From Morality to Metaphysics

This page intentionally left blank

From Morality
to Metaphysics

The Theistic Implications
of our Ethical Commitments

Angus Ritchie

Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom

Oxford University Press is a department of the University of Oxford.
It furthers the University

’s objective of excellence in research, scholarship,

and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries

# Angus Ritchie 2012

The moral rights of the author have been asserted

First Edition published in 2012

Impression: 1

All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above

You must not circulate this work in any other form
and you must impose this same condition on any acquirer

British Library Cataloguing in Publication Data

Data available

Library of Congress Cataloging in Publication Data

Data available

ISBN 978

–0–19–965251–8

Printed in Great Britain by
MPG Books Group, Bodmin and King

’s Lynn

To my parents,
with love and gratitude

Acknowledgements

Ralph Walker

’s inﬂuence on this work is manifold—in the personal and

philosophical inspiration he has provided since my undergraduate days;
in his patience and diligence as a doctoral supervisor, and in his typically
fair-minded insistence that he should have a co-supervisor with rather
different views. As this book is a reworking of my doctoral thesis, my
ﬁrst expressions of thanks are to him and his co-supervisor Sabina
Lovibond (who deepened my engagement with the work of John
McDowell and Philippa Foot in particular). Thanks are also due to
Sabina Alkire, John Deacon, Rob Gilbert, Amanda Greene, Mark
Hopwood, Philip Krinks, Michael Piret, David Staples, Richard Taylor,
and Mark Thakkar for stimulating discussions and wise counsel; to my
doctoral examiners Tim Mawson and John Cottingham for their very
helpful advice and encouragement, and to Peter Momtchilloff for his
assistance in the process of turning my thesis into a book.

It is often argued that faith must be kept out of politics, for reasons both

practical and theoretical. Practically, it is said that religion is a divisive
force. Theoretically, it is claimed that religious beliefs (unlike other kinds
of conviction) are immune to rational debate and correction. I have now
spent over a decade in ministry in East London, all of it in churches
organizing for social justice alongside other faiths. The witness of those
churches has been a crucial inspiration for this work. Community organiz-
ing offers a very practical example of the positive role of faith in public life.
In this book, I seek to complement this with a more theoretical argu-
ment

—showing the openness of religious belief to rational debate.

Through her love, encouragement, and practical support, Jennifer Lau

has contributed more to this work than she knows. It was her misfortune
that my last year of thesis writing was the

ﬁrst year of our relationship, and

the year of its conversion to a book was the

ﬁrst year of our marriage. Her

loss has undoubtedly been my gain, and this acknowledgment comes with
heartfelt love and gratitude. I am also thankful to my parents-in-law,
Chau-Kiong and Kim-Heung Lau, for their hospitality as the

ﬁnal draft

was prepared.

ﬁrst learnt of God and goodness from my family, and from the wider

community in which I spent my childhood. I am deeply grateful to them
all

—to the Church of Scotland congregations in Ardersier and Petty, to

my brother John, and above all to my parents, Garden and Kathleen
Ritchie, to whom this book is dedicated.

Limehouse, East London

The Feast of the Epiphany, 2012

ACKNOWLEDGEMENTS

vii

This page intentionally left blank

Introduction

1. Con

ﬂicting Impulses in Moral Philosophy

What are the raw materials which

‘moral truth’ is constructed? Is it

constituted (or generated) by facts about human sentiments and cultural
conventions? Or is there an objective moral reality, existing prior to those
sentiments and conventions, in virtue of which some moral statements are
true and others are false?

Faced with these questions, most students of philosophy feel two

con

ﬂicting impulses. The ﬁrst might be called the pull of reductionism.

At its crudest level, this re

ﬂects the common conviction that the only

things really

‘out there’ are the entities described by the physical sciences.

If this conviction is correct, then what people call

‘moral truth’ will be

accounted for by reference to human sentiments, preferences and cultural
conventions, and not to an ontologically independent moral reality.

Competing with this

ﬁrst impulse is the pull of objectivism. Students

will no doubt accept that some of their attitudes are culturally conditioned.
However, they will

ﬁnd it hard to do justice to their deepest commitments

without making claims which are more metaphysically ambitious. Sooner
or later, whether in the seminar room or the college bar, the case of the
Holocaust or the Killing Fields will be raised. Few students are willing to
deny that the evil of Hitler or Pol Pot is in some sense absolute.

Such con

ﬂicting impulses are clearly present when students begin a

degree course in philosophy. Oxford University

’s Examiners’ Reports

suggest that three or four years of study do little to resolve the ambiva-
lence. A recent report on the scripts for one of the ethics questions in
Philosophy Finals makes the following comment:

The name most often mentioned in candidates

’ scripts was that of Kant, who

featured in 97% of scripts. The second most often mentioned name was that of
Hitler, who appeared in 91% of scripts. Almost all of this latter group would have

been crude moral relativists if they had not suddenly thought of Hitler halfway
through their answers.

The examiners clearly felt the question demanded a rather more nuanced
response, and that too many students had given an answer which careered
from one extreme to the other

—between the competing ‘pulls’ identi-

ﬁed above. There is of course important intellectual terrain to explore in
between the extremes:a growing number of philosophers have turned their
backs on a wholesale reductionism, but have sought to accommodate the
language of

‘moral truth’ and ‘absolute’ good and evil within a more modest

an ontological framework. Such thinkers are keen to avoid any suggestion
that the universe itself has some kind of purpose behind it, whether a causally
ef

ﬁcacious Good (as in axiarchism) or the will of a benevolent Creator.

Figure 1 places a range of contemporary philosophical positions on a

continuum, depending on how they respond to the con

ﬂicting pulls.

Inevitably, such a diagram is rather rough-and-ready, as the nuances of
each position are forced into just two dimensions. At one extreme is
the error theory of John Mackie. This account is so in

ﬂuenced by the

pull of reductionism that it regards the objectivism in humans

’ pre-

philosophical moral attitudes as entirely mistaken. At the other extreme
are theistic and axiarchic positions, which seek to answer the epistemolog-
ical and ontological questions under discussion with reference to a purpose
which, they claim lies behind the universe as a whole. In the middle,
between the dotted lines, lie positions which seek to accommodate both
pulls. While they differ in many other ways, these intermediate positions
all share two properties. Firstly, they are

‘secular’ in that they deny the

universe has any intrinsic purpose: the existence and nature of the universe
are not to be explained by either an agent (as in classical theism) or the
causal power of the Good (as in axiarchism). Secondly, they all seek to
vindicate the incipient objectivism in humans

’ pre-philosophical moral

attitudes. Although some of these intermediate positions attempt to ac-
commodate these attitudes within in an ontology every bit as modest as
Mackie

’s, what distinguishes all these positions from his is that they refuse

to regard our fundamental, pre-philosophical objectivism as an

‘error’.

I owe this quote to Bob Hargrave

’s online guide to ‘How to Fail Philosophy Exams’, at

<http://users.ox.ac.uk/

ball0888/oxfordopen/Rodin.htm>

FROM MORALITY TO METAPHYSICS

2. Argument of this Book

This book offers an argument for theism. Its claim is that only purposive
accounts of the universe both do justice to our pre-philosophical moral
commitments and explain the human acquisition of belief-generating and
belief-evaluating capacities which track (an ex hypothesi independent)
moral truth. Among such purposive accounts, it claims that theism
offers the most satisfying, intelligible explanation of this phenomenon.

It is important to stress from the outset that this book will not advance a

more general case for theism. Nor will it address the other grounds on
which secular philosophers might reject religious beliefs. The sole purpose
of the book is to highlight a serious and systemic problem facing any
secular positions that attempts to accommodate our pre-philosophical
moral commitments. (In terms of Figure 1, these are the range of positions
which lie between the dotted lines.)

In varied ways, these secular accounts seek to mediate between the

con

ﬂicting pulls of objectivism and reductionism. At one extreme, Simon

Blackburn, Allan Gibbard, Christine Korsgaard, and (in his earlier writ-
ings) Tim Scanlon argues that our fundamental moral convictions can be
accommodated without objectivism. At the other, Philippa Foot, Roger
Crisp, and the later Scanlon seek to combine a fully objectivist account of
moral norms with an explanation of human knowledge of these norms
which does not involve a purposive agent or force. John McDowell

’s

position is harder to categorize, as it rejects any

‘sideways-on’ view

from which the world and our conceptual schemes could be compared.
McDowell describes his general philosophical position as

‘anti-anti-realism’.

NON-PURPOSIVE (‘SECULAR’) ACCOUNT OF UNIVERSE

PURPOSIVE ACCOUNT

ERROR

THEORY

(Mackie)

MORAL QUASI-

REALISM

(Blackburn / Gibbard)

‘ANTI-ANTI–

REALISM’

(McDowell)

OBJECTIVISM

(Crisp / later Scanlon)

THEISM

CONSTRUCTIVISM

(Korsgaard / early Scanlon)

NATURALISM

(Foot)

AXIARCHISM

(Rice / Leslie)

< < Pull of reductionism

Pull of objectivism >>

Figure 1 The continuum of moral philosophy

INTRODUCTION

For these purposes, what matters is that he regards moral truth as every bit as
‘real’ and ‘objective’ as any other kind.

This book claims that all such secular accounts are problematic. Their

individual

ﬂaws ﬂow from systematic weaknesses—weaknesses which

make it doubtful that any future secular account could fare better. These
weaknesses are twofold. In the case of the less objectivist theories (on
the left of the diagram) the concessions made to reductionism leave
themunable to do justice to humans

’ most fundamental moral convictions.

The accounts on the right, which accommodate the pull of objectivism,
face a different problem. These more objectivist theories generate an
‘explanatory gap’.

This brings us to the central contention of the book. My claim is that

all secular theories which do justice to our most fundamental moral
convictions go on to generate an insoluble

‘explanatory gap’. This ‘gap’

consists in their inability to answer the following question:

(Q)

How do human beings, developing in a physical universe which
is not itself shaped by any purposive force, come to have the
capacity to apprehend objective moral norms?

It is important to stress that (Q) concerns the explanation for humans

’

possession of truth-tracking capacities with respect to morality. It does
not concern the justi

ﬁcation we have for relying on the deliverances of our

belief-generating mechanisms. I do not deny that secular thinkers have
some reason to trust the human capacity for moral discernment. My claim
is that all otherwise satisfactory secular accounts generate an explanatory
gap with respect to human knowledge of moral truth.

3. Structure of the Argument

The book has a threefold structure. The

ﬁrst part advances the prima

facie case against those secular accounts which posit an objective moral
order

—namely, that they generate an explanatory gap. The second part

considers the resources secular accounts might have to bridge the gap, or to
avoid it by deploying a weaker conception of objectivity. It argues these
resources are insuf

ﬁcient. The third section develops an argument for

theism

—arguing, ﬁrstly that purposive accounts of the universe are the

only kind which can avoid this explanatory gap, and secondly that theism
is the most satisfying of these accounts.

FROM MORALITY TO METAPHYSICS

I will now sketch out how the argument develops, chapter by chapter.

Part I: The Explanatory Gap Emerges

The argument just outlined only has force if the

‘pull of objectivism’ needs

to be respected. The explanatory gap can very easily be avoided if there is
no objective moral order. So my

ﬁrst task is to explain why we must take

morality to be (to at least some degree) objective; only then does the
explanatory question arise.

Chapter 1 defends our pre-philosophical commitment to moral

objectivism. It is an essential prelude to the main argument, for it sets
the standard by which I will determine which secular accounts are

‘sufﬁ-

ciently

’ objective. This part of the book is not so much a defence of moral

objectivism as an outline of the considerations I take to count in its favour.
Readers seeking a comprehensive defence, rather than the sketch of one,
will need to refer to the thinkers I draw on and cite.

In claiming that humans have the capacity for apprehending an

objective moral order, I will not be making extravagant claims about the
accuracy of our existing moral beliefs. Nor will I presuppose an optimistic
picture of how the con

ﬂicting moral convictions of today’s humans will or

will not converge in the future. The argument of this book just requires
two much more modest assumptions. It assumes that in our moral
re

ﬂection, each of us seeks to approximate to a truth beyond ‘whatever

I think is true

’. I also assumes that this quest is not in vain: which is to say,

that humans have some capacity to attune their beliefs more closely to
thatmoral truth, when they honestly and carefully seek it out. The purpose
of Chapter 1 is to explain why I take these assumptions to be justi

ﬁed.

In Chapter 2, I develop the prima facie case for thinking all secular

moral theories which are

‘sufﬁciently’ objectivist will also generate an

‘explanatory gap’.

This explanatory gap is a distinctive problem for secular accounts of

moral knowledge. For the sake of this argument, I am happy to concede
that human sensitivity to facts about the physical world or norms of
theoretical reasoning can be explained without invoking any purposive
force or agent. Natural selection shows the survival value of having
accurate beliefs about key features of physical reality and the principles
of theoretical reasoning

—even if we take both of these to exist indepen-

dently of the sentiments, beliefs or social conventions of the perceiver. It is

INTRODUCTION

only in the case of evaluative beliefs that the survival value and the
accuracy of beliefs seem to come apart. It is likely that, if their belief-
generating faculties emerged from a combination of random mutations
and selective pressure, humans would have acquires moral convictions
which aided their survival and propagation. But it is not clear why those
convictions would correspond with the moral truth rather than selectively
advantageous simulacra. This is the heart of the explanatory gap: Why are
our capacities with respect to morality capable of tracking truth?

Why do we need an explanation for our truth-tracking capacities with

respect to morality? Why can we not answer the question of how
we apprehend moral truths by giving an anthropological account

—telling

the story of how parents, carers, and communities inculcate in their young
the very moral beliefs we all accept are justi

ﬁed? Showing that we have

reason to trust our belief-generating faculties and telling a story of how
each new generation is inducted into the community of reasoners does not
provide an answer to (Q). For (Q) asked why we are fortunate enough to
have such truth-tracking faculties. We must not confuse an anti-sceptical
argument which justi

ﬁes the trust we place in our faculties with one which

explains their accuracy. In the case of theoretical reasoning, we have the
kind of explanation I am demanding. For, as well as having (i) an argument
to the effect that we must trust our faculties, and (ii) an anthropological
and sociological story about how children are inducted into the commu-
nity of reasoners, we also have (iii) an explanation of why this community
has reliable faculties for reasoning. In the moral case, I argue we have (i)
and (ii) but lack (iii). This is why I claim that the case of moral knowledge
is particularly problematic for secular philosophers, and a distinctive
‘explanatory gap’ arises in this domain.

Part II: Secular Responses

In Part II, I examine some of the most prominent and powerful secular
accounts. I will argue that none of them manages avoid the following
dilemma: either the positions fail to vindicate our pre-philosophical
commitment to objectivism or they generate an explanatory gap. The
particular positions which I consider do not exhaust the conceptual
space open to the secular thinker

—so this book is not offering a ‘knock-

down

’ argument against all possible accounts. It is highlighting a serious

problem that naturalism

’s most prominent advocates have yet to solve, and

FROM MORALITY TO METAPHYSICS

identifying systemic reasons for their failure which make it unlikely
that new variants will succeed.

Chapters 3 and 4 consider secular positions that fall short of objectivism.

The promise in all these positions is that they will do justice to our
pre-philosophical moral commitments without generating the

‘explana-

tory gap

’ which befalls full-blown objectivism.

