy clouds, there-
fore it will rain), or legal (it has been approved by the Queen in Parliament,
therefore it is a valid law), or moral (it shows disrespect, therefore it is wrong),
or semantic (he is your uncle's son, therefore he is your cousin), and many
others. Reasons for believing that a thought is true need not be available to the
person who has it, or at least need not be available to him as reasons for the
thought. This caveat covers, among other cases, the case where belief is direct
and not based on reasons. A person looking at a red bus and thinking "This is a
red bus" does not have reasons for his belief. But there are reasons to think his
belief correct, for example, that he was in an advantageous position to judge
the character and color of the object he saw.
The objectivity of a domain does not presuppose that anyone can have either
a priori or self-evident or incorrigible understanding of what relevant reasons
are like or of the rules of reasoning in this domain. It only presupposes that the
thoughts belonging to an objective domain (a) allow for the application of
judgments based on reasons. Schematically, it allows for beliefs of the
kind "This thought is unjustified for the reason that . . ." where the reason
adduced — even if it is a bad reason, and even if it is of a type that cannot but be
a bad reason — is one that can intelligibly be thought of as a reason in this
200 JOSEPH RAZ
context. And (b) they allow that there can be reasons of an appropriate kind,
that is, reasons whose existence makes such beliefs true.
Two further important conditions derive from Bernard Williams's discus-
sion of the absolute conception of reality (about which more will be said
below). In advancing the absolute conception Williams was concerned with the
conditions for the possibility of knowledge. His discussion is therefore relevant
to objectivity understood as a mark of domains of thoughts that could con-
stitute knowledge. Williams introduces the discussion by saying: "If knowl-
edge is what it claims to be, then it is knowledge of a reality which exists
independently of that knowledge. . . . Knowledge is of what is there any-
way. "ll This thought yields the condition of objectivity which can be formu-
lated as follows:
5. THE INDEPENDENCE CONDITION. Thoughts are objective only if they are

about a reality12 that exists independently of them, that is, whose truth is
independent of the fact that those thoughts are entertained.13
If there are exceptions to this condition they are marginal and degenerate
cases. Trying to refine the condition to accommodate them need not detain us.
Williams follows the above quotation with a complex statement introducing
several ideas, one of which can be called the single reality condition.
6. THE SINGLE REALITY CONDITION. The domain of objective thoughts is sub-

ject to the constraint that thoughts constitute knowledge only if they can all be
explained as being about a single reality.14
The condition is intended by Williams to allow for the acceptance of appar-
ently conflicting claims of knowledge, provided they can be reconciled as
claims about a single reality (e.g., by representing it in different but inter-
translatable schemes of representations), but to exclude claims whose apparent
conflict cannot be thus reconciled. In particular he is keen to rule out suggested
reconciliations that relativize claims of knowledge: "p is true from my perspec-
tive, or relative to me, or from point of view x, even while not-p is true from
your perspective," and so forth. It is a rejection of isolationism, of claims that
every domain of thought must be judged by its own standards of truth, knowl-
edge, and objectivity, which may be very different from those of any other
domain.15 The single reality condition, while allowing for different perspec-
tives and different domains of thoughts, each with its own standards of evi-
dence, et cetera, insists that they must all be reconcilable within one conception
of reality. They must be shown to be not only consistent with each other but
also perspectives or domains that can be seen as being about different aspects
of one world — not an easy condition to spell out,16 but one that is worth
retaining, as it bears on a variety of subjectivist and relativist claims.
We could go on drawing out the implications of these conditions and adding
others to them. I call this approach the "long route" to an explanation of
objectivity, for it does not yield a comprehensive set of conditions, the satisfac-
tion of which is necessary and sufficient for a domain to be objective, and that
provide an adequate explanation of objectivity beyond the need for further
refinement and adjustment. Any set of conditions, however successful they
may be in answering the questions and concerns that led to their formulation,
will need to be refined as new problems emerge. It is in principle impossible to
circumvent the long route by providing a set of necessary and sufficient condi-
tions that will never call for revision and refinement and from which all the
other necessary features of objectivity can be deduced. It is not possible to
argue for this impossibility here. It depends, first, on the fact that propositions
are open to somewhat different interpretations, that they are all inherently
vague and therefore, even if true, may call for further refinement to meet the
needs of new puzzles and hitherto unexplored questions.17 It depends, second,
on the ever-present possibility that new questions will emerge forcing us to
notice and articulate new conditions of objectivity. Perhaps the following could
serve as an example of another condition that does not follow from the above
and that may illustrate how more and more may emerge with emerging reasons
to think of them:
7. THE POSSIBILITY OF IRRATIONALITY CONDITION. That is, that it is possible

to endorse propositions belonging to the domain out of wishful thinking, self-
deception, et cetera.18
Even if the "additional conditions" that may emerge with time apply to the
same domain already delineated by previously spelled-out conditions, they are
not redundant. An explanation of objectivity aims to do more than mark the
boundaries of objective domains. It aims to explain puzzles about objectivity.
The point I am pressing home here is that as new puzzles arise new conditions
may be noticed which help to explain them away.
D. Defending the Approach

Is this "long route" to explaining objectivity adequate? Some may say that
while it brings out necessary conditions of objectivity, it is never going to get to
the real hard questions: are there really things or properties like "rights,"
"tastefulness" or "beauty"? Is it in principle possible for them to exist?
The problem with the approach I outlined, these people might say, is that it
considers only internal criteria of objectivity. Basically it regards thoughts as
objective if we treat them as objective or if the concepts and sentences em-
ployed to express them meet certain conditions, which show that they are
treated as expressing thoughts that can be true or false, and so forth. But we
cannot, so the objection goes, make a thought objective just by treating it as
objective. Whether or not it is objective is not up to us. It depends on how
things are in the world.19 In classifying thoughts as objective we are saying that
202 JOSEPH RAZ
they constitute beliefs, that to hold such thoughts is to have certain beliefs, and
that means that these thoughts can be true or false and that holding them can in
principle amount to knowledge. This shows, the argument proceeds, that
whether they are objective cannot just be a matter of how we treat them. It is a
question of what the world really is like, a question about how things are in the
world.
Here is my problem: Domains of thought are objective, I suggested at the
outset, only if they are subject to disciplines that make them capable of con-
stituting knowledge, and so on. The charge against the way I explained this
route to an explanation of objectivity was that it led to an articulation of what
we take to be the conditions of knowledge, rather than what really are the
conditions of knowledge. It may seem that the only way to justify my position
is to deny that the two can differ. But that would imply that we cannot be wrong
about the conditions of knowledge, which seems too strong a claim. In fact it is
not needed. All I need to argue for is that, necessarily, our knowledge or
understanding of the conditions of objectivity depends on our knowledge or
understanding of the conditions of knowledge. If our knowledge or under-
standing of one of these concepts is imperfect so is our knowledge or under-
standing of the other.
Whether our thoughts are true or not is immaterial to their objective charac-
ter, but they must, among other things, be capable of being true or false, of
bearing a truth-value. Therefore, they cannot amount to knowledge unless they
are about how things are anyway, about the world as it is independently of
these thoughts. That means (given that the character of domains of thought
depends on the disciplines to which we subject them) that they cannot be about
reality unless we treat them as about how things are independently of thought
about them.
As so far stated, these considerations do not show more than that satisfying
the conditions spelled out above, and others derived in like manner, is neces-
sary for a domain of thought to be objective. But their implications reach
further:
First, it is not the case that this approach allows that what is objective is "up
to us." That would mean that we can choose whether to regard matters of taste
or morality as objective. But the account I suggested has no such consequence.
Second, notice that these conditions do more than point to surface syntacti-
cal features, for example, that the thoughts are expressed in indicative sen-
tences, allow embedding, et cetera. They assume the practices that Wright
regards as underlying what he calls "the minimalist" account of truth.20
Wright's account has convinced some writers that nothing of substance de-
pends on allowing that propositions and utterances can be true or false, given
the minimalist account of truth.21 The condition of relevance, however, takes
my explanation of objectivity beyond the conditions that, according to Wright,
have to be satisfied for a truth predicate to apply to a domain.
Third, the condition of relevance makes the objectivity of a domain depen-