Chapter 3 discusses quasi-realism, the position developed by Simon

Blackburn and Allan Gibbard. I will argue that it fails in two key respects.
Firstly, it makes our concern to learn from the insights of others rather
mysterious; whereas for an objectivist it is clear why moral sentiments very
different from our own have epistemic relevance. Secondly, it cannot give
a plausible account of future changes in our moral sensibilities.

Chapter 4 considers Timothy Scanlon

’s account of ethics. Like quasi-

realism, Scanlon

’s early position avoids the metaphysical commitments

which generate the

‘explanatory gap’. However, I will argue, the position

is unable to do justice to the inescapable ethical commitments outlined in
the

ﬁrst chapter. Scanlon draws heavily on Christine Korsgaard for his

meta-ethical account, so the chapter will also consider of her

‘procedural

moral realism

’, arguing it shares the fundamental ﬂaws of quasi-realism.

Finally, it turns to the more objectivist position developed in Scanlon

’s

2009 John Locke Lectures. While this position avoids the problems faced
by Blackburn, Gibbard, and Korsgaard, I will argue that it does so at the
price of generating the explanatory gap.

In Chapter 5, I will consider Philippa Foot

’s naturalist realism, which

treats moral properties as features of the natural world. I will argue that her
account defends a notion of

‘goodness’ too weak for our purposes.

A notion of moral correctness is required which transcends this species-
relative property. Chapter 6 examines John McDowell

’s ‘re-enchanted’

naturalism; a position that builds on Foot

’s and seeks to answer this

objection. Once again, I argue the position falls foul of the explanatory
gap. McDowell seeks to avoid this conclusion, by arguing that the kind of
explanation I am seeking is inappropriate. This returns us to the argument
of Chapter 2: that it is legitimate to ask for an explanation, because natural
selection can account for the way many of our other beliefs come to track
the truth. Chapter 6 concludes with a brief discussion of David Wiggins

’

weaker conception of objectivity in ethics. I will argue that Wiggins

’

position offers further con

ﬁrmation of my central thesis: that the

INTRODUCTION

explanatory gap is only evaded by positions which fail to do justice to
our pre-philosophical commitments.

While Parts I and II of this book consider a wide range of secular

accounts, the list is far from comprehensive. For example, it lacks a detailed
discussion of the

‘Cornell realism’ of Richard Boyd, Nicholas Sturgeon,

and David Brink, or the moral functionalism of Frank Jackson. While
these positions are distinctive in other ways, the objections advanced
against the positions listed above will (if successful) tell against these
and other secular accounts. I will indicate brie

ﬂy in Chapters 2 and 5

where I take my arguments against quasi-realism and Foot

’s naturalism to

tell against Jackson and the Cornell realists respectively.

Offering a treatment of every secular account

—even if such a Hercule-

an labour were achievable

—would be incredibly tedious for reader

and author alike. My argument against each secular account is to press
upon them the same dilemma: either insuf

ﬁcient objectivity to vindicate

our

ﬁrst-order commitments or the generation of an explanatory gap.

To preserve the reader

’s good humour (as well as my own) I have limited

myself to advancing this dilemma against the

‘error theory’ and straight-

forward moral objectivism in Part I of the book, and against four of
the most plausible and distinctive intermediate positions in Part II.

Part III: Theism

The

ﬁnal section of this book advances a positive case for theism.

In Chapter 7, I will defend the legitimacy of teleological explanations,
and in particular explanation in terms of the intentions of an agent. I will
argue they are uniquely are capable of accounting for the way we must
take the characteristics of the objective moral order to shape the convic-
tions and capacities of human beings. In consequence (and in contrast to
their secular counterparts) they are able to explain the human capacity for
apprehending the objective moral order. In Chapter 8 I will argue against
the axiarchism of Hugh Rice and John Leslie, claiming only classical
theism and a non-axiarchic Neoplatonism offer satisfactory explanations.
At the end of the chapter, I argue that classical theism offers the most
intelligible and comprehensive such explanation.

FROM MORALITY TO METAPHYSICS

PART I

The

‘Explanatory Gap’

This page intentionally left blank

Why Take Morality to be
Objective?

1.1 Aim and Structure of the Chapter

The

ﬁrst task of this book is to explain why we should take morality to be

(at least quasi-) objective. It is only if moral truth is in some sense objective
that the

‘explanatory gap’ will arise. If moral truth is constituted or

generated by our sentiments and cultural conventions, then the question
of how our moral beliefs succeed in tracking the truth is easily answered.

Much of the pull of objectivism comes from our pre-philosophical

moral commitments. Confronted with certain phenomena, most obviously
extreme evils, we feel compelled to say their moral properties are (in some
sense) absolute. This chapter will argue that we have good reason to retain
these commitments.

The chapter has a twofold structure. Firstly, it challenges the

‘pull of

reductionism

’, addressing the most popular criticisms of our pre-

philosophical moral commitments. These criticisms concern the (allegedly
problematic) ontology and epistemology of objective moral norms. My
argument will be that even moral sceptics are committed to the view that
there are objective norms of theoretical reasoning with equally

‘problem-

atic

’ properties. Therefore, these alleged problems cannot constitute a

decisive objection.

The second part of the chapter offers a more positive argument which

builds on the same claim: that all humans are committed to the objectivity
of norms of theoretical reasoning. I will argue that we are committed to
such norms because they underwrite practices which are indispensable to
human thought and action. When we come to such practices (e.g. infer-
ence to the best explanation) we have reached

‘bedrock’. I will argue that

human beings are committed to the objectivity of moral norms for much

the same reason. Such objective norms are

‘deliberatively indispensable’—

they are essential to practices which are central to all of our lives. I will
maintain that it is impossible to engage in moral deliberation without
taking oneself to be aiming at a normative truth that goes beyond one

’s

own sentiments or the conventions of one

’s culture. Therefore, outside

the seminar room it is impossible for human beings to avoid a practical
assumption of moral objectivism. (That is the reason for giving this book
the subtitle

‘The theistic implications of our ethical commitments’.)

As I explained in the Introduction, the argument for moral objectivism

can only be advanced in outline form. Such a sketch needs to be given,
because the central claim of this book is that all secular accounts of morality
are either

‘insufﬁciently objective’ or generate an ‘explanatory gap’. This

chapter will establish standard against with these secular theories will be
assessed in the rest of the book; the standard by which I will argue theories
such as Simon Blackburn

’s and Christine Korsgaard’s are ‘insufﬁciently’

objective. The chapter is not designed to convince a reader with a worked-
out case against moral objectivism to revise his or her viewpoint. Rather, it
summarizes the arguments this author

ﬁnds convincing, and on which the

rest of his case will be built

—and refers sceptical readers to the places where

they are advanced more fully.

1.2 Arguments Against Objectivism

J. L. Mackie

’s Ethics: Inventing Right and Wrong remains the classic modern

argument against objectivism.

Mackie thinks the ontology of moral

judgements is problematic for the objectivist (in his terminology, he thinks
there is a

‘queerness’ to such judgements). For it ascribes to such judge-

ments a combination of objectivity (claiming they are there independent
of any agent

’s motivational structure) and bindingness upon the rational

will (claiming they have a

‘built in to-be-pursuedness’).

The objectivity of moral judgements raises serious epistemological

issues. It looks as if we have trustworthy ways of evaluating analytic
statements (i.e. by linguistic analysis), and it is trivially true that we have
reliable ways of evaluating those statements that are empirically testable.
What, then, of this third category into which moral judgements seem to

J. L. Mackie, Ethics: Inventing Right and Wrong (Harmondsworth: Penguin Books, 1977).

FROM MORALITY TO METAPHYSICS

fall? (In fact, some philosophers do take moral statements to be analytic;
I will outline my reasons for rejecting this position below at 2.5.4.)

Mackie thinks that all moral objectivists are ultimately committed to

some kind of intuitionism:

[T]he central thesis of intuitionism is one to which any objectivist view of values is
in the end committed: intuitionism merely makes unpalatably plain what other
forms of objectivism wrap up

. . . [H]owever complex the real process, it will

require some input of this distinctive sort, either premises or forms of argument
or both. When we ask the awkward question, how we can be aware of this
authoritative prescriptivity

. . . ‘a special form of intuition’ is a lame answer, but it

is the one to which the clear-headed objectivist is compelled to resort.

Mackie

’s epistemological objections are reinforced by the extent of moral

disagreement between agents and between cultures. At

ﬁrst blush, non-

objectivist theories look better equipped to explain the continuation of
apparently irresoluble differences in moral judgements. If there were the
same kind of objectivity to truth in ethics as in the physical sciences, one
might expect the former subject to have a similar degree of consensus to
the latter.

1.3 Responses

1.3.1 Analogy with theoretical reasoning

In the face of these objections, why not accept that morality is expressive
rather than objective? Later in the chapter, I will advance the positive case
for objectivism. At this stage I simply want to sketch a reply to Mackie

’s

negative arguments. My response will be to argue we are committed to the
existence of other norms of reasoning with the same ontological and
epistemological properties.

If there are objective, non-analytic norms in other areas of judgement,

this will remove a major motivation for denying there are objective moral
norms. For we would have shown that there are indeed true statements
which share the following pairs of features:

1) They are both descriptions of entities and are also prescriptive to

those rational agents who come to know their truth,

Mackie, Ethics, 38

–9.

WHY TAKE MORALITY TO BE OBJECTIVE

and

2) They are neither analytic nor knowable by empirical research

alone.

Even if we restrict ourselves entirely to the physical sciences, it looks as if
some norms with features (1) and (2) must exist. For it is hard to see how
we make choices between rival hypotheses without invoking such norms
of theory-choice.

I want to suggest that these norms of theoretical reasoning are (in David

Enoch

’s phrase) ‘deliberatively indispensable’.

By this I mean that while

scepticism about these norms is in principle possible, in practice all rational
agents must behave as if they are objective.

Enoch introduces the concept of deliberative indispensability with an

example. A physicist sees the vapour trail in a cloud chamber, and infers
the presence of a proton. The physicist

’s hypothesis (or, rather, the

complex theory of which it is a small part) is the best scienti

ﬁc explanation

for the phenomenon in question. The principles for determining what
counts as a good explanation can be denied without self-contradiction,
and there is nothing

‘beneath’ these principles which we can use to offer a

non-circular justi

ﬁcation. None the less, it is indispensable to our deliber-

ation that we take there to be objective principles which tell us what
conclusions to draw in these cases and thus what the best explanation
consists in. We cannot engage in any serious attempt to make sense of the
world around us without principles of Inference to the Best Explanation
(IBE). Moreover, giving up the attempt to make sense of that world is not
a realistic alternative.

When we choose between rival theories, we rely upon a number of

principles of IBE. We favour those hypotheses which explain the most,
those which are least ad hoc and (arguably) those which are simplest and
most elegant. The precise list of qualities which the best explanation
should display, and the weighting attached to each, may be a matter of
ongoing debate. None the less, if empirical evidence provides us with
grounds for believing anything beyond what is logically deducible from
observation sentences, then some such set of principles must govern our

I owe this phrase to David Enoch,

‘An Outline of an Argument for Robust Metanor-

mative Realism

’, in Russ Shafer-Landau (ed.), Oxford Studies in Metaethics, vol. II (Oxford:

Oxford University Press, 2007), 21

–50.

FROM MORALITY TO METAPHYSICS

inferences. The principles in question must be normative and descriptive,
for they tell us what we ought to accept on the basis of the evidence.

Might the norms of theoretical reasoning be analytic truths? Surely not:

disagreement about their content cannot plausibly be reduced to debates
about the meanings of words. While the principle of Inference to the Best
Explanation is deliberatively indispensable, there is a substantive debate
within the philosophy of science about the precise qualities which earn a
theory the title

‘Best Explanation’. In particular, there is disagreement as to

whether elegance and simplicity count in favour of a theory being the
‘Best Explanation’, and hence provide grounds for thinking that theory
likely to be true. Likewise, there seems to be a substantive disagreement
between philosophers as to whether there is a distinct norm of theoretical
reasoning which validates the Principle of Induction, or whether (as
Gilbert Harman argued) it is possible to account for the logic of scienti

ﬁc

discovery without such a bedrock principle.

For the purposes of this book, it does not matter which side is right in

each of these debates. All that matters is that each dispute is clearly a
substantive one, not merely a matter of linguistic analysis. There is a fact of
the matter about whether the principle of induction ought or ought not to
be relied upon, and a fact of the matter about whether elegance and
simplicity are part of what it is to be the

‘Best Explanation’.

The truth

in question is non-analytic. It appears therefore, that in our theoretical
reasoning we are committed to the existence of synthetic a priori impera-
tives (i.e. ones not reducible to the imperatives of logical inference) which
tell us what conclusions we ought to draw from the evidence generated by
empirical observation and experimentation. In consequence, it seems that
the existence of non-analytic norms of theoretical reasoning is delibera-
tively indispensable for any humans who want to explain and predict
features of the world around them.

What one should say to a sceptic who doubts the existence of any such

objective norms? Perhaps she will agree with us that the norms are indis-
pensable to the project of explaining our immediate observations, but will
then point out we have no independent reason to think the world is explicable.
How do we know our best principles of IBE help us to track the truth?
David Enoch offers a convincing ad hominem response to the sceptic:

Gilbert Harman,

‘The Inference to the Best Explanation’, The Philosophical Review 74

(1965).

WHY TAKE MORALITY TO BE OBJECTIVE

The explanatory project is intrinsically indispensable because it is one we cannot

—

and certainly ought not to

—fail to engage in, it is unavoidable for we are essentially

explanatory creatures. Of course, we can easily refrain from explaining one thing or
another, and it

’s not as if all of us have to be amateur scientists. But we cannot stop

explaining altogether, we cannot stop trying to make sense

—some sense—of what

is going on around us.

Because the practice of explanation is indispensable, and principles of IBE
are indispensable to that practice, we have to take them to be truth-
tracking.

Could the sceptic reply that, while these

‘explanations’ may be prag-

matically useful, that they are no guide to anything objective? The dif

ﬁ-

culty for such a sceptic is that, in the very act of arguing for scepticism she
will have to rely on some synthetic a priori principles. That is to say, any
reasons she has for claiming that the world is

‘unlikely’ to resemble our

theories must itself rest upon an account of what we are entitled to infer on
the basis of any given piece of evidence. If the sceptic

’s view is that ‘we

have no reason to think the world explicable

’, then unless this is a simple

dogmatic assertion (the very thing to which the sceptic claims to be averse)
evidence will need to be provided. Given that the evidence will not entail
the truth of the sceptic

’s claim (for otherwise she would be able to show his

opponents to be caught in a contradiction), the sceptic will be relying on
the very kinds of principle she is supposed to be denying. In short, she will
be caught in a practical contradiction. To assert that anything at all is a basis
for believing anything else, beyond what is provable by deductive argu-
ment, involves an appeal to principles of IBE.

1.3.2 Re

ﬂective equilibrium

The opponent of objectivism may reply that it is not enough for me to
argue that there must be objective norms of theoretical reasoning. Mackie

’s

epistemological worries about such entities need to be addressed more
directly.