dent on the existence of criteria that support or undermine beliefs within that
domain. These criteria could in appropriate circumstances warrant the conclu-
sion that all beliefs within the domain are false or lacking in truth-value
(because a presupposition of their having truth-value fails). Thus this approach
allows in principle for an error theory22 that while admitting that a domain is
the subject of objective thought establishes that there can be nothing in it, that
there can be no rights, no tasteful or beautiful objects, and so forth.
Fourth, the approach allows for the possibility that a domain that appears
objective is not in fact objective, even though it is taken by everyone to be
objective. This can happen if discourse involving it collapses into contradiction
or incoherence, or if it fails to cohere with other established beliefs.
Fifth, the approach allows that our concepts of knowledge, truth, and objec-
tivity may themselves be defective and in need of revision (more on this in
Section HE below).
Sixth, externalist accounts are the better accounts of knowledge. According
to them, whether or not one's belief is justified depends, in part, on how things
turn out in the world, and not merely on satisfying criteria that the agent is
aware of.
The third, fourth, and fifth points establish that the approach I delineated is
not essentially a conservative one, that it is open to challenges that can force
revisions in, or even the abandonment of, established aspects of common
discourse and thought. The last point is but one additional way in which the
criteria of objectivity, derived from the essential properties of knowledge, are
not "too internal."
This six-point reply to the objection may only prompt some of those who
suspect the long route of being too internal to raise another charge. I seem to
have an answer, they would say, because I conflate and confuse a number of
different issues: metaphysical, epistemic, and semantic. The possibility of error
condition concerns semantics, the relevance and the possibility of irrationality
conditions are explicitly epistemic, and the independence and single reality
conditions are metaphysical.
At one level I have no defense against this charge. The list of criteria of
objectivity does include criteria that can readily be classified as semantic,
metaphysical and epistemic, as well as some that may defy such a classifica-
tion. Nor do I wish to raise any doubts about the legitimacy of separating issues
into semantic, metaphysical and epistemic. In many cases such a separation is
useful, and in some it is essential. But not all concepts can be readily separated
in this way, and to my mind the attempts by some writers23 to define separate
semantic and metaphysical senses of objectivity miss the point that the two are
interdependent, and at the most fundamental level these stipulated concepts are
not useful.
Take the relevance condition. It is epistemic in nature, but not exclusively
204 JOSEPH RAZ
so. Someone who does not know anything at all about what may count as
evidence in favor of the proposition that Gubaidulina is a great composer, or
that this chair is not comfortable, or that Everest is the highest mountain on
earth, does not really know what it is to be a composer or a mountain, and so
forth. If upon hearing the claim that Gubaidulina is a great composer someone
replies, "Oh yes? Does she eat fish?" then, barring a peculiar sense of humor,
his remarks reveal that he has not understood what he was told. Such misunder-
standings are often used in popular humor to expose the ignorance and pretense
of social climbers. Examples like this would show that part of understanding
any concept is having some grasp of what counts as relevant evidence showing
that it applies or does not apply.
At the same time, the fact that one thing or another is evidence for the
existence of something, or for its possession of some property, tells us some-
thing about the nature of these existences: material things are those whose
existence and properties can be established one way, mathematical objects and
properties are those that are established in very different ways, et cetera. In
sum, the relevance conditions straddle the divide between the semantic, epis-
temic, and metaphysical. It requires more than knowledge of connections
among concepts. It requires knowledge of the way things are in the world is
relevant to the application of concepts.
The interdependence of the different aspects (semantic, metaphysical, and
epistemic) reveals itself in the way arguments regarding problem cases pro-
ceed. There is a striking correlation between those philosophers who think that
there are "Non-Naturalistic" moral properties and those who believes that
moral predicates refer to such properties. Mackie is the only philosopher who
believes that moral predicates refer to nonnatural properties, but that there are
no such properties, nor can there be. Apart from him, those who believe that
moral statements are "semantically objective" also believe that there are moral
properties, and — some argue — since there are no nonnatural properties, moral
properties must be natural properties. Those who believe that there are no
moral properties opt for a noncognitive interpretation of moral utterances and
so on. The same is true regarding secondary qualities. Even Mackie does not
offer an error theory of secondary qualities or of mathematical properties. In all
these cases one's view of the nature of the properties that exist and one's view
of the semantics of the relevant range of predicates go hand in hand. They are
interdependent and, needless to say, so are the relevant evidential criteria. That
is why I claimed that domain-objectivity combines semantic, epistemic, and
metaphysical criteria.
The interdependence of the semantic and metaphysical is recognized and
explained by Crispin Wright. He argues that, if one believes, as error theorists
like Mackie and Field do, that even though propositions asserting the existence
of ethical (or mathematical or other) properties are false, we need not radically
change our use of them for they serve some other valid purpose, then one has to
explain why not accepting that truth in that domain is constituted by serving
that purpose. In his words: "why insist on construing truth for moral discourse
in terms which motivate a charge of global error, rather than explicate it in
terms of the satisfaction of the putative subsidiary norm [what I called "the
valuable purpose"], whatever it is? The question may have a good answer. . . .
But I do not know of promising argument in that direction."24 That is the sort of
consideration that led me to claim that domain-objectivity combines semantic,
epistemic, and metaphysical criteria.
In spite of these points doubts are likely to linger. While not being conserva-
tive, the criteria I mentioned are all in some sense internal to the ways we treat
thoughts in the domain in question. Is there no possibility that a domain will
meet these criteria and yet fail to be objective simply because of how things
are? In principle the answer is affirmative. Since the long route deals with
problems as they are encountered, and denies the feasibility of producing a
definitive list of necessary and sufficient conditions for objectivity, the (epis-
temic) possibility that factors may emerge that defeat the objectivity of do-
mains of thought that meet the above conditions cannot be ruled out. All we
can do is examine specific doubts regarding the objectivity of practical
thought. The rest of this chapter consists in dispelling two specific sources of
such doubt that have attracted considerable attention.
II. The Role of Parochial Concepts
A. Is Parochial Knowledge Possible?

A couple of terminological stipulations will help us along: "Perspectival
knowledge" is knowledge that can only be expressed with the use of parochial
concepts. "Parochial concepts" are concepts that cannot be mastered by all, not
even by everyone capable of knowledge. "Nonparochial" concepts can be
mastered by anyone capable of knowing anything at all. Any concepts posses-
sion of which requires having particular perceptual capacities (such as color
concepts) and not merely the possession of some perceptual capacity or an-
other, are parochial concepts. For different reasons concepts whose mastery
presupposes interests or imaginative or emotive capacities that are not shared,
nor can be shared, by all creatures capable of possessing knowledge are paro-
chial concepts. We come to understand interests we do not share by relating
them to interests we do share. If an interest is remote from any of ours, and if its
role in the life of the people who have it is ramified and unlike the role of any
interests we have, or know about, then explanation has to be supplemented
with simulation or real habituation before we can understand it or the concepts
that presuppose its understanding. But our capacity for simulation or habitua-
tion is limited. Interests that will evolve only in the future, for example, cannot
be understood in that way, and often they are interests that it is in principle
206 JOSEPH RAZ
impossible for us to understand. By the same token, many of our interests were
beyond the reach of people in previous generations, and so were the concepts
mastery of which depends on understanding these interests.
The impossibility of acquiring concepts that presuppose interests remote
from those we have and that do not yet exist (and therefore cannot be mastered
or discovered by simulation) is sufficient to show that practically all concepts
that can be acquired only by people who understand some nonuniversal inter-
ests are parochial concepts. Regarding all of them, there was a time when they
were beyond the reach of people living at that time. There is, however, yet
another type of limitation on people's ability in principle to master such con-
cepts: habituation and simulation are demanding and relatively slow processes.
You often have to immerse yourself in an alien culture, understand its con-
cerns, religious and other beliefs, come to understand how people normally
react in many (to them normal) situations, share —at least in imagination-their
hopes, fears, and aspirations. All these factors mean that even if there is no
interest-dependent concept one cannot acquire, it is in principle possible for
any person to master only a relatively small number of such concepts and
hence, given the length and diversity of human history, impossible in principle
for anyone living today to master more than a fraction of the interest-dependent
concepts that pertain to different human societies now, or did in the past, even
if we leave aside the future.25
If interest-related concepts are parochial, so are evaluative or normative
concepts, for they are all interest-related (not, of course, in "serving" the
interests of the agent, but in the fact that mastery of the concept requires
understanding some interests or others). This is evident when we consider
relatively specific, so called thick, concepts, like the excellence of an opera or a
novel. But the same is true of our abstract normative concepts such as those
designated by the terms "duty," "obligation," "ought," "a right," "valuable,"
"good," "beautiful," "person," "happiness," "pleasure." The history of these
terms, and the attempt to find terms of comparable meaning in other languages,
show how their meaning mutated over time and how different languages differ
in their abstract normative vocabulary. It is reasonable to conclude that abstract
normative concepts too are parochial.
Doubts have been cast on the objectivity of parochial concepts, on the
possibility of knowledge that depends on their possession and cannot be refor-
mulated without their use. I will examine some grounds for such doubts.
B. Nagel on the Objective and the Subjective

Thomas Nagel regards parochial concepts as subjective but does not deny that
they enable us to acquire knowledge that cannot be obtained without them. In
most respects my views are consonant with his. The most striking apparent
difference is due to the special sense in which he uses the objective/subjective
distinction, which differs from any I have mentioned so far. "A view or form of
thought is more objective than another if it relies less on the specifics of the
individual's makeup and position in the world, or on the character of the
particular type of creature he is." 26 In other words, Nagel calls a view or
thought subjective if it essentially depends on employing parochial concepts
for its expression. The greater its dependence on parochial concepts the more
subjective it is.
For Nagel "objectivity is a method of understanding."27 It seems to be a
method of understanding reality with different aspects of reality being under-
stood depending on the degree of objectivity attained in the method of under-
standing them: "We may think of reality as a set of concentric spheres, progres-
sively revealed as we detach gradually from the contingencies of the self"
{Nowhere, 5). This sentence may suggest that the more objective our method of
understanding the more of reality we come to know and understand. But this is
not Nagel's meaning, or at least it is only part of his meaning. While some
aspects of reality reveal themselves only as they are objectively investigated,
others are lost except when understood through subjective methods. "The
attempt to give a complete account of the world in objective terms . . . inevita-
bly leads to false reductions or to outright denial that certain patently real
phenomena exist at all" {Nowhere, 7). In particular, "[t]he subjectivity of
consciousness is an irreducible feature of reality" (ibid.), which presumably
means that it can only be fully known and understood with the use of subjective
methods, including the use of parochial concepts. Nagel seems to think that
only the mental requires parochial concepts for its understanding.
Although there is a connection between objectivity and reality — only the supposition
that we and our appearances [this seems to mean here: how things appear to us, not how
we appear to others — JR] are part of a larger reality makes it reasonable to seek
understanding by stepping back from the appearances in this way — still not all reality is
better understood the more objectively it is viewed. Appearance and perspective are
essential parts of what there is, and in some respects they are best understood from a less
detached standpoint. Realism underlines the claims of objectivity and detachment, but it
supports them only up to a point. {Nowhere, 4)
But how can we know that no other aspect of reality requires parochial
concepts for its understanding? Perhaps the thought is that since the mental is
subjective it requires subjective concepts for its understanding. But that
thought is guilty of the fallacy of equivocation: Thought is subjective in the
psychological sense that means simply that it is mental. The "method of sub-
jectivity" as a method of knowledge does not mean knowing in a way which
involves mental states, dispositions, or capacities. All knowing involves the
mental in these ways, be it knowledge obtained by more or less objective
methods. "The method of subjectivity" means knowledge that can only be
obtained by the use of parochial concepts. To avoid equivocation the thought
208 JOSEPH RAZ
should be: parochial concepts, and thoughts involving them, can only be under-
stood with the use of parochial concepts. That amounts to claiming that paro-
chial concepts cannot be eliminated. If they are necessary for knowledge of
certain matters, then those matters cannot be known "objectively." That would
not justify the conclusion that only the mental can be understood by the use of
parochial concepts.
Perhaps I am wrong to identify Nagel's understanding of the objective/
subjective distinction with reliance on and the use of concepts that are more or
less parochial? Perhaps all he means is that the mental can be understood only
with the use of psychological concepts? He certainly believes the latter, but that
is not all he is claiming.
The more objective our understanding the more detached it is from our
specific situation, specific capacities, et cetera. To quote Nagel again: the more
objective the thought the less it relies "on the specifics of the individual's
makeup and position in the world, or on the character of the particular type of
creature he is." These specifics constitute what he calls the special perspective
of that creature. Objective knowledge, like any other knowledge, is obtained
within that perspective, but it does not rely on it. It can be shared by people
whose perspectives are different. Subjective knowledge cannot be shared in
this way. It is available only to those who share the same perspective.
My suggestion is that perspectives can be identified by the range of con-
cepts that those who inhabit them can possess. This does not exhaust the
differences between perspectives, but the others follow from the inability of
those who inhabit a perspective to possess certain concepts. The possession of
concepts, let us remind ourselves, involves certain mental abilities, certain
perceptual abilities, and certain experiences. For example, inability to master
mathematical concepts makes one incapable of possessing a whole range of
scientific concepts, the lack of sight makes the understanding of color concepts
incomplete, and those who never desired anything cannot understand what it is
to want something.
So the more objective knowledge is the less does its possession presuppose.
It presupposes fewer common mental abilities, fewer common perceptual abil-
ities, and fewer common experiences. If that is so we can understand why
knowledge of the nature of thought is more objective than knowledge of the
nature of Christianity. To understand the nature of thought one needs to have
the capacity for thought, but other things being equal, this capacity is available
to anyone capable of thinking. These may include Martians, automata, and
others, besides humans. Knowledge of Christianity requires concepts such as
salvation, redemption, and love that need not be available to all who are
capable of knowledge. Hence it is more subjective.
This line of thought can explain some aspects of Nagel's position. But it
does not seem to explain some of its basic tenets. First, it does not explain why
all knowledge of the mental is more subjective. Since only people with a
mental life are capable of knowledge none is excluded from the perspective
that involves at least some mental concepts. Second, it does not explain why
only the mental defies complete objectivity. It does not answer the question:
Why should we assume that there are no aspects of reality that can be known or
understood only by creatures with specific capacities and experiences, that are
incompatible with the capacities needed to understand other aspects of reality?
Nagel seems bereft of an argument which will limit his "subjective method" of
knowledge to the mental only.28
C. Williams and the Priority of Nonparochial Concepts