I have argued elsewhere that our knowledge of synthetic a priori truths

(in both theoretical and practical reasoning) does not come from deductive
argument or experimental observation, but though a process of

‘reﬂective

equilibrium

’.

Such an

‘equilibrium’ is arrived at through a combination of

Angus Ritchie

‘Realism, Ideology, Truth: An Examination of Richard Rortys Critique

of Metaphysical Realism

’ (B. Phil. Thesis, 1996), 47–8.

FROM MORALITY TO METAPHYSICS

(i) singular judgements (which are intuitively compelling to us) and (ii) our
systematization of these judgements into general rules; rules which also
bring them into harmony with the judgements of other people. In a sense,
then, Mackie is correct to say that moral objectivism involves

‘intuition-

ism

’, in that the raw materials of the process of reﬂective equilibrium

include the singular judgements that seem most intuitively compelling.
This is not a precursor to epistemic anarchy. For this combination of
intuition and systematization is a feature of all human reasoning.

In scienti

ﬁc theory-formation, mathematics and ethics alike we con-

stantly attempt to systematize singular judgements into general rules. On
occasion it may be an open question whether to put up with the counter-
intuitive consequences of accepting a rule (be it a principle of theoretical
reasoning or a moral norm), or whether the counter-intuitive case should
be taken as a counter-example to the rule.

In a paper which predates Mackie

’s book by some years, J. R. Lucas

shows that this dialectic between singular case and general principle is
essential in a wide range of subject areas.

His most striking example is

taken from a courtroom where the judge is deciding a contestable point of
law. In such cases, judges do not merely enforce the law. They determine
it. The question then is how such determination is made. As Lucas argues:

[Good judges] do not decide the cases in accordance with some bad rule

—say that

of deciding for the party which bribes them most

—or they would be bad judges.

Nor do they show their impartiality by deciding cases by the toss of a coin in court;
or they would still be bad although now impartial judges. But they do not decide
the case according to some good rule: else the parties would have been able to see
what the decision was going to be and would have settled out of court. So good
judges decide their cases neither according to any rigid rule, good or bad, nor
randomly, that is accordingly to a no-rule. There is thus not an exhaustive
disjunction between being in accordance with some de

ﬁnite rule and being

completely unruly, between the conclusively justi

ﬁed and quite unjustiﬁed.

Lucas

’ argument suggests a way to avoid the dilemma Mackie wishes to

force upon the moral objectivist: either statable rules or an epistemological
free-for-all. In ethics, as in other

ﬁelds, rationality involves singular judge-

ments as well as the deductive application of principles. Although Lucas

The term

ﬁrst comes to Prominence in John Rawls, A Theory of Justice (Cambridge,

MA: Belknap, 1971), 48.

J. R. Lucas,

‘The Lesbian Rule’, Philosophy 30 (1955): 114, 200.

WHY TAKE MORALITY TO BE OBJECTIVE

does not use the term

‘reﬂective equilibrium’ it is just such a process that

his paper describes. In doing so, he anticipates contemporary philosophers
such as Timothy Scanlon (whose 2009 John Locke lectures argue we are
committed to

‘reﬂective equilibrium’ thinking in our theoretical as well as

practical reasoning) and Robert Audi (whose

‘moderate intuitionism’

explicitly includes this process

—and is compatible with the more general

account I am developing).

1.3.3 Commitment to objectivity of truth

The argument above assumes that if science and mathematics are episte-
mologically

‘on all fours’ with ethics, this provides a reason for taking

moral truth to be objective. One could of course draw the opposite
conclusion, and reject objectivism in all areas of knowledge. This is exactly
the line taken by Richard Rorty. Rorty claims that developments in
the philosophy of science and of language (in particular the work of
W. V. Quine and Donald Davidson) show

‘we no longer have dialectical

room to worry about

“how language hooks on to the world”’.

I have argued elsewhere that Rorty

’s position is ultimately self-refuting:

like the sceptic considered above he is caught in a practical contradiction.
For, if we do

ﬁnd Rorty’s denial of objectivism convincing, this will

inevitably be because we judge it well using our best lights of what
constitutes a good philosophical argument. In the very act of accepting
his argument, we will be taking those standards to justify our rejection of
objectivism. The issue here is stated neatly by Hilary Putnam:

If statements of the form

‘X is true (justiﬁed) relative to person P’ are themselves

true or false absolutely, then there is, after all, an absolute notion of truth (or
justi

ﬁcation), and not only of truth-for-me . . . [Otherwise, one would have to

say] that whether or not X is true relative to P is itself relative. At this point, our grasp
on what the position even means begins to wobble.

T. M. Scanlon,

‘The John Locke Lectures 2009: Being Realistic About Reasons’,

Lecture IV (online at <www.philosophy.ox.ac.uk/lectures/john_locke_lectures>); Robert
Audi,

‘Moderate Intuitionism and the Epistemology of Moral Judgment’, Ethical Theory and

Moral Practice 1(1) (1998):15

–44.

Richard Rorty, Philosophy and the Mirror of Nature (Oxford: Blackwell, 1980), 261.

Hilary Putnam, Reason, Truth and History (Cambridge: Cambridge University Press,

1981), 121. Putnam argues for a realism which falls short of the metaphysical realism set out
below in (MR1) and (MR2) and defended in Ritchie, Realism, Ideology, Truth.

FROM MORALITY TO METAPHYSICS

My discussion of Rorty led on to a defence of two theses of Metaphysical
Realism:

(MR1) The world exists, and has certain properties, independently of

either human beliefs or conceptual schemes. Indeed, the
world would exist and have properties even if no human
beings existed at all, and

(MR2) A statement is true if and only if it is an adequate representa-

tion of the way the world is, where

‘the world’ is as construed

in (MR1).

The argument of this book will assume these theses to be correct. (One
quali

ﬁcation ought to be made. In assuming the truth of (MR1) and

(MR2) I am not asserting that

‘the world’ and ‘our conceptual schemes’

can be examined separately and compared. This quali

ﬁcation will not be

enough for John McDowell, whose account of the relationship between
mind and world is

‘anti-anti-realist’ rather than metaphysically realist.

However, in Chapter 6, I will argue that this difference between his
position and mine does not materially affect the case being made in this
book. The

‘explanatory gap’ arises for his account also.)

1.3.4 Extent of moral disagreement

I now want to consider the other objection raised at the start of this
chapter

—the (alleged) fact of widespread moral disagreement.

Once again, there is a parallel with the norms of theoretical reasoning.

Widespread and even intractable disagreement does not rule out the
possibility of objective truth. I have already referred to the debates around
the principle of induction. There is ongoing disagreement as to whether
some version of the principle of induction is among the fundamental
norms of theoretical reasoning. Whatever the merits of each side of the
argument, it looks as if there must be a

‘fact of the matter,’ one way or the

other. There are clearly a wide range of philosophical issues on which
ongoing disagreement occurs without localized relativism looking like a
plausible response. What matters is not whether agreement can always be
attained but whether we have ways of proceeding in the argument.

Ritchie, Realism, Ideology, Truth, 6.

WHY TAKE MORALITY TO BE OBJECTIVE

In evaluating the title of moral enquiry to objective truth, widespread

and ongoing disagreement is not a good sign, but it is hardly a decisive
objection.

The fact that there is disagreement between intelligent and

sensitive reasoners does not imply a relativistic

‘free-for-all’. In philosoph-

ical disagreements such as the debate around induction, we can distinguish
between rigorous and plausible arguments and mere sets of assertions.
What we have in each of these debates are ways of moving forward in
the argument, but as yet no party has won a decisive victory.

In the moral case, agreement goes deeper than is often apparent. Many

of the most oft-cited examples of intractable disagreement relate to the
sanctity of human life. In our own time and place, contested issues include
the permissibility of abortion and euthanasia and the circumstances (if any)
in which wars may be just. If we look to past centuries, and indeed to
different cultures today, there is a lack of agreement on matters such as the
morality of ritual human sacri

ﬁce and the wholesale slaughter of con-

quered communities.

On closer examination, a striking feature of such disagreement, howev-

er serious and impassioned, is that it is usually contained within some
framework of agreement on the value of life. That is to say, the act of
killing another human being requires some justi

ﬁcation—perhaps in

denying that the agent in question is fully human (a key tactic in Nazi
propaganda) or claiming that the killing, while regrettable, is justi

ﬁed by a

greater good (a key part of most justi

ﬁcations of war) or claiming that God

commanded the killing (a key thought in ritual slaughter or the killings of
other tribes authorized in the Old Testament).

Communities recognize

that a justi

ﬁcation must be provided for killing other humans, and as they

come to a deeper understanding of one another

’s identity and culture they

generally

ﬁnd it harder to deny the other’s full humanity. ‘Dehumanizing’

another community usually involves empirical as well as moral falsehoods.
Untruths are circulated about the community

’s habits and allegedly evil

deeds because it is hard for humans who care about doing good to

See Enoch,

‘How is Moral Disagreement a Problem for Realism?’, Journal of Ethics 13(1)

(2009): 39ff.

See T. M. Scanlon,

‘Moral Theory: Understanding and Disagreement’, Philosophy and

Phenomenological Research 55(2) (1995): 353ff.

E.g. in 1 Samuel 15:3.

FROM MORALITY TO METAPHYSICS

recognize the other as human and none the less be willing to harm or kill
them.

In short, it seems that a signi

ﬁcant amount of moral disagreement turns

out not to be about fundamental moral principles, but about their appli-
cation in contested cases. The canvas on which these disagreements are
played out has more uniformity than we might initially appreciate. The
most egregious deviations in moral opinion come when individuals and
communities cease to be interested in a genuine search for the good.

There is a further piece of evidence in favour of this analysis of moral

disagreement: a community which wishes to dehumanize another feels the
need to weave together moral untruths with empirical ones. Communities
genuinely interested in understanding their neighbours and their possible
obligations towards them tend (as they discover more about them) to grow
in their belief in that these lives too have a dignity and worth. We need to
remember what the objectivist is committed to asserting. She need not
claim that all agents are motivated by moral considerations but that moral
considerations have a genuine claim on us that is independent of whether
we recognize it. As we shall see in 1.4.6, there is now a well-developed
literature on self-deception, both among individuals and at a communal
level. In so far as moral disagreement comes down to individuals and
communities biasing their moral beliefs in a self-serving fashion, it does
not undermine either the claim that there is an objective moral order or the
epistemological possibility of apprehending it.

1.4 Positive Case for Objectivism

1.4.1 Deliberative indispensability

So far, this chapter has been engaged in ground-clearing

—addressing the

well-worn objections to moral objectivism that fuel the

‘pull of reduc-

tionism

’. My strategy has been to show that the problems which are

alleged to af

ﬂict moral reasoning also affect other areas of knowledge

Nothing I say here is intended to take a position either way on the arguments which

rage between thinkers who believe in a

‘moral law’ evident to all reasonable agents and those

who believe in more tradition-based modes of enquiry. My own argument is clearly
compatible with the former view

—and it is also compatible with the latter, provided that

(i) there are some signi

ﬁcant moral commitments which all rationally defensible traditions

accept, and (ii) there is some possibility of rational engagement across traditions.

WHY TAKE MORALITY TO BE OBJECTIVE

where objectivism is widely accepted. The epistemological and ontologi-
cal issues which arise with respect to practical reasoning (of which moral
reasoning is a subset) arise also for theoretical reasoning.

In making a positive case for objectivism, I will continue to draw

parallels between theoretical and practical reasoning

—arguing that, just

as objective norms of theoretical reasoning are deliberatively indispensable, so
are moral norms with at least some degree of objectivity. My argument for
the deliberative indispensability of (to some degree) objective moral norms
draws upon, and modi

ﬁes, the case made by David Enoch. As we have

seen, Enoch claims that there are a number of practices in which all human
reasoners engage. In each case, while scepticism is theoretically possible,
there is no credible suggestion that humans should actually abandon these
practices. I have already explained why (following Enoch) I take the
principles of IBE to be deliberatively indispensable. Scepticism about
IBE is theoretically possible, but its practical price would be abandonment
of any human ambition to understand the world around us.

1.4.2 Deliberative indispensability and practical reasoning

Just as IBE is indispensable to our attempts to understand the world around
us, Enoch argues objectivity about the norms of practical reasoning is
indispensable to our practical deliberation. As Enoch observes, when we
deliberate on what we should do, we seem to be trying to

ﬁnd something

out, not simply expressing a preference or taste:

It is worth noting how similar the phenomenology of deliberation is to that of
trying to

ﬁnd an answer to a straightforwardly factual question: When trying to

ﬁnd an answer to a straightforwardly factual question (like what the difference is
between the average income of a lawyer and a philosopher) you try to get things
right, to come up with the answer that is

—independently of your settling on it—

the right one. When deliberating, you also try to get things right, to decide as

—

independently of how you end up deciding

—it makes most sense for you to

decide.

Of course, from the mere feel of our deliberation, relatively little follows.
We also deliberate over whether an action is

‘polite’, and can argue about

whether a joke is

‘funny’. This does not mean that there is anything more

than a locally objective standard of etiquette or humour.

We do not

Enoch,

‘How is Moral Disagreement a Problem for Realism?’, 34, 37.

Cf. Philippa Foot,

‘Morality as a System of Hypothetical Imperatives,’ repr. in her

Virtues and Vices (Oxford: Oxford University Press, 2002).

FROM MORALITY TO METAPHYSICS

suppose that other cultures get something wrong if their rules of etiquette
are different from ours. Nor do we suppose it makes sense for an agent to
say a joke is unfunny if everyone in the world laughs at it (unless, perhaps,
he means that he disapproves of everyone

ﬁnding it funny, and thinks they

ought not to laugh).

What people

ﬁnd funny is constitutive of what is funny. In conse-

quence, while we can imagine an academic research project into

‘Why

people

ﬁnd certain things funny’, a Philosophy D.Phil. on ‘What things

are truly funny

’ would itself be a rather poor joke.

Where standards are clearly culturally relative, two kinds of deliberation

are possible. Firstly, agents can deliberate about what people will take to be
polite or amusing. (Trying to

‘strike the right note’ in a speech involves

seeking to understand the communal standards of etiquette and humour.)
Secondly, agents can deliberate on the stance they will take towards these
communal standards. Perhaps our speaker thinks the standards should be
revised. She may think that contemporary attitudes to humour are unduly
conservative, or that certain forms of etiquette exist only to perpetuate
social inequality. Alternatively, her rejection may be less principled; for
example, she may simply fail to see why she should bother to be good
mannered.

What is an agent doing when she deliberates about the attitude she

should take to such communal standards? She may be trying to work out
what her all-things-considered desires are, and what stance will best ful

ﬁl

them. In a society which disapproved of divorce, an agent might deliberate
over whether to divorce her spouse without any reference to objective
norms. She would have to consider what made her contented and unhappy
within the marriage, about how those around her would react to the
instigation of divorce proceedings, and about her likely happiness at the
end of the process. We can see why this kind of deliberation has a
phenomenological af

ﬁnity with explanatory theorizing. They are both

trying to get to a fact of the matter. In the case of practical reasoning,
Enoch and I have yet to show that this

‘fact of the matter’ need involve

something independent of an agent

’s desires or social conventions. That is

the question to which I will now turn.