Williams in discussing secondary qualities introduces the requirement that all
knowledge must be capable of being expressed without the use of parochial
concepts. I will call this the priority condition, meaning the condition that a
thought is objective only if it can be expressed without the use of parochial
concepts. Here is his argument:
Can we really distinguish between some concepts or propositions which figure in the
conception of the world without observers, and others that do not? Are not all our
concepts ours . . . ? Of course; but there is no suggestion that we should try to describe a
world without ourselves using any concepts, or without using concepts which we,
human beings, can understand. The suggestion is that there are possible descriptions of
the world using concepts which are not peculiarly ours, and not peculiarly relative to our
experience. Such a description would be that which would be arrived at, as C. S. Peirce
put it, if scientific enquiry continued long enough; it is the content of that "final opinion"
which Peirce believed that enquiry would inevitably converge upon, a final opinion . . .
independent not indeed of thought in general, but of all that is arbitrary and individual in
thought.29
The scientific representation of the material world can be the point of convergence of
the Peircean enquirer precisely because it does not have among its concepts any which
reflect merely a local interest, taste or sensory peculiarity. . . . This extended conception
would then be that absolute conception of reality.30
And Williams contrasts this view with that of a critic who maintains that
"scientific theories are a cultural product which it would be senseless to sup-
pose could be freed from local relativities."31
The motivating concern leading to the priority condition seems to be the
need to establish the credentials of our knowledge claims and of our concepts.
We need to know that through our concepts we can understand the world as it is
independently of us, that our thoughts are no mere shadows cast by our own
concepts. Therefore, any view or thought that claims to be knowledge must
meet a test that is concept-independent, or — given the impossibility of thought
without concepts — that is at least independent of any concepts that are not
accessible to everyone. That test is the Peirceian convergence of all those who
210 JOSEPH RAZ
are engaged in pure inquiry. In the nature of things if the convergence is to

encompass all those capable of knowledge, then it must exclude thoughts that
essentially depend on parochial concepts for their expression. Such parochial
thoughts will not be available to all inquirers and therefore will not be able to
be subjected to the test of convergence.
The priority condition has been criticized as incoherent and inconsistent
with its aim of establishing some limited credentials of knowledge that relies
on parochial concepts.32 In later writings Williams abandons it and concedes
that the understanding, as well as the explanation, of knowledge relying on
parochial concepts cannot be accomplished without such concepts.33 This
entails also abandoning the ideal convergence condition in Peirce's version that
Williams endorsed earlier. There is knowledge regarding which no con-
vergence of all competent inquirers is possible. "Alien investigators," to use
his phrase, may be unable to understand our perspective, even those aspects of
it that yield knowledge, if that knowledge is dependent on parochial
concepts.34
Convergence is in any case a suspect condition.35 A degree of convergence
is entailed by the fact that agreement in meaning presupposes a degree of
agreement in judgment. This does indeed vindicate testing agreement in mean-
ing by looking for agreement in judgment. But the convergence that test entails
is limited to those who share the same meanings. This cannot be a requirement
of a Peirceian convergence of all competent inquirers, since, by definition, they
do not share understanding of parochial concepts. Within local communities
who share parochial concepts, such as "inflation," "chivalry," and their like,
there will indeed be a degree of convergence in judgment. But not enough to
eliminate disagreements and disputes. In particular, beliefs whose justification
typically involves complex reasoning regarding a diverse range of consider-
ations will always be liable to disagreement even among those who share the
same meaning and a similar environment (and therefore a similar evidential
starting point).
Most theorists who recommend convergence as a criterion of objectivity
emphasize that it is the convergence of competent inquirers that is achieved
when they follow the relevant reasons. The required convergence can never be
the convergence of all. But who is competent is not a neutral question, indepen-
dent of one's other epistemic claims. Rather, it is defined by competence to
understand and apply the criteria of objectivity and, in particular, the condition
of relevance. Therefore, if parochial concepts are essential to some knowledge,
then convergence may possibly be attainable, but it will be confined to those
who can master these concepts, not the Peirceian convergence of all competent
inquirers. Ideal convergence cannot be a requirement that rules out parochial
concepts or removes them to a lesser status. Rather settling the question of the
status of parochial concepts is necessary before one can establish what sort of
convergence is necessary.
There is another, perhaps more radical, limitation on any convergence re-

quirement. Would the criterion of relevance meet a convergence requirement
limited to those who master the relevant concepts? Not quite, it would have met
the requirement but for the fact that it allows for indeterminacy of reasons that
may lead rational inquirers, even when they share the same premises, to
diverge in their conclusions on those occasions where it would be rational to
believe a certain proposition and also rational to doubt it. The manifestations of
the underdetermination of reasons are widespread and familiar. For example,
two people hear the same testimony and one believes it while the other does
not. In many (though not in all) everyday situations like this both have to admit
that the other is not irrational, while insisting, of course, on the rationality of
their own stance. Common as underdetermination by reason is, it has not
received adequate attention in epistemology. This is not the occasion to discuss
it in detail,36 though I will return to the point later.
D. The Siren Call of Epistemic Absolutes

Why does Williams think that parochial concepts are more suspect than non-
parochial ones? They are concepts "which reflect merely a local interest, taste
or sensory peculiarity." What is it for a concept to reflect merely a local
interest? It has become an oft-repeated philosophical example that the Innuits
have more terms designating types of snow than are known to any other
language. Are the concepts of types of snow they have an example of concepts
of a "mere local interest"? This is how Putnam understands Williams, for he
thinks that "grass" is similarly a local concept.37 They may well be concepts
that not every inquirer can master. Their mastery may well presuppose famil-
iarity with some minutiae of Innuit life, which give their ways of distinguishing
types of snow their point, and which are beyond the comprehension of Mar-
tians. As mentioned earlier, perceptual concepts are another class of parochial
concepts. Since not all people possess the same sense organs not all of them
can master the concepts possession of which requires perceiving the world
through particular senses.
It is not my purpose to argue that scientific theories are likely to find
concepts like "grass" or "sweet" particularly useful.38 Suppose that they do
not. Is that relevant to the preconditions of knowledge or to the nature of reality
or the world? Accountants have no more use for the notion of quarks than
physicists for the notion of inflation. Does it follow that inflation is a mere
illusion? Clearly not. The interest of Innuits in snow may lead to the emergence
of a range of parochial concepts, which are no use to the rest of the world. But
that does not show that they do not pick out real features of the real world. In
fact it is hardly conceivable that they could express local or, indeed, any tastes
or interests if they did not. The same considerations apply to perceptual con-
cepts. That certain creatures lack a certain perceptual capacity may prevent
212 JOSEPH RAZ
them from becoming aware of certain aspects of the world that can be known
only by those who have that capacity. It does not show that those who have
those senses are not able through them to perceive how things really are, thus
acquiring (perceptual) knowledge the others lack.39
Clearly some concepts have no use in formulating any true thoughts about
how things are (other than about what use people made of those concepts). The
credentials of concepts are not beyond question. But I doubt that parochial
concepts need more defense than nonparochial ones. "Ether," "alkahest," and
other scientific or pseudoscientific concepts, nonparochial if any are, fail the
test. They are no help in acquiring knowledge. Some of these concepts are
empirically empty. Others are conceptually incoherent or inconsistent with
basic scientific laws.
Nonparochial concepts, because they are in principle accessible to all, have
one obvious advantage. When they can be used to test parochial concepts and
parochial epistemic principles, then they can be used to test whether those
parochial concepts and principles are free of local biases. Such taints will be
exposed once the concepts and principles that incorporate them are judged by
principles that are accessible to people not blinkered by those interests, and
whose vision is not distorted by the peculiarities of any specific perceptual
capacity or alleged capacity.
It does not follow that all parochial concepts or principles can be thoroughly
tested in this way. Moreover, pointing out this advantage is very different from
giving nonparochial concepts and principles special priority, a special status
among the conditions for the possibility of knowledge.40
It seems to me that the thought that they could play such a criterial role is a
mutation of a Cartesian ideal. Not that Williams's absolute conception involves
a commitment to self-evident knowledge, but it is committed to the thesis that
the objectivity of thought is underpinned by an epistemic principle stating that
under ideal conditions knowledge is undisputed and clear for all to see. Two
ideas combine here, one explicit and one implicit: The first is that under ideal
conditions there will not be any deep disagreement. (If a disagreement occurs it
will be quickly resolved —by looking at the evidence as it were). The second is
that in the ideal conditions epistemic tests will not lead us astray. In the real
world, it is acknowledged, we are at the mercy of epistemic luck. We may often
be epistemically justified, indeed required, to accept beliefs that are in fact
false. But under ideal conditions this will not happen.
These ideal conditions may never be realized, but only an epistemology that
subjects itself to this test, that accepts that what is true will be recognized as
such under ideal conditions, can underpin a claim to objectivity. Only domains
of thought that can stand this test (i.e., that out of the thoughts which belong to
them, under ideal conditions competent inquirers will necessarily converge and
accept the same thoughts and reject the same thoughts) are domains of objec-
tive thought. While we acknowledge, the thought is, that in real life we are at
the mercy of epistemic luck since we may be rationally justified in accepting

false propositions as true, no domain of thought is objective if it is necessarily
always subject to epistemic luck, if there are no circumstances in which the gap
between justified belief and true belief is closed.
There is no cogent reason to accept this yearning for freedom from epis-
temic luck. In all areas of knowledge there are limits — imposed by conditions
of intelligibility — to the possibility of error. But beyond that we are at the
mercy of epistemic luck. This is an inevitable result of the fact that (again with
the same exception already mentioned) epistemic justification is path-
dependent. This means that what is rational for one person to accept (given his
situation and history) may well be rational for another person to reject (given
his situation and history). It follows that whether or not they have knowledge
depends, for both of them, on the luck of their situation. There is no reason to
think that this path-dependence can be overcome by everyone occupying the
same starting point. There are good reasons to think that it is not possible for
everyone to occupy the same starting point (this will involve, e.g., time travel,
changes in perceptual organs, as well as complete change of culture). There are
also good reasons to think that there is no starting point such that those who can
occupy it can know everything that can be known. If some knowledge is
parochial then that possibility is ruled out. If people must diverge in their
epistemic baggage then path-dependence is a necessary feature of human
existence. One that cannot be overcome under any conditions, however ideal.
There is therefore no reason to make submission to a luck-free ideal test a
condition of objectivity.
E. Epistemic Anxieties and the Long Route

I suggested that domains of thought are objective if they meet conditions like
the possibility of error, the relevance condition, the possibility of irrationality,
the independence condition, and the single reality condition. The list can be
extended indefinitely. The method of the "long route" consists in tracing
various truisms associated with the logic of truth and knowledge. We seek them
out as the need arises, that is, as we encounter questions and doubts they help to
answer.
I raised a doubt about this method, namely, that it is "too exclusively
internal." We can agree that whatever meets its conditions is thought by us to be
objective, but is it necessarily so in reality? The thought that this method is "too
internal" is the thought that there are, or at least that there could be, precondi-
tions for anything being capable of being true or false that are due to how
things are and are not reflected in the logic of our concepts.
The quest for ideal convergence, or for the testing of parochial concepts by
nonparochial ones, which are available to all, are manifestations of the same
anxiety. They are attempts to break out of the inner and to connect with how
214 JOSEPH RAZ
things really are in the world. They do so by claiming that knowledge must be
anchored in the universal, in those concepts all can have and those claims of
knowledge round which all competent inquirers will converge in ideal condi-
tions. By replying that the worry is illusory, that by exploring the preconditions
of our concepts of knowledge and truth we are willy-nilly exploring features of
reality I was navigating a middle course. On the one hand is the Scylla of
denying that the nature of reality, knowledge, truth, and objectivity are just
what is revealed through use of the concepts we happen to have of them, and
requiring some absolute test satisfaction of which by a concept or a principle
guarantees that they reach to reality "as it really is." 41 Against this is the
Charybdis that denies that we can ever have any concepts of reality, knowl-
edge, or truth different from those we happen to have, or that the nature of
reality or knowledge can be different from what these concepts (our concepts
of them) reveal. Our concepts are what we measure by. A thought that is not a
thought (i.e., does not conform to our concept of a thought) is not a thought.
Knowledge that is not knowledge (i.e., not an instance of our concept of
knowledge) is not knowledge.
The middle road allows that our concepts can be subjected to rational
evaluation, which may lead to revision. But it denies that the evaluation is in
the light of any absolute test, like the tests of convergence. We judge our
concepts in light of our concepts and beliefs. While it is necessary that in
general they will be vindicated, some of them, some aspects of even the most
basic of them, may turn out to be incoherent or unsustainable, given whatever
else we know. If we are lucky we improve our understanding of reality and of
knowledge. But we do not do so by satisfying one or several master tests.
Similarly, some of our epistemic procedures may be misguided. What is taken
to be good or adequate evidence for a class of conclusions may turn out to be
inadequate. Considerations taken to be good reasons for certain positions may
be irrelevant, or confused, or just insufficient to justify them.
The history of the practice of science provides examples of how epistemic
standards change, often in the light of rational reflection and criticism, but
without any master test that is held constant and governs the changes. Various
people have suspected that concepts such as the divine, the supernatural, time
travel, miracle, action at a distance, beauty are not all that they seem to be.
Arguments have been put forward to show that their use has to be abandoned
because they are incoherent, or that they have to be reformed in some funda-
mental ways.
If that is so then even the necessary features of our concepts of objectivity
and knowledge may stand in need of revision and can be revised through
reasoning that follows the long route. To deny that is to assume that while the
use of these concepts can reveal difficulties with concepts like "beauty," that
pressure is one-sided: Our concept of "beauty" comes under pressure for being
in tension with the presuppositions of knowledge, but our concept of knowl-
edge is not under pressure for being in conflict with necessary features of
"beauty." But there is no reason to think that. So long as we remain holistic in
our approach to the clarification of the structures of thought the possibility of a
need to revise our notions of knowledge and truth remains.42 I am not saying
that they need revising, only that it is possible that they do.
To acknowledge this possibility is to acknowledge the possibility that there
are preconditions for anything being true or false, or capable of being known,
which are not reflected in our concepts, that is, our concepts as they are.
Therefore, the long route is not "purely internal" and self-vindicating. I con-
clude that it may be that the long route is all we need.
As doubts about the objectivity of this or that domain arise we inquire a n d -
if successful — we establish conditions that settle the issue one way or another.
In other words we need a focus for an inquiry. The focus establishes which
results might be relevant and helpful. There is no way of producing an exhaus-
tive list of such conditions, for we can never anticipate all the doubts that can
arise concerning the objectivity of this or that domain.
Put it another way: there is no interesting comprehensive theory of objec-
tivity. But there can be fruitful inquiries into the objectivity of one area or
another when specific doubts arise regarding their status.
III. The Illusion of the Authority of the Social
A. Thick Concepts and the Social Connection

In elaborating on his absolute conception Williams adds another condition to
those mentioned so far. It is set out in the words I italicize in the following: "If
knowledge is what it claims to be, then it is knowledge of a reality which exists
independently of that knowledge, and indeed (except for the special case
where the reality known happens itself to be some psychological item) indepen-
dently of any thought or experience. Knowledge is of what is there anyway. "43
So far as I know Williams does not develop this extension of his idea, but in
one form or another it is a common condition in many discussions of realism. It
is often set out in its strong form: knowledge must be of what is there anyway,
independently not only of the thoughts and experiences of the agent, but of any
thought or experience of anyone. This condition would rule out the possibility
of knowledge of many properties of artifacts and of many socially constituted
persons (organizations, cities, etc.), facts and events.
Think of an existing chess club,44 with its rules and customs. They are not
merely created through acts involving thoughts and experiences, they are also
maintained in existence through the continued intentions of the club's mem-
bers to carry on with their participation in the club. To be sure, some people
would want to say that such clubs, or their rules and customs, do not form a part
of what is really real. But if they are right there must be other reasons for that
216 JOSEPH RAZ
view. It does not follow from the idea that knowledge is of what is there
anyway.
Williams's condition fares no better if it is narrowed down to saying that
knowledge must be of a reality that exists independently of any thoughts or
beliefs about its existence. The example of the chess club refutes this condition
as well, since the existence of the club depends on some people believing (or
having believed) that it exists.
Objects of knowledge can sometimes even depend on the agent's own prior
thoughts and experiences. The motivation for the condition of independence,
namely, that knowledge is of what is there anyway, does not justify the blanket
exclusion of all cases of this kind. For example, yesterday the chairperson of
the chess club may have passed a regulation binding on the club. Its existence
and content depend on his intentions in passing it. They also depend on his
beliefs about what he was doing, that is, his belief that he is making a rule for
the club. There is nothing in that to disqualify him from having knowledge of
the existence or content of rules he made any more than there is reason to deny
that other people can have that knowledge. Moreover, his knowledge of the
regulation is not knowledge of his own thoughts and intentions. It is knowledge
of a rule of the club. Nor need it derive from his memory of his own thoughts.
He may have forgotten that it is his regulation, and learn about it from col-
leagues. He, like anyone else, can have knowledge of the regulations of the
club. The underlying rationale of the independence condition does not warrant
Williams's claim that if knowledge is to be of what is there anyway it must be
independent of what is there in virtue of people's or the agent's own thoughts
and experiences. The objectivity of no domain of thought depends on this
condition.
But is not that conclusion premature? It may be thought that while in
general views and thoughts about matters whose existence depends on psycho-
logical or social facts are objective, there is a special case against the objec-
tivity of evaluative thought arising out of its possible dependence on social
facts. If it turns out—it may be argued-that if value propositions are capable of
being true or false then their truth-value depends on social facts, then norma-
tive or evaluative thought cannot be objective.
This particular objection to the objectivity of practical thought concentrates
on what I called the relevance condition. Practical thought appears to meet the
other conditions we canvassed. The real doubt about its objectivity is: does it
meet the relevance condition? Are there grounds that are not merely persuasive
but logically relevant to the confirmation or discontinuation of any practical
thought? The most promising attempt to provide an affirmative answer relies
on the use of thick concepts. But thick concepts seem to depend on a shared
culture with its shared acceptance of various values. Hence it seems that the
thick concept solution to the problem of the relevance condition makes the
truth of evaluative propositions depend on the social facts of shared views.
This is where the objection is directed: if it turns out that if practical proposi-
tions are true or false they are so in virtue of social facts, then they are not
capable of being either true or false.
The objection is a serious one. Thick concepts are indeed crucial to any
attempt to establish the objectivity of practical thought and its conformity with
the relevance condition. Consider the sort of reasons people give in support of
such propositions: Simon should be respected because he acted with great
dignity. Sarah deserves promotion for she solved a complicated problem with
subtlety and ingenuity. Robin acted well; he kept his head under pressure and
showed cool judgment and discrimination in handling his boss. This is a great
film: it is easy to understand without being trivial, and it combines humor and
wisdom. Typically just about all our evaluative reasoning is saturated with
thick concepts of a variety of kinds. Both conclusions and reasons for them are
typically expressed by the use of thick evaluative concepts. An account of the
relevant reasons that support or undermine evaluative or normative proposi-
tions will largely consist in an explanation of the relations between thick
concepts. But, and that is where the objection starts, mastery of thick concepts
depends on shared understandings and shared judgments. These shared judg-
ments both enable us to understand the meaning of thick evaluative terms, and
incline us to accept the legitimacy of their use. There is no independent way of
validating the legitimacy of the use of thick concepts. Hence, the validity of
evaluative propositions, if it depends on thick concepts, depends on shared
understandings and judgments, that is, on social facts. The truth or correctness
of value propositions cannot, however, depend on social facts. Such depen-
dence will make value judgments contingent, for the facts they depend on are
contingent, and arbitrary, for whether or not one has cogent reasons to accept
them will depend on the evaluatively arbitrary fact of one's membership in one
culture or another. Worst of all, if the truth conditions of evaluative proposi-
tions are contingent social facts, then they cannot be normative; they are
merely statements of those facts whose existence renders them true.
Since the normative cannot depend on social facts it cannot depend on thick
concepts, and given the absence of any other plausible account of the way
practical thought meets the relevance condition, we must conclude that it lacks
objectivity.
B. Types of Social Dependence