1.4.3 Deliberative indispensability and objective practical norms

Are there certain considerations which should bear on me, whatever I do or
do not happen to desire? If so, is one purpose of our deliberation to weigh
these considerations correctly? Enoch answers

‘yes’ to both questions:

WHY TAKE MORALITY TO BE OBJECTIVE

[W]hen you allow yourself to settle a deliberation by reference to a desire, you
commit yourself to the normative judgment that your desire made the relevant
action the one it makes most sense to perform. So even with desires at hand, you
still commit yourself to a normative truth.

Enoch is claiming the following: If after deliberation I take a course of
action because I desire that outcome, I am committed to the judgement
that the desire has made the action the one it makes most sense to perform.
I think this claim is excessive. In this section, I will explain why, before
advancing my own (less extreme) version of the argument.

To evaluate Enoch

’s claim, let us consider this example: I am deliberat-

ing over whether to meet someone I

ﬁnd rather tiresome, on the

grounds that I have promised to do so, or whether to stay at home, on
the grounds that I would prefer this. If I decide to stay at home, am I really
committed to the view that it would have made less sense to go and meet
the person in question? Intuitively, it seems more plausible to say that I can
continue to agonize and debate about what to do, even once I am clear
about which course of action is morally the best. In eventually settling
upon a morally inferior choice, it is not obvious that I am committing
myself to the view that there is any kind of wider truth of practical reason
to the effect that staying at home and doing what I want is

‘the action . . . it

makes most sense to perform

’ (as if the moral ‘ought’ were outweighed by

other considerations in a wider,

‘all-things-considered ought’).

The relationship between deliberation and action in such cases is com-

plex and contested. Enoch

’s account seems to imply a highly rationalistic

view of this

—the kind of position taken by Socrates, and defended by

R. M. Hare in the opening sections of The Language of Morals. Hare argues
that we do not simply look at what someone says in order to work out
what they believe:

If we were to ask of a person

‘What are his moral principles?’ the way in which we

could be most sure of a true answer would be by studying what he did

. . . . [it is

when] faced with choices or decisions between alternative courses of action,
between alternative answers to the question

‘What shall I do?’, that he would

reveal in what principles of conduct he really believed.

Enoch,

‘How is Moral Disagreement a Problem for Realism?’, 37–8.

R. M. Hare, The Language of Morals (Oxford: Clarendon Press, 1952), 1.

FROM MORALITY TO METAPHYSICS

One might think this account fails to capture the nuances of what goes on
in situations of akrasia (weakness of the will). It seems plausible to say that
we do sometimes deliberate over whether to act morally, without com-
mitting ourselves to the view that there is a

‘fact of the matter’ about

whether we ought, all-things-considered to act morally.

We do not need to come to a conclusion on this issue in order to move

forward in the wider discussion. As I will show below, Enoch

’s overall

conclusion (about the deliberative indispensability of normative realism)
can be defended even if we accept a claim about practical deliberation
weaker than the one he has made. All that we need claim is that, in so far as
an agent

’s action is the product of a practical deliberation, she judges the

action to make sense. If after deliberation she knowingly chooses the
morally wrong action, it must make sense in some other way (perhaps
because it generates pleasure or protects one of her fundamental interests).
It does not, however, need to make more sense than any other alternative.

The crucial question, then, is whether an agent need accept that there

are any objective moral reasons for action. If practical deliberation is
sometimes simply a matter of determining what my all-things-considered
desires are, or what social convention demands, why could it not always be
restricted to these kinds of fact? Why also postulate objective moral facts?

When an agent deliberates about the stance she will take towards a rule

of etiquette, or whether to instigate a divorce, she will not only ask
whether her actions will offend or please others. She will also ask whether
their responses are justi

ﬁed. This question of justiﬁcation is even more

central when we come to deliberate upon our own desires. It is pretty
implausible to claim that the only thing I ever do when re

ﬂecting on my

desires is to discern my own all-things-considered desire. My deliberation
also involves considering whether my desires are excessively self-centred
or shallow, and how they ought to be weighed against the desires and
needs of others. It looks as if I am trying to get something right, and trying
to revise my desires in the light of the fact of the matter about what is or is
not right.

If this line of thought is correct, it still does not commit us to Enoch

’s

more extreme thesis

—that when I decide to act on my desires, I commit

myself to a normative truth. As I have observed, agents do not only
deliberate about what is the right thing to do. They also deliberate about
whether to do the right thing. Moral decision-making seems to include
both cases where we are solely concerned to work out what is the right

WHY TAKE MORALITY TO BE OBJECTIVE

thing to do and cases where we are concerned as to whether to do the right
thing. The latter case cannot be straightforwardly factual. If I choose
to do the wrong thing I do not seem committed to claim that, had
I chosen to do the right thing, I would have made a mistake (by ignoring
an overriding reason for action).

I conclude from this that we can sometimes deliberate without thinking

we are trying to get to a fact of the matter. However, it is impossible to
make sense of much of the deliberation all re

ﬂective human beings

regularly engage in, whatever their philosophical persuasion, without
taking there to be some further standard by which they judge both their
desires and the attitudes of their wider culture. It is not clear how the non-
objectivist is to account for the prima facie independence of this standard.

On the other hand, even if we grant that there are reasons for agents to

act whose truth is independent of the their desires or opinions, it is not
obvious that any of this will lead to moral objectivism. One might accept
that there are objective norms of practical reasoning but deny that any of
these norms are, in the last analysis, moral. Moral objectivism is a stronger
claim, and one I will defend in the section which follows.

1.4.4 Objectivity of moral norms

In this book, I will be using

‘moral’ with a rather broader scope than some

philosophers. I will use the word to refer to all the reasons for action that
ﬂow from the existence of objective value or obligation.

In conse-

quence, when I claim that there are

‘objective moral norms’ all I mean is

that there is a fact of the matter about how one human being ought to
behave towards others (or how he

‘has reason’ to behave towards them),

and that the content of these obligations (reasons) is not wholly prudential.
That is to say, we have reasons for, and obligations to, care about one
another which do not

ﬂow from the beneﬁt that care will ultimately give

to us.

When

‘morality’ is deﬁned in this broad sense, it seems fairly clear that

the deliberation all re

ﬂective agents engage in makes use of moral, as well

as more generally practical, norms. Abandoning the belief that these
standards are objective would have a huge impact on our deliberative
practices. This impact is signi

ﬁcantly underplayed by Mackie in his defence

Thus, in so far as aesthetic values provide reasons for action, these reasons would also

count as

‘moral’.

FROM MORALITY TO METAPHYSICS

of an error theory. Mackie seeks to construct a

ﬁrewall between one’s

meta-ethics and both the content of

ﬁrst-order moral judgements and the

intensity with which they are held.

This seems implausible: if, in the

ﬁnal

analysis, we view our moral judgements simply as expressions of our
sentiments and preferences, or the conventions of our culture, rather
than an articulation of an absolute claim made upon us by the moral
order, this cannot but affect our

ﬁrst-order attitudes.

Ronald Dworkin makes this point with characteristic force:

If someone says that soccer is a

‘bad’ or ‘worthless’ game, for example, he may well

concede on re

ﬂection, that his distaste for soccer is entirely ‘subjective’—that he

doesn

’t regard that game as in any ‘objective’ sense less worthwhile than the game

he prefers to watch. Though he has a reason for not watching soccer, he might say,
no one whose tastes are different has the same reason.

So when I say that the badness of abortion is objective

. . . it would be natural to

understand me as explaining that I do not regard my views about abortion in the
same way

. . . The claim that abortion is objectively wrong seems equivalent, that is,

in ordinary discourse, to

. . . [the claim] that abortion would still be wrong even if

no one thought it was

. . . I mean that abortion is just plain wrong, not wrong only

because people think it is.

The cost of an error theory of morality is that it collapses these two kinds of
‘badness’ into one. In consequence, we can make no sense of that aspect of
our deliberation which relates to whether actions are justi

ﬁed. In Dwor-

kin

’s example, the person who says soccer is ‘worthless’ is merely referring

to his all-things-considered desire. For such a person, a game of soccer
would have no worth. By contrast, the person who says abortion is

‘bad’ is

making a claim about what actions there are objective reasons to perform
or desist from, and our deliberation on these matters does not simply
concern our all-things-considered desires.

Of course, none of this proves that our pre-philosophical practices are

justi

ﬁed. What it does show is that the consequences of adopting an error

theory are far deeper than Mackie has allowed. As Dworkin argues,
adopting such a theory will undermine our ability to claim anything at
all is

‘just plain wrong, not wrong because people think it is’.

Mackie, Ethics, ch. 5.

Ronald Dworkin,

‘Objectivity and Truth: You’d Better Believe It’, Philosophy and

Public Affairs 25(2) (1996): 98.

WHY TAKE MORALITY TO BE OBJECTIVE

There are of course arguments from within our normative discourse in

favour of scepticism about particular judgements. Many people are scepti-
cal about whether sexual acts are intrinsically good or bad, and see consent
and harm as the only criterion for value judgements.

Likewise, some

people object to attempts by one culture to impose its values on another.

In both cases, one set of normative judgements (

‘acts between consenting

adults are only wrong if someone is harmed

’ and ‘it is wrong for one

culture to impose its values on another

’) is used to deny the validity of

another (

‘homosexuals are intrinsically disordered’ and ‘the values of

western society ought to be imposed on other cultures

’). Local scepticism

ﬂows from speciﬁc worries about the reliability of a moral judgement, and
assumes a backdrop of generally truth-tracking moral faculties. I judge this
attitude (be it disapproval of homosexuality or be it

‘cultural imperialism’)

to be a product of an erroneous social conditioning which has distorted my
view of the moral truth. But such a judgement relies on my capacities for
discerning moral truth to be, more generally, truth-tracking. Local scepti-
cisms rely on a backdrop which is non-sceptical.

It would be a natural, and I think correct, extension of Dworkin

’s

argument to say that a view which denies there are any reasons for action
grounded in the needs of others rather than our own preferences is one we
ought not to hold. The problem with grounding of morality in non-moral
reasons

—as Mackie’s error theory must—is that it is morally, rather than

meta-ethically, wrong. There is no non-circular justi

ﬁcation at this point:

we have reached bedrock.

That one cannot argue someone into our most fundamental moral

commitments does not undermine them, any more than our commitment
to the principles of Inference to the Best Explanation would be under-
mined by us encountering a sceptic who failed to see why anything
needed to be explained at all. In the moral case, as Raymond Gaita
observes:

The fact that blackboards can be

ﬁlled with what are called sceptical arguments is

what sustains the illusion that it is a serious intellectual option. If anyone seriously
asserted them in his own name we would judge him to be wicked and we would

E.g. Scanlon,

‘Moral Theory’, 352.

The tensions within such a view are discussed in Will Kymlicka, Multicultural Odysseys:

Navigating the New International Politics of Diversity (Oxford: Oxford University Press), 6f.

FROM MORALITY TO METAPHYSICS

believe his wickedness to be the reason such arguments carried any weight with
him.

Gaita argues that certain views ought to be

‘unthinkable’—that there is

something morally wrong with needing a justi

ﬁcation for the most funda-

mental of our ethical commitments. In applying this argument to this
discussion of Mackie, I am not impugning the morality of Mackie or his
fellow reductionists. Indeed, is their very loyalty to our shared and funda-
mental moral commitments that stops them pushing their meta-ethical
arguments to their logical conclusion. Mackie very much wants to avoid
ﬁrst-order conclusions which are ‘unthinkable’ and ‘beyond consider-
ation

’. This is why a large proportion of Ethics: Inventing Right and Wrong

is devoted to an attempt to show that a reductionist meta-ethics has
positive and progressive

ﬁrst-order consequences.

For the reasons

given in this chapter, I think he fails.

The signi

ﬁcance of Gaita’s observation is that it defuses one potential line

of response to this failure. We do not need to give credence to the response
that it is our

ﬁrst-order practice (rather than the meta-ethics) which needs

to change. There is some evidence that Mackie would (at some level at
least) agree, for as we have seen, he tries very hard to downplay the negative
impact of his meta-ethics on our

ﬁrst-order practice.

In the sections which follow, I will press home this critique of the error

theory, by arguing that it cannot escape a nihilistic view of humans

’ dignity

and value, and also that such nihilism has a human cost. The error theory
deprives us of the motivation for emancipatory changes in our moral
outlook, and also of one of the principal bulwarks against oppression.
The Oxford

ﬁnalists cited in the introduction to this book were not so

remiss in recalling Hitler half-way through their examination answers.
I will argue that a more consistent mindfulness of the evils of Nazi
Germany should make us wary of the

‘illusion’ that any moral scepticism

is a

‘serious intellectual option’.

1.4.5 Nihilism of error theory

I will make the case for the inherent nihilism of the error theory by way of
a thought-example. It is borrowed from Roger Crisp

’s defence of (what I,

Raymond Gaita, Good and Evil: An Absolute Conception (London: Routledge, 2004),

316.

Mackie, Ethics, Parts II and III.

WHY TAKE MORALITY TO BE OBJECTIVE

though not he, would term) moral objectivism in Reasons and the Good.
(Crisp

’s position is objectivist about practical norms, but not about what he

terms

‘morality’. But on the wider deﬁnition of morality given above,

where all

‘reasons for action that ﬂow from the existence of objective value

or obligation

’ count as ‘moral’ reasons, Crisp is an objectivist. For he is a

hedonist who holds that pleasure and the avoidance of pain are the sole
bearers of objective value.

)

Crisp imagines two agents discussing what they have reason to choose.

As so often in philosophers

’ thought-examples, a rather limited range of

possibilities are imagined to be on offer

—in this case, listening to pleasant

music or being tortured. The dialogue runs as follows:

You: Why do I have a reason to choose the music rather than the torture?
Me: Because the music

’s enjoyable while the torture will be agonizing

You: But why do I have a reason to choose what

’s enjoyable over what’s

agonizing?

Me: Because enjoyment is good for you, and suffering is bad for you.

Here, it seems, we really have reached epistemic bedrock:

‘If the dialogue

continues with you asking the question

“Why do I have a reason to

promote what

’s good for me?” I confess that all I can say is “Well, you

just do.

” But that is not because I have not already given you the best and

ﬁnal answer available.’

To hold an agent has a reason, which is independent of sentiments

and cultural conventions, for choosing what is pleasurable over what is
agonizing is to make a substantive claim. It may seem to us to be an
utterly self-evident claim, but it is hard to see how it could be an
analytic truth. To say that something is pleasurable is not the same as
saying that it is preferable. In acknowledging that pleasure and the
avoidance of pain provide me with prima facie reasons for action, I am
making a substantive claim that can be denied without fear of self-
contradiction.

The mystery lies in why anyone (unless in the grip of a peculiarly self-

punishing set of moral or religious convictions

) would deny the obvious

Roger Crisp, Reasons and the Good (Oxford: Oxford University Press, 2006), 15f.

Crisp, Reasons and the Good, 38.

Crisp, Reasons and the Good, 50.

See, for example, Robert M. Adams

’ discussion of the views of the seventeenth-century

quietist François Fénelon in

‘Pure Love’, in his The Virtue of Faith and Other Essays in

Philosophical Theology (Oxford: Oxford University Press, 1987), 174

–92.