To evaluate the objection we need to investigate the ways, if any, in which
value judgments depend on shared understandings and judgments. This re-
quires relying on assumptions that cannot be explored here. The observations
that follow will, therefore, be tentative and subject to much further clarification
and further supporting arguments.45 I will be relying on one crucial distinction,
the distinction between practices that create or sustain the existence of goods
218 JOSEPH RAZ
and values and practices that are conditions for access to such good or values.
On the one hand, it is possible that values and goods are created or maintained
in existence by social practices and the shared beliefs and understandings that
are part of them. I will call practices that bring into existence goods or values,
or sustain them in existence, sustaining practices. On the other hand, it is
possible that shared understandings affect not the existence of values and
goods, but our ability to learn of them and perhaps to benefit from them. Such
practices control access to the values and goods concerned. Quite likely the
social dependence of values and goods takes many and varied forms. I will
briefly delineate four fairly typical types of case, starting with the closest to
dependence of value on practice.
"Socially Created" Goods: Local Goods

The existence of some goods seems to be clearly socially dependent. It used to
be important for young women, of a certain class, to walk only with short and
measured steps, or for men to wear wigs when out of their homes. There were
social advantages to behaving in accordance with conventions of manners and
disadvantages attached to flouting them. Conventions of manner, fashion, and
deportment vary over time and space. A particular form of manner or dress can
lose the meaning it has and acquire a different, even contradictory meaning. It
can have different, even mutually exclusive meanings in different places at the
same time. Moreover, within a single country it can have one meaning in one
subculture and another, even incompatible meaning in another subculture. In
the same place and at the same time what earns kudos in one class may earn the
disrespect of another, et cetera.
In matters of fashion and manners what is valuable depends on what people
do, that is, on conventional conduct, but also on shared attitudes and reactions,
on a shared meaning associated with different modes of conduct, deportment,
or dress. The dependence is multifaceted. There is more than one value to be
realized. In matters of fashion, for example, there are the values, say, of being
at the cutting edge of advanced fashion, the value of keeping up with new
fashions, the value of merging in the crowd, of being thoroughly conventional
and nondistinctive, the value of being charmingly old-fashioned, the value of
emphatically asserting one's indifference to fashion, the value of being asser-
tively nonconventional, and many more. When we come to manner the number
of different meanings different modes of conduct and address can display is
much greater.
The practices and shared meanings of a society determine not only the
benchmark of the normal, relative to which the conventional, defiant, cutting
edge, et cetera are defined. They also determine which meanings or values
(e.g., that of conventionality, defiance, cutting edge) exist relative to the
benchmark. Some cultures multiply meanings that turn on subtle distinctions.
In others only a few distinctions determine different meanings. In some so-

cieties meanings are fixed and rigid, attached to easily identifiable external
cues; in others they are more flexible, more complex in manifestation, allowing
for more individual variations and for easier mutation over time.
"Socially Created Goods": The Temporally Unbound Variety
The type of good or social meaning I have in mind here may differ from local
goods in degree only, but a difference there is: Take a game, for example,
chess, with the goods that playing it makes possible (being good at chess, being
a good chess tactician, ingenious at the end game, etc.). Chess was created at a
particular time. It makes no sense to think of it as having existed from the
creation of the world and only discovered at a certain time. It was invented,
created, and developed, not discovered. So it too is socially created or socially
constituted, and like all socially created goods and meanings it has an origin in
time, and its existence is contingent rather than necessary. It could not have
been invented. But in one respect chess is unlike manners. Once invented it is
with us forever. Of course it can be forgotten. Times may come when no one
will know what chess is, no one will know how to play it. Some games that
used to be played by people during the flowering of Maya civilization in the
Yucatan, for example, are lost to us. But that is mere loss of knowledge, loss of
access to a good. We can rediscover how to play these games. There is a sense
in which once the game is invented it remains in existence for ever, or at least
for as long as it is logically possible for it to be (if lost) rediscovered.46
Many other goods belong to this kind. Think of New York Jewish humor.
Enjoying it, being good at telling good New York Jewish jokes and witticisms,
or being able to display that sense of humor in one's conversation is valuable
and admirable. Like chess, New York Jewish humor and other goods of this
kind are generated by a particular culture and did not exist beforehand. This
sense of humor developed and emerged during a particular period in the history
of New York. It did not exist as a timeless form of humor to be discovered by
New York Jews in the nineteenth or twentieth century. Like all socially con-
stituted goods, New York Jewish humor enjoys a contingent existence. It might
not have come into existence. But, unlike fashion and like chess, once it has
been developed it remains in existence forever (or for as long as it can be -
should it be lost — rediscovered). It is quite likely that times will come when
people will not understand this form of humor. It also likely that a time will
come when there will be no one around capable of understanding it. But these
are contingent facts, which do not limit the possibility that it will be re-
discovered. What goes for New York Jewish humor goes for classical ballet,
the novel, opera, and a large number of other goods that are culturally con-
stituted in similar ways.
220 JOSEPH RAZ
Wherein lies the difference between the two categories of socially created
goods? In the possibility of enjoying them outside the cultures that bred and
sustained them. Suppose you discover a game that has long been forgotten or a
form of music no longer practiced in your culture. It is in principle possible to
discover what they were, to come to understand them and the cultures that
sustained them. That would involve understanding what was good in them.
Once one understands those goods one can enjoy them, one can engage in them
in the right way, bringing one the intrinsic benefits that they brought people in
the cultures where they were practiced. Naturally this requires further condi-
tions, one needs to be good at games, have willing partners to play with, be
musical, et cetera. But in principle those goods though created and practiced in
one culture can be enjoyed by people in other cultures, even by individuals who
live in cultures where the game or form of music did not take root. Not so with
regard to local goods. While we can learn and appreciate the fashions or
manners of previous times or faraway places, we cannot enjoy their benefits
unless they become the fashions and manners of our society.47
Goods That Are Not Socially Created — with Limited, Culturally

Conditioned Accessibility
Sunsets are not constituted by social practices and shared meanings, nor are
beautiful sunsets, nor the beauty of beautiful sunsets.48 There were beautiful
sunsets on earth at least since its atmosphere acquired its present constitution.
There probably were beautiful sunsets before the emergence of animal life on
the earth. It is possible, however, that there were periods when people did not
enjoy beautiful sunsets and did not find them attractive in any aesthetic way at
all. It is also likely that after people developed an aesthetic response to natural
phenomena, they differed in what they found valuable in nature and in the
values they found in what was valuable. Possibly there are psychological
universals that determine a degree of similarity in sensitivity across cultures.
But perusal of works of art suggests that whatever the similarities the
differences in aesthetic responses to, and appreciation of, nature among cul-
tures are considerable. Moreover, the responses to, and appreciation of, the
same natural phenomenon (e.g., sunsets) may be incompatible, in that no
person can appreciate all aspects of the beauty of a sunset at the same time,
without being ambivalent about the phenomena he is reacting to. For example,
the beauty of the sea, the same sea in the same conditions, can be perceived as
due to its tranquil and harmonious character by a holiday maker on an after-
noon walk, and due to its hidden awesome power by a fisherman who knows its
fickleness.
It is possible that the sea has all these qualities. This is easy to explain if
these are relational qualities, for example. If anything is beautiful if and only if
it looks beautiful, beauty is a property of appearance, a matter of how things

appear. The appearance of things is an objective matter about which mistakes
can be made. 49 One may, for example, mistakenly think that a good-looking
person is rather plain. Yet some properties of the appearance of things are
relational: some may relate to how they appear in daytime and others to their
nocturnal appearance, et cetera.
Objective though they are, not everyone can perceive the way things appear.
To be capable of perceiving the appearance of things one may need to be
acculturated in suitable ways. Let us think of the values that became prevalent
with the Romantic Movement in the eighteenth century. We can understand
them and relate to them, as we are the children of the Romantic Movement.
They are easier to appreciate for some than for others, and perhaps no one can
appreciate all of them any more. But since Romanticism, they have become in
principle available to people. Romanticism made forms of appreciation of
nature possible that were not possible before. But Romanticism no more
changed nature, or the appearance of nature, by creating new valuable features
in it than the development of humans or other animals with color vision added
color to objects in the world. Red is not made red by the existence of people
who can perceive it as red, and beautiful sunsets are not made beautiful by the
existence of people who can appreciate their beauty. Culture does not create the
beauty of nature. It merely enables us to become aware of it, to come to
understand and enjoy it.
Goods with Universal, Culturally Conditioned Access

Some goods may be even more independent of social practices: not only are
they not created by social practices, access to them is not restricted by social
practices either. Even though sunsets are not man-made, nor is their beauty, the
appreciation of their beauty is made possible by culture. One may feel that they
are as local and time-bound as all the socially created goods briefly discussed
above. Whether culture determines access rather than existence can be felt to
be of little significance. What matters is accessibility. Consider the relation of
value to personhood. Persons are rational beings, that is, they possess the
ability to perceive that some things are good or bad in various ways and to
respond appropriately. They should find out enough about value to enable them
to conduct themselves sensibly. They bear responsibility to find out enough
about value, but that responsibility is limited by accessibility. They cannot
come to appreciate normative aspects of the world that are not accessible to
them. Three points seem to bear this out:
1. People cannot be blamed for not being guided by values they could not
know about.
222 JOSEPH RAZ
2. People's conduct cannot be morally wrong for disregarding or violating

values or rules they could not in principle know.
3. It is, of course, possible that lack of access to values will impoverish a
person's life and will render it less rich and admirable, and so forth, than it
could have been had that person the ability in principle to understand and
engage in more valuable activities. Similarly, lack of access to values may
coarsen a person's character.
Access to a value, rather than its existence, is decisive in evaluations that

presuppose responsibility, be they evaluations of people's actions, life, or
character. Furthermore, for evaluations of life and character that do not presup-
pose responsibility, evaluation of them as limited, or coarse, or rich and re-
fined, it does not matter whether the absence of a good from a person's life was
due to its nonexistence or to its inaccessibility. Either way, the absence of the
good from a person's life may be coarsening or impoverishing. These observa-
tions would justify the view that so far as the kinds of values and goods
considered so far go there is no significant normative difference between
socially created values and those that are not so created, but whose accessibil-
ity is socially dependent. If this is all there is to say on the subject, then a fairly
strong form of social relativism might be vindicated. Are there any normative
considerations which are universally accessible, that is, which are within the
reach of all people, of whatever time or place, at least in principle?
A full answer to the question is beyond the scope of the present discus-
sion.50 For our purposes suffice it to say that one objection to the possibility of
universally accessible normative considerations is mistaken. Consider the sort
of normative considerations often advanced as universal and timeless. For
example: It is always wrong to murder an innocent person.51 If we can judge
people to have acted wrongly in having committed murder we must assume
that it was possible for them to know that they should not perform the actions
that are the murder of innocent people. But we need not assume that they must
have been able to understand the concept "murder," "person," "innocent,"
"intention," and "killing," which we use to articulate and explain this rule.
Arguably, even before these concepts were available to people they had other
ways of categorizing mental states, other ways of marking transgressions, other
ways of marking animals that, as we now know, belong to the species Homo
sapiens. Their concepts and generalizations may have been based on false
beliefs, and we may find them inadequate in many ways. But they may have
enabled them to know that the acts that in fact constitute murder of innocent
people are wrong. That is enough to establish that they had the access to the
norm that is a precondition of being able to blame them for its violation. If
people at all times had access to the norms, in one form or another, then we
have here an example of a consideration, to which there is universal access,
though that access is culturally determined.
C. Social Dependence and Objectivity

What conclusions regarding the objectivity of value can be drawn from this
fourfold classification, assuming that it is basically sound? First, we need to
distinguish the social dependence of values from the social dependence of
access to them. Second, while all but the most abstract values, goods, and other
normative considerations are referred to by the use of thick terms, only some of
them are socially created.
My observations at the beginning of this part52 showed that social depen-
dence does not necessarily pose a threat to objectivity. I believe that most
people worry less about the objectivity of local goods, as I called them, than
about the objectivity of socially created goods that are not local or those that
are not socially created at all. The doubt about the objectivity of value does not
arise from the contingency of the social on which all value allegedly depends,
but from the suspicion that since knowledge of what is valuable depends on
mastery of "thick" concepts, it is suspect.
One additional source of suspicion that values that are socially constructed
cannot be objective should be mentioned. It is sometimes assumed that if some
good or value is socially created and sustained, then the reason for its being
good is that it is accepted as good, that people think it is good, or variants on
these. This is not at all so. Consider chess: neither the reasons for making one
move rather than another in the course of the game nor the reasons for playing
the game involve any appeal to the social practice that created the game.
Similarly, we can give reasons for thinking one joke better than another, or one
piano sonata better than another, but they will not and should not include an
appeal to the fact that these goods are socially created. A joke is good because
it is funny, because it has more than one sting, because it takes the mickey out
of the pompous, who deserve the treatment, because it points out human
foibles, and so forth. A piano sonata is good because it is full of musical ideas,
ingenious variations, suspense and surprise, tension and its resolution, because
it mixes moods, speaks to the emotions, and so on. The fact that a good is
socially constituted is no more reason for its value than the fact that a chair is a
product of human design is a reason for thinking it a good chair. The same goes
for all artifacts and for all socially constituted goods (see the qualification in
the next paragraph). It must be so, for otherwise it would not be possible to
distinguish good social creations, or artifacts, from bad ones.
It is true that practices and shared beliefs are (part of the) means by which
we identify goods that are socially created or maintained. But identifying what
is the good is not the same as explaining what it is that makes it good or why it
is good rather than bad.
Local goods may be thought to be a special case since they tend to be
conventional goods. That is why some feel less suspicious of them. They feel
that we understand the nature of conventions. After all they have received the
224 JOSEPH RAZ
seal of approval of game theory. Moreover, it may be thought that with conven-
tional goods the explanation for what is good about this mode of conduct, or
dress, et cetera, is clear: it is good because that is how people generally behave,
how they dress, and so on. This is, however, at best only part of the explana-
tion.53 The rest has to account for the reason for conforming to the conven-
tional norm. We understand conventions only when we understand the good
they are thought to achieve and whether they in fact lead to good or ill.
The values served by conventional goods explain what, if anything, is good
in the conventions and enable us to distinguish between good and bad conven-
tions. I emphasize this point for it shows that the explanation of the goodness of
any good or valuable thing or option has to be relatively independent of the
social practices that create that good. Only thus can we acknowledge that social
conditions can also lead to bad practices, which will be mistakenly taken by
their participants as good ones. In "Moral Change and Social Relativism"54 I
suggested a different argument to the same effect, namely, that the inherent
intelligibility of values means that they have to be somewhat independent of
social practices. The two arguments are two aspects of the same thought.
These arguments do not show that normative properties are not socially
created. They show that they are not all conventional, and that their normative
nature must be explained normatively and cannot be explained in nonnorma-
tive terms. But there can be social practices or other social phenomena that can
only be identified in normative terms, and such social practices can be said to
give rise to new normative properties consistently with the arguments above.
We learn that not all goods are socially created not from the argument above
but by examining the nature of the various goods (as was sketched in the
previous section). The preceding argument is meant to show that even if all
goods are socially created (and they are not) it does not follow that the reasons
that explain why they are good, what makes them good, consist in an appeal to
the fact that the relevant social practices exist. They must consist in pointing to
good-making properties of the social practices concerned.
These remarks may lead to the opposite objection, namely, that socially
created goods are not socially created at all. The goods involved in them are
timeless, only the ways they are realized are local. But while that may be true
of some instances, it is not generally true. For example, observing rules of
polite behavior tends to make one feel comfortable in social situations and
facilitates interaction with others. That good is neither conventional nor so-
cially created. It is timeless. But to say this is to abstract from the differences
between different societies and different codes of good behavior. The abstract
value of facilitating interactions may be the correct value to apply to the
situation, yet it leaves out information that only a concrete and localized
description of the good of polite behavior in one or another social setting can
convey. Concrete socially created values must be subsumed under abstract
universal values, or they will be unintelligible. But they are no "mere" in-
stantiations of universal values. They are distinctive specific goods, which can
be enjoyed only if created by social practices.55
D. Objectivity and Thick Concepts