FROM MORALITY TO METAPHYSICS

truth of Crisp

’s ‘best and ﬁnal answer’. While it has no further non-circular

defence, it is surely a central example of the kind of belief yielded by
the process of re

ﬂective equilibrium, a process I have already shown we

have to rely on when defending principles such as induction or inference
to the best explanation. It is a process we are equally justi

ﬁed in deploying

here.

Once we have established that our own pain and pleasure provide

reasons for action, it is a fairly straightforward matter to show (by the
re

ﬂective equilibrium process) that the pain and pleasure of others also

provide such reasons. Once again, Crisp makes the case by way of some
unlikely thought-examples:

Two Doors 1.

You are confronted by two doors. If you do not pass through one

or other of them you will suffer an extremely painful electric shock. If you pass
through door A, nothing further will happen. If you pass through door B, some
other person, a stranger who is out of sight, will suffer an extremely painful electric
shock. Once you have passed through either door, you will entirely forget what has
happened.

As Crisp argues, anyone who denies they have reason to pass through door
A rather than door B is

‘in the grip of an egoistic theory to the point where

they cannot appreciate what to most of us is obvious.

’

A second example is given to show that that the pain of others is not

silenced as a reason by a diminution in our own well-being:

Two Doors 2.

You are confronted by two doors. If you do not pass through one

or other of them, you will suffer an extremely painful electric shock. If you pass
through door A, you will experience a minor twinge in your leg, but nothing
further will happen. If you pass through door B, you will not experience the
twinge, but some other person, a stranger and out of sight, will suffer an extremely
painful electric shock.

Crisp, I think rightly, concludes that all agents have reason to pass through
door A.

To claim that you have stronger reason to pass through door B than through door
A is almost as implausible as the claim that in Two Doors 1 there is nothing to choose

Crisp, Reasons and the Good, 132.

WHY TAKE MORALITY TO BE OBJECTIVE

between the doors. The well-being of others, in other words, can ground reasons
to act which override the reason we have to promote our own well-being.

This example is of course extremely contrived

—and may seem a rather

laborious way to make an obvious point. However, the obviousness of the
point has not stopped a number of philosophers from denying it. Given
the epistemological argument outlined above (in favour of the re

ﬂective

equilibrium process) we have good reason to accept Crisp

’s conclusion.

The fact that one can, without self-contradiction, deny that we have
objective reasons for the reduction of pain and the enhancement of
well-being, for ourselves and for other agents

—that is to say, reasons

which are in no way constituted by our own preferences or the conven-
tions of our culture

—does not give us any grounds for entertaining such a

denial. Outside the seminar room, no-one views these kinds of moral
conviction as simply another kind of preference. As Crisp observes:

A form of realism which begins by re

ﬂection on cases such as that of the choice

between torture and music, then moves on the basis of that re

ﬂection to the view

that there is a normative reason to promote one's own well-being, and thus answers
the normative question in the af

ﬁrmative, cannot be accused of any logical error.

All that [its critics] can say is that it is a mistake to believe

—just on the basis of

ﬂection on the case in question—that there is a reason to choose music over

torture

. . . Again we have reached bedrock, but it might be worth recording that

when I have described this case to non-philosophers and asked the normative
question, everyone turns out to be a realist

. . . The realist does not begin by

assuming that there are certain actions we have reason to do, but arrives at that
conclusion by re

ﬂection on certain cases.

A more justi

ﬁed objection to Crisp’s account would be that it asserts too

little, not that he asserts too much. Crisp claims that only pleasure and the
avoidance have objective value. For reasons I will elaborate later (at 2.5.5),
I think these hedonistic principles cannot be the sole foundation of ethics.
This

ﬂaw in Crisp’s more detailed account of ethics does nothing to detract

from the argument just considered. Crisp

’s case for objectivism in ethics

seems every bit as sound as the arguments we must accept for the objec-
tivity of principles of IBE or for the objective validity of the principle of
induction.

Crisp

’ thought-experiments highlight the nihilism inherent in the error

theory. I will now go on to argue in more detail that this nihilism has a
corrosive impact on

ﬁrst-order ethics.

FROM MORALITY TO METAPHYSICS

1.4.6 Objectivism and the

‘expanding circle’

‘Inventing right and wrong’ transforms our moral deliberation from a
search for the truth (where we may need to examine and discard our
prejudices and biases, and seek insights from others) to an elaboration of
our all-things-considered preferences. It is quite implausible to deny that
such a change in our conception of the purpose of moral deliberation
would radically change the practice. Mackie accepts it would lead to some
change, and indeed seeks to present the error theory as being more
hospitable to progressive social forces than objectivism.

One can see why, at

ﬁrst blush, this might seem plausible. There is an

obvious conservatism in my argument for objectivism, with its dismissal of
certain changes in our

ﬁrst-order moral beliefs as ‘unthinkable’ and ‘be-

yond consideration

’. Which social changes should count ‘progressive’ is of

course a moot point, but one part of what constitutes moral progress is
usually taken to be the

‘expansion of the circle’ of moral concern (an image

owed to Peter Singer)

to include those (humans and perhaps non-

humans) whose dignity, rights, and needs have previously been obscured.
In this book, I am going to assume without further defence that a concern
for the weak and vulnerable, at least within the human race, is a funda-
mental

ﬁrst-order moral commitment. This assumption will rule out

certain forms of utilitarianism which make concern for the weak and
vulnerable only contingently desirable. (For example, an egalitarian policy
might be justi

ﬁed on the grounds that the marginal increase in utility when

resources are given to the poorest is greater than when given to the richest.
However, in circumstances where the richest gained more utility per pound
spent on them than the poorest, such a utilitarian theory would commend
allowing some to live in destitution whilst others had far more than
enough.

)

There are two reasons why these

ﬁrst-order commitments favour ob-

jectivism. Firstly, objectivism provides a rationale for re-examining our moral
beliefs. That is to say, it requires us to take seriously the proposal that the
communal status quo might need revision on behalf of those whom the
current consensus excludes. This offers the possibility of an

‘expanding

Peter Singer, The Expanding Circle: Ethics and Sociobiology (New York: Farrar, Straus and

Giroux, 1981).

See Amartya Sen

’s criticisms of utilitarianism in his Development as Freedom (Oxford:

Oxford University, 1999), 67f.

WHY TAKE MORALITY TO BE OBJECTIVE

circle

’ of moral concerns as we come to recognize the claims of those who

have previously been thought less than fully human. The objectivist must
also face the question of moral claims non-humans may or may not have
upon us

—a theme to which we will return in later chapters.

How one determines the moral signi

ﬁcance and claims of different

agents is itself a complex question. But precisely because it conceives of
ethics as a matter of discovery rather than invention, objectivism permits a
conception of moral progress through an increasing openness to the lives
and insights of others.

While I have been arguing against an ontological moral scepticism,

there is another kind of scepticism which can help expand the circle of
our moral concerns. Sabina Lovibond discusses this

‘protesting’ scepticism,

arguing it can offer an inchoate protest against the restrictive moralizing of
a narrow and hypocritical society. Although it may initially strike us as
amoral, it has the potential to expand our moral horizons:

Phenomena as diverse as the Dada and surrealist styles in art, the Baader-Meinhof
style in politics, and assorted

‘repulsive’ styles in popular culture may all be

understood as ironic responses to what is perceived as too cohesive moral, political
or aesthetic rationality. The exponents of these styles may be thought of as taking
up the challenge which society, in their view, presents in due course to each of its
members

—a challenge relating to the consensual scheme of values, which is felt to

be on offer only as a package: either you buy it, or you accept the status of a moral
and intellectual outlaw

. . . to opt deliberately for outlaw status is merely to answer

the question in the terms in which it was posed.

Ironically, it is precisely because the objectivist takes there to be a truth
beyond his or his culture

’s existing horizons that he can make sense of the

kind of protest Lovibond describes. Its real signi

ﬁcance lies not (as it

initially appears to) in the denial of all moral truth but in its rejection of
an excessively constricting consensus.

This is an issue I have explored in more detail elsewhere. Drawing on

work on

‘false consciousness’ by Denise Myerson and Jon Elster,

I have

identi

ﬁed three ways in which a moral objectivist could and should accept

Sabina Lovibond, Realism and Imagination in Ethics (Oxford: Blackwell, 1983), 187.

Ritchie, Realism, Ideology, Truth, ch. 5; Denise Meyerson, False Consciousness (Oxford:

Clarendon Press, 1991); Jon Elster, Sour Grapes (Cambridge: Cambridge University Press,
1983) and Making Sense of Marx (Cambridge: Cambridge University Press, 1985).

FROM MORALITY TO METAPHYSICS

that vested interests may systematically bias the ordinary processes of belief
formation:

1. Theory Generation. The processes by which new theories are gener-

ated can be biased. There may be systematic disincentives (in terms
of popular opprobrium and reduced opportunities for personal pre-
ferment) to develop accounts which challenge vested interests;

2. Theory Propagation. Once a theory has been formulated, the parts

which will be emphasized and elaborated on, whether in academia
or in the wider community, may also be affected by vested inter-
ests;

and

3. Theory Evaluation. Because the evaluation of theories is not by

deductively valid argument, and relies instead on the process of
re

ﬂective equilibrium, there is a possibility of a systemically skewed

evaluation of the strengths and weaknesses of rival theories, depend-
ing on whose interests they serve.

A number of philosophers, particularly in the Continental tradition (the
most prominent being Michel Foucault),

have taken such possibilities of

systemic bias to undermine the claim that any of our discourse is directed
towards truth. However, the very notion of

‘bias’ requires us to take there

to be a

‘truth’ that is something more than the outcome of the play of rival

interests. As I have argued above, it is the image of morality as a matter of
discovery and not invention that provides the motor for challenging and
overcoming the biases in our thinking which may oppress and exclude. It
is because moral beliefs seek to articulate an absolute demand upon us that
we have to consider how our ethical horizons might need to be expanded.

This second way in which objectivism underwrites our

ﬁrst-order

commitments is that it guards against a contraction in the circle of our moral
concerns. Here we must return to the argument of the Oxford Finalists, for
the evils of Nazi Germany provide the most dramatic example of the ways

Jon Elster cites the example of Robert Nozick

’s theory of distributive justice, where his

advocacy of a

‘nightwatchman state’ has been seized on and his insistence on the need for a

principle of recti

ﬁcation of past injustices (with potentially radical consequences in the USA

for the indigenous and the African American communities) has been largely ignored. Elster,
Making Sense of Marx, 475.

Michel Foucault, Power/Knowledge: Selected Interviews and Other Writings 1972

–1977, ed.

and trans. Colin Gordon (London: Longman, 1980).

WHY TAKE MORALITY TO BE OBJECTIVE

in which a society may redraw the circle in a way that abandons a whole
section of its membership.

When asked to explain her own motivation for moral objectivism,

Philippa Foot drew on the experience of the Nazi terror:

It was signi

ﬁcant that when I came back to Oxford in 1945, that was the time when

the news of the concentration camps was coming out. This news was shattering in a
fashion that no one now can easily understand

. . . . [I]n the face of the news of the

concentration camps, I thought

‘it just can’t be the way Stevenson, Ayer, and Hare

say it is, that morality in the end is just the expression of an attitude.

’ . . .

[A]ccording to these theories, there is a gap between the facts, or grounds, for a
moral judgement and that judgement itself. For whatever reasons might be given
for a moral judgement, people might without error refuse to assent to it, not
ﬁnding the relevant feelings or attitudes in themselves. And this is what I thought
was wrong. For, fundamentally, there is no way, if one takes this line, that one
could imagine oneself saying to a Nazi,

‘but we are right, and you are wrong’ with

there being any substance to the statement.

The motivation here, as with the Oxford

ﬁnalists, is retrospective: objec-

tivism makes room for the absolute evil which must be ascribed to the
Holocaust. An even more signi

ﬁcant motivation for objectivism is

provided by the impact meta-ethics may have had on

ﬁrst-order ethics

in German society in the 1930s.

Heinrich Rommen was a legal scholar who

ﬂed Nazi Germany in 1938.

Writing in the midst of the Nazi terror in 1936, he saw a connection
between the dominant meta-ethical position and

ﬁrst-order moral judge-

ments. It was in what he termed

‘positivism’—that is, the view that in law

and ethics right and wrong are stipulated and not discovered

—that Rom-

men located the regime

’s source of legitimation:

Modern totalitarianism is

. . . the ﬁnal outcome of positivism as a general philoso-

phy, as an intellectual atmosphere, as a scienti

ﬁc method raised to the level of the

absolute and divine. The position that law is will has come to mean that the human
will is freed from all universal ideas, from any objective moral order beyond class
interests, beyond nationalist or racial programs, beyond economic considerations,
beyond unlimited evolutionary progress.

Alex Voorhoeve,

‘The Grammar of Goodness: An Interview with Philippa Foot’,

Harvard Review of Philosophy 11 (2003): 33

–4.

Heinrich Rommen, Die ewige Wiederkehr des Naturrechts (Leipzig: J. Hegner, 1936); in

English as The Natural Law: A Study in Legal and Social History and Philosophy, trans. Thomas
R. Hanley (Indianapolis: Liberty Fund, 1998), 134

–5.

FROM MORALITY TO METAPHYSICS

According to Rommen, it was in the notion of an objective moral order
that the resources for resisting Nazism were to be found. Positivism had
dismantled the bulwarks against the Nazi terror, and an emancipatory
critique of the Nazi worldview required the terra

ﬁrma of moral truth:

[R]esistance to totalitarianism, in which the end results of positivism appear as
ethical and intellectual nihilism, had to look for support beyond any mere national
tradition or status quo ante and base itself on something superior to history, race,
class, scienti

ﬁc method, and the like.

When

. . . the totalitarian revolutions had succeeded by formally legal methods,

whence could a positivist, whether judge or jurist, derive a critical norm that would
enable him to pass judgment on the legitimacy of the legally correct totalitarian
revolution? Or how could a positivist determine the intrinsic injustice of a formally
legal act of the now totalitarian government?

. . . Any criticism of, or resistance to,

totalitarianism had consequently to

ﬁnd a deeper juridical basis of criticism or

resistance than the mere actual will of the state formulated with legal correctness
and enforced with an irresistible power.

It is for the historian rather than the philosopher to evaluate the causal
claims which Rommen is making. However, the fact that the connections
which this chapter has drawn out (between ontological scepticism in
meta-ethics and the corrosion of our

ﬁrst-order moral commitments)

were also being drawn in the midst of the Nazi terror should give us
pause. I hope it is now clear why I take moral objectivism to provide the
raison d

’être for our most fundamental ﬁrst-order moral commitments, and

why I think any position which undermines the former is in grave danger
of undermining the latter. As I have indicated, my intention is not to
impugn the goodness of those who deny the objectivity of the claims of
morality. My claim is rather that the very

ﬁrst-order commitments these

philosophers share with the wider community make them reluctant to
pursue their meta-ethical positions to their logical and practical conclu-
sions.

1.4.7 Varieties of scepticism

The line of argument I have advanced in this chapter is relentlessly
practical

—which raises a very obvious objection: the fact that we have

to assume something to be true does not make it true. My response is to
accept this criticism wholeheartedly, and simply to point out that there are
a whole range of things we have to assume to be true in order to get any
rational discussion going at all

—and therefore a range of forms of

WHY TAKE MORALITY TO BE OBJECTIVE

scepticism which may be impossible to refute, but are practically uninhab-
itable as systems of belief. In this sense, scepticism about moral norms or
about Inference to the Best Explanation is

‘on all fours’ with the worry

that a malign genie might be manipulating our deductive inferences. One
cannot defend the reliability of our process of inference without relying
upon it. Far from undermining the case for relying on our capacity to
make valid deductive inferences, it

’s very indispensability is the best

argument we have for trusting that capacity.