Can the objectivity of "thick" concepts be defended? Suspicion of their use can
now be seen to be independent of any doubts arising from the alleged social
dependence of all value. But suspect they remain. I have already dismissed
some of these suspicions. It is true that what we believe may well depend on
which evaluative concepts we became familiar with first and which evaluative
beliefs we acquired first. That is a result of the path-dependence of epistemic
justification, a feature of epistemic justification in all fields of knowledge.
Furthermore, it is true that there is no hope of a Peirceian convergence regard-
ing evaluative beliefs. There is no reason to think that there is a meaningful
ideal situation in which the views of all competent believers will converge. Yet
again, we saw reason to doubt the cogency of the requirement of such con-
vergence as a precondition of the possibility of knowledge. Both these observa-
tions mean that there is an inevitable contingency not only about our evaluative
beliefs, but also in the account of their epistemic justification. What we know
and what we do not know is partly a matter of the accident of our circum-
stances, and even the best epistemic justification possible cannot rid our beliefs
of an element of luck. But epistemic luck is a feature of the conditions of
knowledge in general. It is not a circumstance special to evaluative beliefs, and
it does not negate the possibility of knowledge.
There remains, however, another, though related, source of doubt in the
objectivity of evaluative beliefs, if that objectivity depends on deploying
"thick" concepts. There is the familiar charge that the thick concepts of
different cultures, or of different moral or religious outlooks, are incommensu-
rate. Concepts that belong to one of these systems of thought cannot be ex-
plained in terms of concepts belonging to another. These thick concepts, the
argument proceeds, are essential for the expression of the systems of thought to
which they belong. The views of each system cannot, therefore, be expressed
or explained using the concepts of another system. This leads the objector to
the conclusion that even when one can be confident that one's evaluative
beliefs are justified from within one's own system of thought, one has to
acknowledge that that is only a relative justification. Incompatible beliefs
enjoy equally cogent justification within other systems of thought. Further-
more, the arguments refuting the other system of thought as a whole to one's
own satisfaction are themselves relative to one's own system of thought. The
rival system is not only impregnable in its own terms, but it quite likely
contains within itself cogent arguments refuting one's own system of thought.
According to the objector it follows that there is no rational way to adjudi-
cate between different systems of thought.56 This would seem inconsistent
226 JOSEPH RAZ
with the objectivity of evaluative thought. For, the objector points out, many
conflicting evaluative thoughts have equal credentials. Therefore, none of
them can be correct or true. The argument is not merely that they cannot be
known to be correct or true. That would have been the case had there been
nonrelational grounds establishing one rather than the other which people
cannot, not even in principle, come to know.57 But, according to the objector's
argument, the failure is not of ability to know. There are no such grounds,
therefore there is nothing to know.
Much has been written about the coherence of the incommensurability
claim that underlies the objection. I will put such doubts to one side. The
objection is also vulnerable on other grounds. It relies heavily on the inability
to explain the concepts of one culture, or system of thought, using those of
another. But how important is this limitation? The argument is often discussed
as revolving not on the relations between concepts belonging to two different
systems of thought, but on the relation between the meanings of terms in two
languages. In the development of languages, however, new terms are often
added whose meaning cannot be precisely explained using the other terms of
the language. We learn their meaning in part through use, and by ostension, and
only in part by locating their meaning relative to that of other words.58 We do
the same when we encounter concepts in a foreign culture. Words are added to
the language often borrowed from the other language, whose meaning is
learned not exclusively through explanation, but by ostension, and through
habituation to aspects of the alien culture that they signify.
There is little doubt that often we fail to understand concepts embedded in a
culture or system of thought that is alien to us. But is there any reason to think
that even given favorable conditions we could not master them? That they
cannot be exhaustively or satisfactorily explained using our concepts does not
establish that conclusion, for we can learn them directly, by being exposed to
their use, rather than through translation. Sometimes we could do so by actu-
ally living in the alien society; at other times it is possible to learn their
meaning by learning about that society and reconstructing in imagination, or
simulation, its ways and beliefs.
To claim that members of one culture cannot in any way come to understand
the concepts of another amounts to the implausible claim that people have the
capacity to acquire the concepts of one culture only. Once they have done so
their conceptual capacity is exhausted or perhaps blocked. This seems an
implausible supposition.
I can think of only one possible reason for it. Complete understanding of a
concept — and it is important to remember that understanding is a matter of
degree — involves knowing its relations with all the concepts one understands.
Take, for example, the relation "being incompatible with." One does not under-
stand what it is to be blue if one does not know that it excludes being red, nor
what it is to be just if one does not know that it excludes being unfair. It follows
that in a way the more concepts one has already acquired the more difficult it is
for one to acquire new ones, for one would have to understand whether they are
or are not compatible with those one has already mastered. Could it not there-
fore be the case that for this reason those who mastered the concepts of our
culture cannot understand those of ancient alien civilizations? Our inability to
understand their concepts is no sign of reduced mental ability. Rather, to
understand their concepts we have to know the degree to which they are
compatible with ours. Members of those ancient civilizations themselves did
not have that knowledge. Since they did not know our concepts they did not
have to know how our concepts relate to theirs in order to understand theirs. We
do, and this is why we cannot understand their concepts.
At this point the argument that systems of thought are incommensurate
because their concepts are not mutually translatable has been abandoned in
favor of the assertion that systems of thought are incommensurate because no
one can master the concepts of more than one of them. The new contention is,
however, implausible. It flies in the face of the evidence. There were and are
people who inhabited more than one culture and understood both. It also
overlooks the fact that our own culture contains concepts derived from
different systems of thought, which have not merged together. While some of
us do not have use for concepts such as grace, sacred, blessed, prayer, and
others, for the most part we manage to understand them, when we try. Finally,
the conclusion exaggerates the conceptual insularity of different cultures. To be
sure there are many culture-specific concepts, that is, concepts that evolved in
one culture and have no parallels in others. But they are embedded in a
conceptual framework that includes many concepts bridging the cultural gap or
having at least near relatives in other cultures. It seems safe to dismiss the
thesis that no person can master concepts of more than one culture and, with it,
the thesis that there are past or present cultures whose concepts cannot in
principle be understood by us. 59
It may be worth pointing out that the considerations I advanced here are
consistent with the view expressed in Part II about the essentially local charac-
ter of evaluative concepts, or of many of them, and of the nonexistence of a
Peirceian ideal condition from which everything can be known. There I was
arguing that it is impossible, even if only because of the impossibility of
knowing the future, for anyone to master all the evaluative concepts. Here, I
argued that there is no general reason to think that no one can come to know
and understand more than one of the existing systems of thought. This conclu-
sion can be reinforced to the effect that no past or present human culture and its
concepts are beyond the comprehension of people for whom it is an alien
culture. There are no human cultural islands that cannot be understood by
anyone other than their members.
At this point one may be tempted to resuscitate the challenge to the objec-
tivity of evaluative thought, while abandoning the conceptual incommen-
228 JOSEPH RAZ
surability thesis. One may argue that (1) all justification is internal to a system
of thought, and that (2) there is no way of adjudicating between beliefs belong-
ing to different systems of thought, and since (3) some beliefs belonging to
different systems of thought are incompatible and cannot both be true, on pain
of contradiction, it follows that no beliefs are true. As before, this argument
presupposes (1) that evaluative beliefs are true only if there are reasons that
explain how it is that they are true, and (2) that evaluative reasons are in
principle available to people, though not necessarily to everyone. For reasons
we cannot explore here these seem reasonable assumptions.
This challenge to the objectivity of evaluative thought is ill-founded. It is
true that all justifications are relative: they are all addressed to an actual or
potential audience, addressing the doubts of that audience and relying on what
that audience accepts. The relativity ofjustifications is itself a result both of the
parochial character of (many) evaluative concepts and of the path dependence
of epistemic justifications. So long as all the different justifications are consis-
tent, that relativity does not undermine the objectivity of evaluative thought.
Nor does it follow from the relativity of justification that there is no way of
adjudicating between incompatible thoughts or beliefs.
More generally I suspect that not infrequently the sense that incompatible
evaluative propositions have equal credentials and therefore neither can be true
results from a failure to distinguish two types of incompatibility. Two evalua-
tive propositions can be (1) inconsistent with each other, or they can (2)
express values, virtues, or ideals that cannot both be realized in the life of a
single person. The second kind of incompatibility does not undermine the
objectivity of evaluative thought. It merely leads to value pluralism.
The first kind of incompatibility is relevant to the debate on objectivity.
Given the two presuppositions I accepted it must be the case that only one of
two inconsistent propositions can be adequately supported by reasons. But we
have to distinguish epistemic from constitutive reasons (though they may
overlap). Two people can have adequate epistemic reasons for accepting each
of two inconsistent propositions. These could be as banal as that different
people, whose judgment they have good reason to trust but whose beliefs are
inconsistent with each other, have advised them. Even constitutive reasons can
conflict as there may be something to be said for both sides of an evaluative
dispute (constitutive reasons are often prima facie). But evaluative proposi-
tions and their negations cannot both enjoy adequate or completely vindicatory
support by reason. None of the considerations I canvassed in this article sug-
gest that inconsistent evaluative propositions do enjoy adequate or vindicatory
support by reasons.
This has been a long and mostly negative discussion. There may be reasons
for rejecting the objectivity of evaluative thought. I offered various consider-
ations to show that many arguments offered as justifying such rejection do not
succeed in doing so. The constructive contribution of my arguments is in
showing how doubts about objectivity can arise out of a misunderstanding of

objectivity and of justification. If you set us an unachievable target you will be
able to show that we fail to achieve it. It is instructive to see how we were
victims of an overdemanding and overrigid conception of justification and
objectivity. But the misguided allure of these "high ideals" is not confined to
evaluative thought. Those who succumb to them are entrapped by confusions
in other areas of thought as well.
Notes
1. Unless otherwise indicated, when referring to what people think, I have in mind the
thought (or thought-content) rather than the act or activity of thinking.
2. Bear in mind that it would be odd actually to talk of people's impartiality unless someone
challenged it or unless the circumstances of the action invite doubts about it.
At the hands of some contractarian or constructivist philosophers impartiality extends
in another direction as well: in matters of ethics what impartial judges accept is thereby
made true. But I will not explore the contractarian uses of the notion.
3. Though nothing much depends on the way domains are demarcated. The criteria of
objectivity determine whether a domain is objective or not, whichever way it is carved.
4. For example, Marmor, "Three Concepts of Objectivity" in A. Marmor (ed.), Law and
Interpretation (Oxford: Clarendon, 1995).
5. D. Wiggins, "Objective and Subjective in Ethics, with Two Postscripts About Truth" in
B. Hooker (ed.), Truth in Ethics (Oxford: Blackwell, 1996), p. 36.
6. Wiggins's own view is explained in "What Would Be a Substantive Theory of Truth?" in
Z. van Straaten (ed.), Philosophical Subjects: Essays Presented to P. F. Strawson (Ox-
ford: Oxford University Press, 1980).
7. Responsibility, rather than more narrowly moral responsibility.
8. Whether this applies to epistemic objectivity in all its variants is open to doubt. Perhaps
the capacity to be epistemically objective (one version of epistemic objectivity noted
above) is an exception. The question whether or not one is responsible for not being
objective in that sense involves considerations of aspects of responsibility that cannot be
pursued here.
9. And this may account for the tendency to classify them as subjective.
10. Some would make this the first and most important condition of objectivity. One way of
formulating it is: A domain is objective if thoughts belonging to it deal with matters about
which there are propositions that are true or false (or correct or incorrect, or more or less
correct or incorrect). I state the condition in this way, rather than by saying that those
thoughts express propositions that are true or false, in order to allow for the possibility of
thoughts that are objective, and belong to an objective domain, but express propositions
that are neither true nor false.
11. B. Williams, Descartes (London: Harvester, 1978), p. 64.
12. I am using a realist mode of expression ("about a reality") in a weak sense, in which we
could say of arithmetical propositions that they are about numbers and their relations,
without being committed to Platonism about numbers. This weak sense is explained in
the second half of the independence condition.
13. In "What Would Be a Substantive Theory of Truth?" (Philosphical Subjects), Wiggins
suggests as a mark of truth that a statement's being true or not is independent of any
230 JOSEPH RAZ
particular subject's means of appraising its truth-value. This goes beyond the indepen-
dence condition. Practical or evaluative thought does not conform with this condition
under some natural interpretations of it, but is objective. Some of the discussion in the
sequel bears on the reasons for rejecting Wiggins's condition.
14. Many would claim — Williams observes — that we are now familiar with the situation of
doing with less than an absolute conception, and can, as modern persons and unlike the
ambitious or complacent thinkers of earlier centuries, operate with a picture of the world
that at the reflexive level we can recognize to be thoroughly relative to our language, our
conceptual scheme — most generally, to our situation. But it is doubtful to what extent we
really can operate with such a picture, and doubtful whether such views do not implicitly
rely, in their self-understanding, on some presumed absolute conception, a framework
within which our situation can be comprehensively related to other possible situations.
{Descartes, p. 68.) We should agree that the notion of "relative reality" does not make
much sense. There is no interesting sense in which something can be the case relative to
us, but not be the case relative to others, unless, as Williams observes, such claims
presuppose a common frame of reference.
15. Such an isolationist position seems to be endorsed by Dworkin in "Objectivity and Truth:
You'd Better Believe It," Philosophy & Public Affairs 25 (1996), p. 87.
16. It is not clear whether as formulated by Williams the condition would rule out perspec-
tivalist accounts of truth or of reality that are not self-contradictory. For a claim that there
can be coherent perspectivalism see S. D. Hales, "A Consistent Relativism," Mind 106
(1997):33.
17. This also explains why there is safety in numbers. The more criteria we have the firmer
our grasp of the concept, for cumulatively they eliminate the possibility of misunder-
standings relative to known questions. On the other hand they open up more possibilities
of hitherto unthought of questions regarding which they are vague and need refinement.
18. Cf. B. Williams, "Truth in Ethics" in Hooker (ed.), Truth in Ethics, p. 25. (Note that this
example merely illustrates the possibility of additional, logically independent, condi-
tions. It does not illustrate any puzzles they may be called for to dispel). My espousal of
the long route is in line with Dworkin's position in "Objectivity and Truth." There he
warns against taking it for granted that conditions for the objectivity of one domain (say
relating to material objects) apply also to all other domains (say that which relates to
mathematics or to values).
19. This objection is based on claims endorsed by several philosophers, including Dummett,
Wiggins, Williams, and Crispin Wright in Truth and Objectivity (Cambridge, Mass.:
Harvard University Press, 1992).
20. Truth and Objectivity, chaps. 1 and 2.
21. Cf, for example, B. Williams, "Truth in Ethics," pp. 19—20 and elsewhere.
22. In the sense with which the phrase was introduced by J. Mackie in Chapter 1 of Ethics:
Inventing Right and Wrong (London: Penguin, 1977).
23. For example, Marmor, "Three Concepts of Objectivity," and Pettit and Leiter in the
present volume.
24. "Truth in Ethics," in Truth in Ethics, p. 3. It may be that Wright's argument here is the
argument Dworkin tries to make in his "Objectivity and Truth: You'd Better Believe It."
25. Given that mastery of concepts is a matter of degree, everything I say in the text has to be
modulated to allow, e.g., for a more imperfect mastery of a greater number of concepts.
26. The View From Nowhere (New York: Oxford University Press, 1986), p. 5. He also says
that, "To acquire a more objective understanding of some aspect of life or the world, we
step back from our initial view of it and form a new conception which has that view and
its relation to the world as its object. In other words we place ourselves in the world that is
to be understood" (ibid., p.4). Here Nagel seems to be saying that a belief about the world
is more subjective than a belief about my having a belief about the world. But my beliefs
about the content of the theory of quantum mechanics are not more subjective than my
belief that I have certain beliefs about quantum mechanics. They are simply different. I
will assume that the quotation above is simply an unhappy formulation of the idea I
quoted in the main text.
27. Ibid., p. 4.
28. This issue is relevant to the standing of perceptually dependent properties, e.g., color
properties, and I will not pursue it here.
29. From "A Critical Review of Berkeley's Idealism," in Charles S. Peirce, Selected Writings
(Values in a World of Chance), ed. Philip P. Wiener (New York: Dover, 1966), p. 82;
Descartes, p. 244.
30. Ibid., p. 245.
31. Ibid., p. 248.
32. See H. Putnam, "Bernard Williams and the Absolute Conception of the World" in
Renewing Philosophy (Cambridge, Mass.: Harvard University Press, 1992), p. 80, and J.
McDowell, "Aesthetic Value, Objectivity, and the Fabric of the World" in E. Schaper
(ed.), Pleasure, Preference, and Value (Cambridge: Cambridge University Press, 1983),
P . i.
33. Ethics and the Limits of Philosophy (London: Fontana Press/Collins 1985), p. 140.
34. Ibid.
35. It is illuminatingly discussed by Wright in Truth and Objectivity, e.g., pp. 88—94.
36. C. Wright's condition of cognitive command {Truth and Objectivity, pp. 92-93)
disregards the existence of rational underdetermination in its assertion that a discourse
exhibits cognitive command only if all disagreements can be explained by "divergent
input." On pp. 95ff, Wright has an interesting discussion of the related topic of the degree
to which reason can be permissive only, but he does not fully adjust his conclusions to
allow for its insights.
37. "Objectivity and the Science-Ethics Distinction," in Putnam, Realism with a Human
Face (Cambridge, Mass.: Harvard University Press, 1990).
38. For a critique of the absolute conception that argues that science too needs perspectival
concepts, see Putnam, "Objectivity and the Science-Ethics Distinction." Note that to
establish his case Putnam has to show not only that different systems of representations
can be used in science, but also that they employ what I call parochial concepts.
39. There may or may not be reasons to think that some perceptual concepts, those designat-
ing the so-called secondary qualities, designate relational properties of things. For exam-
ple, there may or may not be reason to think that they designate how things when
observed strike the observer, rather than how they are when not observed. Williams
advances some arguments to that effect, and their examination is irrelevant to my argu-
ment here. The only relevant point is that the fact that perceptual concepts are local is not
such an argument.
40. This is one of the points made by McDowell, "Aesthetic Value, Objectivity, and the
Fabric of the World."
41. It may be worth noting that the rejection of this option is not just an endorsement of the
232 JOSEPH RAZ
"Neurath's boat" understanding of the pursuit of knowledge. Those who accept that we
must start from here, i.e., from the beliefs we happen to have, can be wedded to an ideal
convergence test or some other test of the kind here rejected.
42. Perhaps it is helpful to warn here that this reference to a holistic approach is not an
endorsement of an epistemic approach based on coherence in any shape or form. I have
explained some of the reasons for this in Ethics in the Public Domain (Oxford: Clarendon
Press, 1994), essay 13: "The Relevance of Coherence."
43. Williams, Descartes, p. 64
44. I am assuming a club which is not legally incorporated.
45. They will also be crude in not drawing some elementary distinctions. I will refer to
goods, values, valuables, norms, considerations which determine that an action is right or
wrong, etc., without trying to distinguish between them. For present purposes such
distinctions are immaterial. Furthermore, I will use examples, implying that they are
examples of genuine goods, or values. Their purpose is to illustrate abstract general
points. It is not my suggestion that everything that is endorsed by a local culture is good
for that reason, and the examples can be substituted with others by those who doubt their
credentials. Finally, I will consider only beliefs about intrinsic goods and will avoid any
reference to instrumental goods.
46. It is true that chess may be forgotten and then be reinvented rather than rediscovered.
That would be the case if its reemergence is independent of knowledge of its previous
existence. But this seems to be consistent with my claim: So long as it can be re-
discovered there is something to rediscover.
47. Two qualifications do not erode the contrast I draw in the text. First, there are "deviant"
ways of enjoying local goods in alien cultures: it could, e.g., mark one as an eccentric.
Similarly, there is usually a whole network of meaning associated with all goods —
socially constituted or not — which is lost when they are moved from one context to
another. Playing chess in seventeenth-century Paris invokes different meanings than
playing in Washington Square in New York City today.
48. I will call any good, norm, or value that is not socially constituted "universal," even
though it may not be strictly universal, as its existence may be conditioned by other
factors.
49. See generally, P. M. S. Hacker, Appearance and Reality (Oxford: Blackwell, 1988).
50. See my discussion of the universality of value in "Moral Change and Social Relativism,"
in E. F. Paul et al. (eds), Cultural Pluralism and Moral Knowledge (Cambridge: Cam-
bridge University Press, 1994). Among other things I discuss there the possibility of
applying judgements that presuppose responsibility, to people, their life, character or
actions, by values that were not created until a later time. In the present discussion I
ignore that possibility.
51. I present a simplified and inaccurate norm, to avoid complicated substantive moral
issues.
52. See "Thick Concepts and the Social Connections."
53. At best for, as was noted in discussing local goods, typically a single conventional
standard provides an opportunity for a variety of goods, some of which depend not on
conformity but on deviation from it (e.g., the good of being an unconventional person, or
one with a strong individual taste).
54. In Cultural Pluralism and Moral Knowledge.
55. I have argued this and some other points made in the current section and the next one in
"Moral Change and Social Relativism."
56. My formulation of the objection is not meant to capture the precise argument of any
writer, though it borrows from points advanced by a number of philosophers. For exam-
ple, some aspects of such a view were elaborated by A. Maclntyre in his Gifford lectures.
Davidson's "On the Very Idea of a Conceptual Scheme" is the best-known attempt to
refute such views. But see Hacker, Appearance and Reality.
57. The argument presupposes, correctly, that evaluative beliefs are not about brute facts.
What value things have is an intelligible matter. Necessarily if something has a value in
one way or another then there are reasons which explain this fact.
58. The Oxford English Dictionary defines "nerd" as "An insignificant or contemptible
person, one who is conventional, affected, or studious; a 'square', a 'swot.'" Surely this
misses out a lot that is essential to the understanding of the term in contemporary,
especially American culture. The OED instances the following (from an ad in the New
York Times, 1978): "The nerdiest nerds on TV are really smart cookies," a quote made
mysterious rather than explained by their definition.
59. This remark is confined to cultures of creatures who share our perceptual capacities and
emotional make-up.

Joseph Raz - Notes On Value and Objectivity

Uploaded by

Copyright:

Available Formats

Joseph Raz - Notes On Value and Objectivity

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

Joseph Raz - Notes On Value and Objectivity

Uploaded by

Copyright:

Available Formats

Objectivity in Law and Morals

Cambridge University Press

© Cambridge University Press 2001

This publication is in copyright. Subject to statutory exception

First published 2001

Library of Congress Cataloguing in Publication data

ISBN 978-0-521-55430-5 hardback

I. The Long Route

B. Some Senses of the Objective/Subjective Contrast

such people or causes are dear to them.2 Epistemic objectivity as a capacity to

C. Some Conditions of Domain Objectivity

1. THE POSSIBILITY OF KNOWLEDGE CONDITION. Only if a domain is objective

2. THE POSSIBILITY OF ERROR CONDITION. A domain is objective only if

3. THE POSSIBILITY OF EPISTEMIC OBJECTIVITY. Only if a domain is objective

4. THE RELEVANCE CONDITION. Thoughts are objective only if there can be

i. to talk of thoughts as being correct, true, or as being errors and mistakes.10

The fourth condition allows us to apply the distinction between what is

5. THE INDEPENDENCE CONDITION. Thoughts are objective only if they are

6. THE SINGLE REALITY CONDITION. The domain of objective thoughts is sub-

7. THE POSSIBILITY OF IRRATIONALITY CONDITION. That is, that it is possible

D. Defending the Approach

Third, the condition of relevance makes the objectivity of a domain depen-

II. The Role of Parochial Concepts

A. Is Parochial Knowledge Possible?

B. Nagel on the Objective and the Subjective

C. Williams and the Priority of Nonparochial Concepts

are engaged in pure inquiry. In the nature of things if the convergence is to

There is another, perhaps more radical, limitation on any convergence re-

D. The Siren Call of Epistemic Absolutes

the mercy of epistemic luck since we may be rationally justified in accepting

E. Epistemic Anxieties and the Long Route

III. The Illusion of the Authority of the Social

A. Thick Concepts and the Social Connection

B. Types of Social Dependence

"Socially Created" Goods: Local Goods

In others only a few distinctions determine different meanings. In some so-

"Socially Created Goods": The Temporally Unbound Variety

Goods That Are Not Socially Created — with Limited, Culturally

it looks beautiful, beauty is a property of appearance, a matter of how things

Goods with Universal, Culturally Conditioned Access

2. People's conduct cannot be morally wrong for disregarding or violating

Access to a value, rather than its existence, is decisive in evaluations that

C. Social Dependence and Objectivity

D. Objectivity and Thick Concepts

showing how doubts about objectivity can arise out of a misunderstanding of

You might also like