At this point, it is worth mentioning a very different form of scepticism:

the Pyrrhonian variety. Pyrrhonism maintains that the irresoluble nature of
disputes in both theoretical and practical reasoning provides ground for the
suspension of judgement. We simply are not warranted in drawing de

ﬁ-

nite conclusions on the basis of the evidence set before us, and hence
should live in a way that commits us as little as possible to one answer or
another.

The arguments given above do not immediately tell against Pyrrho-

nism. However, is vulnerable to a closely related attack. The fundamental
objection to the Pyrrhonian position is that it under-estimates the practical
nature of the questions posed in ethics. As we have just seen, the Nazi
regime posed such practical questions to those who lived under it, and did
so in a particularly stark form. Our answer to such practical questions is
given by what we do

—not merely by the propositions we accept, and the

degree of intellectual assent we are willing to afford them.

Taken as an answer to the question

‘How should I behave?’ the Pyr-

rhonian position is far from neutral. In this sense, it does not withhold
judgement. Instead, it commends a very speci

ﬁc course of action: the

cultivation of mental imperturbability (ataraxia). In evaluating Pyrrhonian
scepticism, we must not just ask whether it offers a consistent, non-
refutable position. It clearly does. We must also ask whether it is remotely
plausible that the cultivation of ataraxia is the right response to the world in
which we live. The real

ﬂaw in Pyrrhonian scepticism is that it masquer-

ades as neutrality (

‘As moral arguments are irresoluble, let us withhold

judgement

’) when in fact it commends a practical set of actions (‘Cultivate

ataraxia

’). It is against this unashamedly practical standard that it should be

measured. As a response to the world around us today, let alone that of
Nazi Germany, it is surely found wanting.

FROM MORALITY TO METAPHYSICS

1.5 Conclusion

The aim of this chapter has been to clarify the assumptions on which the
rest of the book will rest. I have outlined the reasons why I think any
acceptable meta-ethical theory must vindicate our most fundamental pre-
philosophical commitments. I have further argued that the only theories
which meet this standard are to some degree objectivist: that is to say, they
take morality to amount to something more than a human invention
which will vary with changing sentiments and cultural conventions.

This is the standard I will use in Part II of the book to evaluate the

various secular theories of ethics. Those theories attempt to accommodate
this requirement of objectivism in different ways. For example, while
Simon Blackburn

’s quasi-realism regards morality as entirely constructed

out of human sentiments, it seeks to accommodate the requirement of
objectivism by denying that moral truth will vary as those sentiments
change.

The central contention of this book is that all secular theories which

successfully accommodate this requirement of objectivism go on to gen-
erate an

‘explanatory gap’. That is to say, such theories cannot explain how

humans come to acquire belief-generating and belief-evaluating capacities
which track this (to some extent) objective moral truth. The next chapter
makes the prima facie case for thinking that, unless the universe is taken to
have some purpose behind it, objectivist theories of ethics will generate
such an explanatory gap.

WHY TAKE MORALITY TO BE OBJECTIVE

The Gap Opens

Evolution and our Capacity for Moral
Knowledge

2.1 The Aim of the chapter

This chapter presents the central argument of the book: that secular moral
objectivism generates an

‘explanatory gap’ with respect to human knowl-

edge of ethics. It will make the case with reference to the strong form of
moral objectivism expressed in (MO1) and (MO2) below. These two theses
of moral objectivism deliberately parallel (MR1) and (MR2), the two theses
of metaphysical realism which were cited in the previous chapter:

(MO1)

The moral order has certain properties, independently of
either human beliefs or conceptual schemes. Indeed, the
moral order would exist and have properties even if no
human beings existed at all;

and

(MO2)

A moral judgement is true if and only if it is an adequate
representation of the way the moral order is, where

‘the

world

’ is as construed in (MO1).

The chapter only seeks to establish a prima facie case against secular posi-
tions. Many of the philosophers whom I will discuss in Part II would not
accept anything like as strong an objectivism as (MO1) and (MO2). My
claim will be that these positions fall onto the other horn of my dilemma:
they escape the explanatory gap at the price of failing to vindicate our pre-
philosophical moral commitments.

Of course, moral objectivism is compatible with the thought that all goods and obliga-

tions pertain to human beings. In their absence, the

‘objective moral order’ might simply

consist of statements of the form

‘if human beings existed, they would be obligated to {x, y,

z}, or that, their

ﬂourishing would consist in {a, b, c}.’

2.1.1 Initial assumptions

Before going any further, I should add one further detail to the position under
consideration. It would be possible to accept these two theses of moral
objectivism whilst holding that all of our

ﬁrst-order moral beliefs are hopelessly

misguided.

This combination of views does not generate an explanatory gap.

The gap only arises when the following commitments are combined:

(1) Robust moral objectivism [that is, (MO1) and (MO2)];
(2) Secularism [that is, the rejection of a purposive account of the

universe, of the kinds offered by theism and axiarchism], and

(3) The belief that humans, through the exercise of their normal belief-

generating and belief-evaluating capacities, are able to apprehend
the objective moral order.

The arguments advanced in the previous chapter count in favour of (3) as
well as (1). The case made for objectivism was based on the deliberative
indispensability of our fundamental moral commitments. This involves
not only believing there are moral truths, but that human beings have a
capacity for discerning these truths.

Assenting to (3) does not commit us to the view that the contemporary

moral consensus is near the truth in all respects. Nor need we take a
‘Whiggish’ view of the history of ethics in which all changes in moral
beliefs constitute improvements. All (3) commits us to is the view that
humans have the capacity to apprehend the moral truth. Given that we do
not all currently agree on what that truth is, we must believe that our
epistemic capacities provide room for improvement over time (e.g.
through rational argument or through the re

ﬁnement of our sensibilities).

This is of course is compatible with the belief that, at certain points in
history (including perhaps our own) humans have become more morally
degenerate and less interested in getting to the truth. That is a matter for
historians and sociologists, not meta-ethicists. What we must assume,
given the arguments of the previous chapter, is that when humans are
genuinely engaged in the demanding enterprise of coming to more honest
and sensitive ethical beliefs, they have the capacity to acquire beliefs which

Sharon Street

’s argument for moral antirealism falls foul of this objection. See her ‘A

Darwinian Dilemma for Realist Theories of Value

’ in Philosophical Studies 127: 1 ( January

2006), 109

–166.

THE GAP OPENS

to some extent re

ﬂect the objective moral order. In this chapter, I argue

secular philosophers cannot explain why humans have such a capacity.
Whilst they can explain humans

’ acquisition of moral sensibilities and

practices of reasoning, this does not tell us why those practices and
sensibilities have the property of tracking the truth.

2.1.2 Structure of the argument

My argument will be developed in three stages:

- Firstly, I will outline some basic assumptions I am making about the

nature of explanation, and what constitutes a good explanation of a
phenomenon;

- Secondly, I will argue that the verisimilitude of our moral capacities is

something that requires an explanation; and

- Thirdly, I will consider the obvious naturalist candidate for such an

explanation of our moral capacities, namely evolutionary biology.
I will argue that it is in principle incapable of accounting for the
capacity of human thought to track the objective moral truth.

2.1.3 Assumption of explanatory realism

The argument of this chapter assumes the truth of explanatory realism.
That is to say, it assumes there is a fact of the matter as to whether

(4)

The obtaining of state of affairs Y would be a good explanation for
phenomenon X

is true, and also, quite separately, whether

(5)

The obtaining of state of affairs Y is the true explanation for X.

This allows for the possibility, which we will consider in the

ﬁnal chapter,

that a state of affairs, Y might be a good explanation

—indeed, perhaps the

only good explanation

—for a phenomenon X, but that there might be

other reasons for believing Y to be false. In such a case, one might accept
(4) while rejecting (5), believing instead that X had occurred by chance.

I take explanatory realism to be deliberatively indispensable. If the

argument outlined in the previous chapter is correct, we rely upon
norms of theoretical reasoning (such as the principles of Inference to the
Best Explanation and the Principle of Induction) because they too are
deliberatively indispensable. Reliance on principles of Inference to the

FROM MORALITY TO METAPHYSICS

Best Explanation requires us to assume there are facts about whether Y
would be a good explanation for X, and this gives us grounds for the
(logically distinct) belief that Y is therefore likely to be the reason for X.

If propositions (1) to (3) are true, the following phenomenon requires

explanation. Why is it that human beings have the epistemic capacity to
apprehend a moral order which would exist and have the properties it does
whatever they believed? If we assume explanatory realism, there will be
facts of the matter as to (i) what would constitute a good explanation of
this state of affairs and, separately, (ii) whether any such explanation is the
reason for the state of affairs. (As we shall see, it will be important to bear in
mind the possibility that there is no reason for the state of affairs.)

2.2 What Requires Explanation?

Before I consider any speci

ﬁc candidates for being the explanation, I need

to clarify exactly what needs to be explained. It is important to distinguish
the explanatory demand I am making from two other demands with
which, in my experience, it is often confused. For any cognitive capacities
of human beings

—such as our perceptual capacities {a, b, c . . . }, capacities

for theoretical reasoning {g, h, i

. . . }, and capacities for practical reasoning

{m, n, o

. . . }—we can ask three different questions about their genesis and

justi

ﬁcation. These are:

(D1)

What is the justi

ﬁcation for our faith in their reliability?

(D2)

What is the historical explanation of their development?

and

(D3)

What is the explanation for their capacity for tracking truth?

In the moral case, (D1) has already been answered. Chapter 1 explained
why I take us to be justi

ﬁed in assuming moral objectivism, and in accord-

ing a prima facie trust to our

ﬁrst-order commitments. The responses given

to Mackie and the argument which was made from deliberative indispens-
ability assume neither theism nor axiarchism. The claim I defended in that
chapter was that whatever our beliefs about any purpose which the universe
might or might not have, we have reason to be moral objectivists. Further-
more, we have reason to take the outcome of a

‘reﬂective equilibrium’

process in ethics to yield beliefs track the truth. More detailed

THE GAP OPENS

epistemological questions, concerning the particular roles of ratiocination,
intuition, and sensibility within the process of justifying moral beliefs,
constitute a separate issue within moral theory. This issue is itself best
addressed through the re

ﬂective equilibrium process—but the broad ac-

count I am defending is compatible with approaches which places the main
emphasis on intuitions (e.g. Audi

’s) or on ratiocination (e.g. Scanlon’s

current position).

It is precisely because I take the fundamental convictions which emerge

from the re

ﬂective equilibrium process to be justiﬁed—that is, I take them

to have a non-accidental correlation with objective moral norms

—that

(D3) arises. Consequently, it is important that we do not confuse the
demand for an explanation for the reliability of our moral beliefs with
the demand for a justi

ﬁcation of our trust in the human capacity to acquire

and modify our moral beliefs in a way that tracks truth. The explanation is
only needed because the justi

ﬁcation is at hand.

With respect to (D2), there will, of course, be an anthropological and

sociological story to be told, of how the particular practices of moral
reasoning have emerged historically, and the role of ethical education in
inculcating these practices into successive generations of humans. The
explanation I am demanding is not an explanation of the speci

ﬁc moral

judgements and principles we have come to hold, but rather of the fact
that these beliefs track the truth.

Granted that we are entitled to assume (in the absence of speci

ﬁc reasons

for scepticism) the reliability of our cognitive capacities (i.e. their posses-
sion of the property T of being truth-tracking), and that we have historical
accounts of how perceptual capacities {a, b, c

. . . }, capacities for theoreti-

cal reasoning {g, h, i

. . . }, and capacities for practical reasoning {m, n,

. . . } have emerged historically, we still have no explanation of why they

have the property T.

In his essay on

‘Objectivity and Truth’ Dworkin challenges this explan-

atory demand. He offers two reasons to reject (D3)

—ﬁrstly, that we do not

need such an explanation to ground our moral objectivism, and secondly,
that there is no credible explanation of the kind being requested. As I have
argued above, the

ﬁrst of these claims is correct. Because the explanation

I am seeking is distinct from a justi

ﬁcation, it follows that we might be

See footnote 8 below.

FROM MORALITY TO METAPHYSICS

justi

ﬁed in our moral objectivism even if it was absent. None the less, I will

argue, the absence of an explanation is problematic.

We can see this more clearly by means of a thought-example which

Dworkin provides. He invites us to consider the

‘bleak’ causal thesis that our

moral beliefs might be caused by factors which do not generally track the truth.
This is a situation in which we have both a justi

ﬁcation for our moral beliefs

(that is to say, (D1) is met) and a historical explanation of the development of
the processes by which we form moral beliefs (so (D2) is also met), and yet we
lack an explanation for the truth-tracking nature of those faculties. He goes on
to give a speci

ﬁc example. People who are not well off often believe

(3)

Justice requires a redistributive taxation system.

Let us suppose we can explain all human moral convictions by the
following law:

(4)

Humans have whatever moral convictions will maximize their
material wealth

Once someone who believes in redistribution realises that (4) obtains,
should this undermine his belief in (3)? Dworkin thinks not, arguing as
follows:

[W]hy shouldn

’t you count it as a piece of luck—a special example of what

Bernard Williams has called moral luck

—that your self-interest and justice here

coincide? You realize that you would have had a very different view if your own
self-interest had favoured that different view instead (just as you realize now,
independently of the bleak thesis, that you would have had different views if you
had been born into a very different culture). But that fact alone cannot undercut
your conviction about justice, and therefore about your moral luck. You have to
think something

—either that you have an obligation to vote for higher taxes or

that you do not

—and you have no reason, just in what the bleak thesis declares, to

turn your back on all the reasons . . . that inclined you to think that you do.

This is correct, in that the justi

ﬁcations for believing (3) lie in the moral

arguments one can adduce in its favour, and (4) does not rule out the
possibility that the objective reasons stack up in favour of (3). In the case of
indispensable practices (such as IBE and moral re

ﬂection), I have argued

above that only localized scepticism can ever be justi

ﬁed. We only need to

Ronald Dworkin,

‘Objectivity and Truth: You’d Better Believe It’, Philosophy and Public

Affairs 25(2) (1996): 125.

THE GAP OPENS

take the sceptic seriously if he can point to an alternative position that
would be more justi

ﬁed. In the moral case, then, the sceptic may call for a

revision of one area of our moral beliefs if he has evidence that our belief-
generating faculties are biased in a particular direction. But the general
thought that they are unreliable does not constitute a proposal for a better
moral view. It is a counsel of despair, and should be rejected in favour of
continuing to do the best we can with whatever epistemic resources are
available to us.

While I agree with Dworkin that, if the

‘bleak’ thesis were true we

would still have no choice but to place some trust our moral faculties,
I think that in doing so, we would be forced to postulate a large-scale and
quite inexplicable correlation between our faculties and the moral order.
Dworkin underestimates quite how astonishing the correlation would be.
He argues that, if (4) were true

[y]ou would be in much the same state of mind as the theologians of predestination
. . . who supposed people saw the truth only because they had been elected by God
. . . Of course you would be disappointed to learn, from God or later history, that
you had been led astray by self-interest. But not that you had been led in the right
direction.

I think this analogy fails. On the predestinarian

’s account, there is a reason

why the deliberations of the saved are reliable, even if it is not accessible to
the damned. By contrast, if all our reasoning on moral issues were disguised
self-interest, and if such self-interest were not a reliable tracker of truth,
then every day would present us with a new and inexplicable correlation
between a process which only hit the truth at random and the objective
moral order. It would be as if we had won a new epistemic lottery every
time we bent our minds and sensibilities around another moral problem.

As I have explained, I think Dworkin is right when he observes that

‘No

matter what we learn about the physical or mental world, it must remain an
open question, and one that calls for a moral rather than any other kind of
judgment, how we ought to respond.

’

Whatever the situation with respect

to explanation, we have to make moral judgements by the best lights we
have. However, if the

‘bleak thesis’ is true, they are very dim lights indeed.

This will have an impact on our

ﬁrst-order moral judgements. Contra

Dworkin,

‘Objectivity and Truth’, 126.

Dworkin,

‘Objectivity and Truth’, 128.

FROM MORALITY TO METAPHYSICS

Pyrrhonism, it cannot lead us to suspend judgement across the board. None
the less, it might lead us to hold (at least some of) our convictions rather
more tentatively.

Of course, there is little reason for us to adopt a theory of moral belief-

formation as cynical as (4). But, even without (4), the position looks rather
bleak for the secular moral objectivist. Unless we have the positive answer
to (D3) which I am demanding, we still have to view the reliability of our
moral cognitive faculties as a matter of immense and inexplicable good
fortune.

2.3 What is Involved in an Explanation?

What might a positive explanation look like? Dworkin only considers two
kinds of answer. One (his preferred response) simply leads us back through
the reasons for our moral beliefs. I show you why I take a certain moral
judgement to be true by leading you through the factors that warrant that
judgement. Either you agree that these factors could plausibly warrant
the judgement (in which case you are no longer surprised that I make the
judgement) or you fail to see why these factors point to this judgement (in
which case you may want a different kind of explanation

—wishful thinking,

class interest or so on

—for my judgement). It should be clear from the

arguments offered above why this kind of justi

ﬁcatory answer is not enough.

The other kind of answer he considers is the

‘moral-ﬁeld thesis’, where

moral truths have a direct causal impact on our beliefs. This, he says,
implies

that the universe houses, among its numerous particles of energy and matter, some
special particles

—morons—whose energy and momentum establish ﬁelds that at

once constitute the morality or immorality, virtue or vice, of particular human acts
and institutions and also interact in some way with human nervous systems so as to
make people aware of the morality or immorality of the virtue or vice.

He claims that this latter thesis is not only an absurd piece of physics, but is
also ethically objectionable. It implies that the wrongness of torture con-
sists in the

‘morons’ surrounding it, rather than the characteristics which

we would normally cite when asked why it is a wicked practice. Further,

Dworkin,

‘Objectivity and Truth’, 104.

THE GAP OPENS

he claims, leaves open the logical possibility that

‘morons’ might have

clustered around different sets of practices.

In his eagerness to parody alternative views, Dworkin seems to miss a

crucial distinction. No one in the debate is suggesting a simplistic physical
ontology of moral imperatives. It is clear that

‘morons’ could at best be

constant correlatives of morality and immorality: it would always be an open
question in a new case whether the presence of such particles continued to
correlate with moral properties. The

‘moral-ﬁeld thesis’ could at best be a

theory about belief-formation. The claim would be that morally right acts
happened to have certain physical correlates which then led us to believe
them to be morally right. In a less vulgar version, such a thesis actually looks
quite plausible. Perhaps acts which are morally right are also survival-
enhancing

—not by deﬁnition, but in a systematic and non-coincidental

manner. Here we can see a path opening up for a credible explanation of the
accuracy of our cognitive capacities with respect to ethics, for it is one which
will explain both why a belief-generating mechanism emerges in human
beings and why the mechanism is capable of tracking the truth.

2.4 Teleological Explanation

This type of explanation is rather different from the law-like form which
Dworkin parodies. Physics and chemistry are dominated by nomothetic
explanation

—a form which seeks to subsume information about the be-

haviour of particular objects into generalized laws. While later writers have
proposed a number of modi

ﬁcations and reﬁnements, the classic theory of

nomothetic explanation is given by Carl Hempel.

On this account, we

explain a phenomenon by showing that its occurrence is a logical conse-
quence of a set of starting conditions and a

‘covering law’.

It is hard to see how this form of explanation could account for the

ability of humans to cognize the moral order. We have no reason to
suppose that there will be some

‘moron’-like physical property underlying

all virtuous actions and choices which both explains the genesis of and is
detected by humans

’ moral faculties.

Carl Hempel,

‘Explanation in Science and in History’, in David-Hillel Ruben (ed.),

Explanation (Oxford: Oxford University Press, 1993), 17

–41.

FROM MORALITY TO METAPHYSICS

More promising for our purposes is the teleological form of explanation.

Charles Taylor describes it as follows:

To offer a teleological explanation of some event or class of events, e.g. the behaviour
of some being, is, then, to account for it by laws in terms of which an event

’s

occurring is held to be dependent on that event

’s being required for some end.

A teleological explanation will therefore explain a particular event or state
of affairs a by showing that it is either (i) part of the end-state Y which a
system X brings about or (ii) part of the means by which system X brings
about end-state Y. To be a complete account, the teleological explanation
will also have made it intelligible why the system X yields outcome Y. To
show that Y was

‘required for some end’ does not explain why Y happened

unless we have also given an account of the means by which Xs are capable
of generating Ys, and the reason Xs tend to generate Ys.

2.4.1 Agent explanation

Until the nineteenth century, the most promising answer to this further
question was thought to be the intentional action of an agent. Intentional
action met the above requirements, by giving explanations with the
following form:

(6)

Agent X wanted to bring about state of affairs Y,

(7)

Agent X knew that the best means to bring about Y involved
actions {a, b, c

. . . },

(8)

Agent X had the power to take actions {a, b, c

. . . },

therefore

(9)

Agent X took actions {a, b, c

. . . },

which explains why

(10)

Y occurred.

This explanation is as yet incomplete, for as yet we lack an account of
why agent X wanted to bring Y about. For reasons I will give later

Charles Taylor, The Explanation of Behaviour (London: Routledge & Kegan Paul, 1964), 9.

That is not to say the explanation offered below is the only kind of intentional

explanation. There will be other forms of intentional explanation where the agent

’s beliefs

are incorrect. This ideal form covers cases where the agent

’s beliefs are correct.

THE GAP OPENS

(in Chapter 7) I take the following to provide a satisfactory answer to this
question:

(11)

Agent X knew state of affairs Y would be objectively valuable,

therefore

(12)

Agent X had good reason to bring about Y,

which explains why

(6)

Agent X wanted to bring about Y.

Explanations (2) to (12) would therefore be one example of how inten-
tional explanation can provide a complete teleological explanation of an
event or state of affairs, and so would answer (D3).

In the case of our faculties of perception, theism provides an explanation

of exactly this form. When William Paley preached that

Of the bene

ﬁcial faculties, the contrivance is often evident. Ask after our eyesight,

the anatomist will show you the structure of the eye, its coats, humours, nerves,
and muscles, all fabricated and put together for the purpose of vision, as plainly as a
telescope or microscope for assisting it, and in the very same way

he was putting forward an account of the development in human beings of
a reliable optical system in terms of the motivation of a benevolent deity to
create rational agents capable of perceiving and interacting with their
physical surroundings.

2.4.2 Natural selection

One of the great attractions of evolutionary biology is its promise of
explanations which are as satisfying as Paley

’s, but which do not invoke

a purposive agent. Given that biological mutations are purely random, a
species

’ adaptation to its environment is explained not by any agent’s

intentional planning but by the fact that it is those genetic mutations
that aid its ability to survive and multiply which will themselves be most
extensively replicated.

William Paley, in Matthew Eddy and David M. Night (eds), Natural Theology: Or,

Evidence of the Existence and Attributes of the Deity, Collected from the Appearances of Nature,
Oxford World

’s Classics edn (Oxford: Oxford University Press, 2006), 237.

FROM MORALITY TO METAPHYSICS

Natural selection is only quasi-teleological. It does not suggest that the

telos which is achieved is brought about because that telos has any causal
power (as in axiarchism) or through the actions of a divine agent. Rather,
natural selection offers an account of the following form:

(13)

If we assume the genes of a species, initially in state S, undergo
random mutations, selective pressures will ensure that, in the long
run, those mutations which are conducive to the survival and
replication of the species will prevail,

and

(14)

Trait Y was conducive to the survival and replication of the
species,

therefore

(15)

Genetic mutations which led to Y prevailed,

therefore

(16)

Trait Y prevailed within the species.

We can contrast the intentional explanation given by William Paley for
the reliability of the optical system with the explanation given by evolu-
tionary biologists Dan-Erik Nilsson and Susanne Pelger, and popularized
by Richard Dawkins.

The authors model the way in which random

mutations of a light-sensitive patch on an animal would be likely, given
selective pressures, to yield a focused-lens eye.

In terms of the schema just outlined, the initial state S is the existence of

an animal with a light-sensitive patch. Nilsson and Pelger show (by way of
mathematical modelling) that if all subsequent mutations are random, then
over a period of time n one would expect a chain of mutations (each of
which is favoured because it is an optimal response to selective pressures) to
yield the focused-lens eye. The time period n depends on the gene

’s rate of

mutation. Given the life-cycle of the initial animal with the focused-lens
eye, and on the assumption that genetic mutations across generations are

Dan-Erik Nilsson and Susanne Pelger,

‘A Pessimistic Estimate of the Time Required for

an Eye to Evolve

’, Proceedings of the Royal Society of London, Biological Sciences 256 (1994): 53–8;

and Richard Dawkins,

‘Where D’you Get Those Peepers?’, New Statesman and Society, 16

June 1995, 29.

THE GAP OPENS

random rather than goal-directed, the authors

’ mathematical model ex-

plains why the focused-lens eye might have been expected to develop
from such a light-sensitive patch in the amount of time natural historians
estimate it actually took.

Selective pressures in the wider environment would lead to less adaptive

mutations dying out, and those which were optimal for propagation
becoming more numerous. Across many generations, a process which is
not genuinely teleological (in that the genetic mutations are purely ran-
dom, and are not themselves caused by any telos) would none the less tend
towards a particular outcome (namely, the emergence of the focused-lens
eye). This would occur through a chain of genetic mutations which were
each selectively bene

ﬁcial, and hence enabled their bearers to survive and

replicate in greater numbers than those who bore less bene

ﬁcial mutations.

In this way, the schema provides a fully intelligible explanation for the

ﬁnal

state of affairs, just as Paley

’s does, but it does so without postulating any

intentional activity.

For the purposes of this argument, I am inevitably simplifying the

accounts of natural selection given by contemporary biologists. Not all
scientists would accept that the process of evolutionary adaptation is this
effective in moving species towards selectively optimal mutations. In
particular, Stephen Jay Gould and Richard C. Lewontin have criticized
the tendency to assume that all the causally signi

ﬁcant traits in evolution at

any stage will be ones speci

ﬁcally selected for at the previous stage.

They

suggest instead that many causally signi

ﬁcant traits will have arisen as mere

side-effects of the traits previously selected for. The popular account of
natural selection which we see in the writings of Richard Dawkins and
Daniel Dennett seeks to account for nearly all signi

ﬁcant traits by claiming

that their emergence was

‘selected for’ (that is, the trait has appeared

because it helps the creature survive and replicate). Gould and Lewontin
maintain, by contrast, that many such traits are simply incidental: signi

ﬁ-

cant traits often emerge, not because they are selectively bene

ﬁcial but

because they are side-effects of other traits which are selectively bene

ﬁcial

(much in the way that the spandrels of the Basilica of San Marco in Venice
are side-effects of the arch shape which has in fact been

‘selected for’ by the

See S. J. Gould and R. C. Lewontin,

‘The Spandrels of San Marco and the Panglossian

Paradigm: A Critique Of The Adaptationist Programme

’, Proceedings of The Royal Society of

London, Series B 205(1161) (1979): 581

–98.

FROM MORALITY TO METAPHYSICS

architects). Thus, Gould and Lewontin argue, it is a mistake to seek a
quasi-teleological explanation for all signi

ﬁcant traits, for the process by

which such traits emerge is more haphazard than Dawkins and Dennett
would have us believe.

Because of these considerations, Gould and Lewontin regard natural

selection as having a relatively smaller impact, and conclude that the
overall outcome will be less strongly goal-directed than my initial sketch
suggested. If they are correct, the capacity of natural selection to explain
the prevalence of selectively optimal traits will be weaker than I implied.

I will pass over their detailed arguments, because the signi

ﬁcant factor for

my overall case is the fact that natural selection can in principle explain the
prevalence a trait in terms of its selective optimality. It does not matter to
my argument whether such an explanation is correct for all human traits.
My overall claim is that it is legitimate to press (D3) with respect to the
truth-tracking quality of humans

’ moral faculties because we see in natural

selection a way in which (D3) can be answered for our truth-tracking
capacities for theoretical reasoning and our truth-tracking capacities with
respect to the physical world.

Nilsson and Pelger

’s work clearly enables us to answer (D3). The issue

they are addressing is not how we justify faith in our perceptual capacities
(as D1 requests), nor simply how we give a natural history of the particular
stages of the emergence human beings with these capacities (which would
meet D2). In their account of the development of human perceptual
capacities, evolutionary theory explains why our capacities track the
truth about the physical world around us. What would otherwise be
deeply puzzling as the product of a set of random mutations is rendered
intelligible as the product of a larger evolutionary history in which the
ability to track the truth is selectively advantageous.

I think Daniel Dennett summarizes the debate quite fairly:

‘To see what the issue is,

consider the popular retort of software engineers when somebody

ﬁnds a ﬂaw in their

program:

“It isn’t a bug, it’s a feature.” In other words, it’s supposed to be that way; it was

designed to work like that. Now is everything observable in the biosphere a

“feature,” an

“adaptation,” as an evolutionary biologist would call it? Are there no bugs, no undesigned
bits, no historical accidents? Of course there are. Everybody agrees on that. Everybody does
not agree on how important these non-adaptations are; the greater the role you give to
natural selection, the more

“adaptationist” you are, and the Gould/Lewontin essay was an

attempt to swing opinion the other way.

’ (Daniel C. Dennett, ‘The Scope of Natural

Selection

’, Boston Review (October/November 1996), online at <http://bostonreview.net/

BR21.5/dennett.html>).

THE GAP OPENS

An analogous explanation is possible in the case of theoretical reasoning.

Natural selection is the obvious candidate for an explanation of the
development within humans of truth-tracking capacities with respect to
the fundamental principles of deduction, Inference to the Best Explana-
tion and (if it is a distinct principle) induction. It is highly plausible to claim
that humans will be better able to survive and multiply if their cognitive
faculties enable them to come to true beliefs regarding explanation,
inference and induction. Among competing groups of humans (and
their evolutionary precursors) such abilities would be bene

ﬁcial to the

survival and

ﬂourishing of the collective. Thus we can see why there might

well be a harmony between the principles of reasoning which are truth-
yielding and those which emerge from random genetic mutations and
selective pressure. In consequence, natural selection offers a plausible story
of how humans come to have truth-tracking capacities for theoretical
reasoning.

In both the cases of physical perception and theoretical reasoning, the

quasi-teleological explanation offered by natural selection does not require
us to postulate a benevolent deity or a casually ef

ﬁcacious Good. All that

has to be established is that the verisimilitude of the faculty in question is
bene

ﬁcial for the species’ survival and multiplication.

2.4.3 Implications

The possibility of an evolutionary explanation of humans

’ other cognitive

capacities undermines Dworkin

’s argument for resisting (D3) in the moral

case. His claim was that (D3) was an unreasonable and excessive demand
for explanation. But the cases of both perceptual beliefs and theoretical
reasoning show that (D3) is not only a demand that can legitimately be
made but that it is a demand that can often be met. There is clearly
potential for an explanation that is neither a rehearsal of the justi

ﬁcations

Ralph C. S. Walker, The Coherence Theory of Truth: Realism, Anti-Realism, Idealism

(London: Routledge, 1989) and Alvin Plantinga, Warrant and Proper Function (Oxford:
Oxford University Press, 1993) offer more radical objections to the possibility of natural
selection explaining our truth-tracking capacities with respect to theoretical reasoning. If
either argument succeeds, then mine is obviously undermined. I will assume for the sake of
this thesis that they fail

—for if either does succeed, it will be cold comfort to the secular

positions I am criticizing. If one succeeds, there is an

‘explanatory gap’ in secular thought of an

even greater magnitude than I am suggesting.

FROM MORALITY TO METAPHYSICS

for our moral beliefs nor a

‘moron’-type explanation. The justiﬁcation

advanced in Chapter 1 for trusting our

‘bedrock’ principles of theoretical

reasoning made no reference to these evolutionary considerations. For the
arguments against scepticism stand quite independent of the explanation of
the genesis of our cognitive capacities. (D1), as we have seen, is logically
distinct from (D3). What natural selection offers in these other cases is not
merely a description of the speci

ﬁc faculties that evolved but an explana-

tion of the faculties having the property of tracking truth.

2.5 The Explanatory Gap Opens

Having shown that (D3) is a legitimate demand to make, the second half of
this chapter will argue that it cannot be met by secular moral objectivists,
and hence that their position generates an explanatory gap.

Whereas

there is a plausible correlation between maximal survival value and truth in
the case of perceptual beliefs and theoretical reasoning, no such correlation
is plausible in the moral case.

On the account given by evolutionary biology, it is not the fact that

moral beliefs are correct which leads to them being selected for. Rather it
is the fact that they are conducive to the

ﬂourishing of the collective.

However, it might seem plausible to see that quality as having a non-
coincidental connection with these beliefs

’ objective truth.

Following this line of thought, one might argue that the

‘explanatory

gap

’ could be ﬁlled, not by saying that the evolutionary process tracks the

rightness of moral beliefs, but that the process and objective rightness both
track

‘collective ﬂourishing’. If this were so, the harmony between the

beliefs produced by process and that which is objectively right would not
be a cause for surprise.

I shall call someone who advances this argument a

‘moral Darwinian’.

There are two immediate problems which face the Darwinian account of
the harmony between what is useful and what is objectively right.

As well as distinguishing this explanatory gap from any justi

ﬁcatory gap, it needs to be

distinguished from the gap in the human capacity to act morally which is argued for in John
Hare, The Moral Gap: Kantian Ethics, Human Limits and God

’s Assistance (Oxford: Oxford

University Press, 1997).

THE GAP OPENS

2.5.1 What is

‘ﬂourishing’?

The

ﬁrst problem is that the term ‘ﬂourishing’ is highly ambiguous. We are

used to deploying it as an evaluative term. To say that something is
‘ﬂourishing’ implies that things are well with it. Biologically, however,
all we have good reason to believe will be encouraged by natural selection
is a species

’ survival and replication. There is no guarantee that the qualities

which lead to multiplication will have any other excellence about them.

Evolutionary biology may be able to account for the value we place on

pleasure, for we will, on the whole, evolve to

ﬁnd things pleasant if they

are conducive to survival and replication. Sexual intercourse is an obvious
example. Of course, our cognitive apparatus does value things which do
not maximize survival, replication, and pleasure. But it is hard to see how
this fact will support the moral Darwinian

’s case. For all the valuations we

make which are not immediately analysable in terms of survival, replica-
tion, and pleasure will fall in to one of two categories. Either (i) there is a
less obvious way in which they do promote survival, replication, and
pleasure or (ii) there is not. If (i) is true, then moral Darwinism is vulnera-
ble to the objection I make below in 2.5.2, namely that a values system
based on these three qualities alone is morally objectionable. By contrast, if
(ii) is true, then the valuations in question have no evolutionary purpose.
To borrow Gould and Lewontin

’s terminology, such valuations are ‘span-

drels

’. Spandrels are thrown up by the natural process but they lack any

direct connection with genetic survival and multiplication. In so far as
valuations resemble spandrels, they are essentially aimless by-products of
the evolutionary process. Unless we have a wider teleological account

—

say a theistic account on which humans

’ apparently haphazardly generated

valuations are providentially aligned with the objective moral order

—we

have no reason to suppose that these valuations have any non-random
connection with that objective moral order.

2.5.2 Moral objections

The second problem is that a value system based solely on survival, replication,
and pleasure yields results that are quite immoral. Philip Kitcher makes the
point well with the following thought-experiment:

FROM MORALITY TO METAPHYSICS

result from resentment at past cruel treatment, from recognition of the miserable
lives that offspring would have to lead, from sickness, or whatever.) Under these
circumstances [the principle of maximizing the survival of the collective] requires
the willing person to coerce the unwilling person, using whatever extremes of
force are necessary

—perhaps even allowing for the murder of those who attempt

to defend the reluctant one.

The moral Darwinian might reply that the moral principles and attitudes
which guide us through normal cases are not well-suited for such a
contrived example. Can we be so sure what would be right in this extreme
situation? Even if we accepted this highly questionable response, the
problems facing the moral Darwinian would be far from over. For it is
not only in such extreme scenarios that the moral principles and attitudes
that have prevailed in the re

ﬂective equilibrium process do not judge

collective

ﬂourishing to be an overriding good. A huge number of the

dif

ﬁcult moral issues we deal with in daily life—especially in politics and

international relations

—are concerned with the trade-offs which have to

be made between the

ﬂourishing of different human beings.

This was the issue at the heart of the

‘sociobiology’ controversy in

the 1970s. The publication of E. O. Wilson

’s Sociobiology: A New Synthesis

provoked a vitriolic response precisely because he failed to make the
distinction between what maximizes a species

’ replication and what is

right.

It appeared that, in calling for a morality based on evolutionary

imperatives, Wilson was advocating the abandonment of the weak and
vulnerable, raising the spectre of a Nazi-style attitude to eugenics. As we
shall see, Wilson

’s considered position was not that of the moral Darwinian.

None the less, his

ﬁrst, highly public, attempt at accounting for ethics in

terms of natural selection did not drive a suf

ﬁciently robust wedge between

that which is right and that which is useful. It was all too evident that, in far
less extreme cases than Kitcher

’s post-Holocaust example, the imperative

to maximizing genetic survival and replication was at odds with the com-
munity

’s most fundamental moral commitments.

It would of course have been possible for Wilson to follow the Darwini-

an to its logical conclusion, and to argue in favour of the abandonment of

Philip Kitcher,

‘Four Ways of “Biologicizing” Ethics’, in his In Mendel’s Mirror: Philo-

sophical Re

ﬂections on Biology (Oxford: Oxford University Press, 2003), 328.

Ullica Segerstråle, Defenders of the Truth: The Sociobiology Debate (Oxford: Oxford

University Press, 2000). See also E. O. Wilson, Sociobiology: The New Synthesis (Cambridge,
MA: Harvard University Press, 1975).

THE GAP OPENS

those communal moral commitments which contradicted evolutionary
imperatives. But, if the argument of Chapter 1 is correct, we already have
justi

ﬁcations for our ‘bedrock’ moral principles. What we are seeking from

the Darwinian is an explanation of how our (ex hypothesi largely truth-
tracking) capacities for moral reasoning come to be entrenched in our
cognitive apparatus. It would be wholly unsatisfactory to answer by claim-
ing natural selection could explain why we would come to an utterly
different set of moral judgements. The fact natural selection would account
for a set of moral judgements we do not hold, and have every reason to
believe false, would hardly inspire us to think (D3) had been answered.

I should stress that none of the thinkers under discussion in this chapter

has ever made such a proposal. I am not attempting to besmirch such
thinkers by suggesting they would be sympathetic to a Nazi-style ethic
which exalted survival and replication over against the needs and rights of
the weakest and most vulnerable members of the human community.
Rather, my argument is that the very fact they recoil from such a proposal
means they share a fundamental

—and, by the arguments of Chapter 1,

justi

ﬁed—set of moral convictions which ﬂow from a cognitive apparatus

whose accuracy cannot be explained by natural selection alone.

2.5.3 Evolution and revision in Darwinian ethics

The obvious riposte for a moral Darwinian to make is that, whilst the
imperatives of collective survival are not identical to moral truth, they do
provide its foundation. The Darwinian could argue that human beings
have now reached a level of conscious re

ﬂection on ethics—a reﬂection in

the past inhibited by superstition and indeed religion

—that they can now

choose a more humane set of norms than the ones natural selection alone
would generate. This is the position at which E. O. Wilson had arrived
when he wrote the following (in 1983):

Only by penetrating to the physical basis of moral thought and considering its
evolutionary meaning will people have the power to control their own lives. They
will then be in a position to choose ethical precepts and the forms of social
organization needed to maintain the precepts.

Charles Lumsden and E. O. Wilson, Promethean Fire (Cambridge, MA: Harvard

University Press, 1983), 183.

FROM MORALITY TO METAPHYSICS

Richard Dawkins makes a similar claim in The God Delusion.

The sugges-

tion both writers are advancing is that the human capacity for moral
re

ﬂection can emerge from natural selection and then pull itself up, as it

were, by its own bootstraps

—removing those elements of the value-system

bequeathed by the evolutionary process which are cruel and excessively
competitive.

This response is unconvincing. As we have seen, the only component of

our existing moral capacities whose more than lottery-like accuracy is
accounted for by natural selection is precisely the component which tracks
collective survival and replication. From the theory of natural selection

—

whether in its

‘Panglossian’ adaptationist form, or in the more modest

version outlined by Gould and Lewontin

—we have no explanation of

the verisimilitude of those of our capacities which generate moral judge-
ments which con

ﬂict with the imperative to maximize species-replication. If

Wilson and Dawkins are right, we correct the inadequate principles which
natural selection gives us by relying on other parts of cognitive apparatus.
The problem is that their accounts give us no explanation whatsoever of
why those parts of our apparatus are able to track the objective moral truth.

Two responses are open to the moral Darwinian. Both of them seek to

root our capacity for truth-tracking moral judgement in our ability to
grasp the norms of theoretical reasoning.

2.5.4 Moral truths as analytic?

The most extreme response would be to argue that all true moral judgements
are analytic statements (that is to say, they are statements the negation of
which entails a logical contradiction.) If moral judgements were analytic
statements, then our capacity for moral knowledge would

ﬂow directly

from our capacities for deductive reasoning. If this were true, the moral
Darwinian could offer precisely the explanation of our truth-tracking
capacities with respect to moral reasoning that I advanced with respect
to theoretical reasoning at 2.4.2.

One philosopher who takes moral truths to be analytic is Richard

Swinburne. This may come as a surprise, for it is far from his intention
to advance the Darwinian position. When Swinburne makes the claim
that moral judgements are analytic, he is preoccupied by a completely

Richard Dawkins, The God Delusion (New York: Random House, 2006), 222.

THE GAP OPENS

different set of concerns, namely the relationship between the necessity of
moral truth and the omnipotence of God. None the less, the claim he
makes is precisely the one the Darwinian needs.

Swinburne elucidates his position as follows: by

‘a logically necessary or

analytic statement

’ he means ‘a statement, the denial of which states

nothing which it is coherent to suppose could be true.

’ I take it that

Swinburne means that the negation of the statement involves a logical
contradiction. (I make this assumption, not merely because this is the usual
philosophical usage, but because it is only on such a reading that Swin-
burne

’s argument provides the Darwinian with any kind of respite. Unless

‘analytic’ means ‘logically necessary’, humans will need some kind of
truth-tracking capacity distinct from our other capacities for theoretical
reasoning, in order to work out which moral statements are

‘coherent’ and

which are not.)

Swinburne offers the following examples of allegedly analytic moral

truths:

Among the necessary moral truths one would expect to

ﬁnd general principles of

conduct such as that one ought to care for one

’s children, not punish the innocent,

not tell lies (together with whatever quali

ﬁcations are needed).

It is hard to see how these examples can qualify as (in the usual sense)
analytic truths, for there seems no logical contradiction involved in asserting
that one ought not to care for one

’s children or that there are circumstances

in which the innocent will need to be punished. In the

ﬁrst case, there

seems no prospect whatsoever of recasting this as a statement one must
accept on pain of self-contradiction. In the second the nearest logically
necessary statement would be is that

‘the innocent do not deserve punish-

ment

’. However, this statement would only be analytic precisely because it

made no substantive assertion. It would not say that people who have
behaved thus-and-so do not deserve to be punished, but that if behaving
thus-and-so leaves one innocent, then one deserves not to be punished. And
that is hardly informative. Swinburne

’s third example (that ‘one should not

tell lies

’) is robbed of substantive content in precisely by the fact that he

needs to add

‘(together with whatever qualiﬁcations are needed).’

R. G. Swinburne,

‘Duty and the Will of God’, Canadian Journal of Philosophy 4 (1974):

217

–18. This same argument is repeated in his The Coherence of Theism (Oxford: Clarendon

Press, 1993), ch. 11, and Responsibility and Atonement (Oxford: Clarendon Press, 1989), 126 ff.

FROM MORALITY TO METAPHYSICS

As T. J. Mawson observes, in a critique of Swinburne

’s argument,

‘Necessarily, if one makes whatever qualiﬁcations are needed to make a
given thing logically necessary, then that thing will be logically necessary.

’

Mawson goes on to dissect Swinburne

’s position in more detail:

In discussion, Swinburne has claimed that the quali

ﬁcations spoken of were meant

to be taken as ones given in

‘non-moral terms’. However, if so, the qualiﬁcations

could never be suf

ﬁcient to make it analytically/logically necessary that one ought

not to tell lies given those quali

ﬁcations, just as any description of the planets in

non-mathematical terms could never be suf

ﬁcient to make it analytically/logically

necessary that there be nine of them.

In short, Swinburne is seeking to have his cake and eat it. For the purposes
of his wider dialectic, Swinburne needs to show that moral truths are both
analytic (so that they impose no limitations upon God

—an issue we return

to below at 7.3.1) and substantive (for otherwise they cannot guide our
actions):

Given that necessary moral truths are [for Swinburne] analytically or logically
necessary and thus cannot guide our actions, providing us with

‘general principles

of conduct

’ is exactly what we should expect necessary moral truths not to do. Why

then does Swinburne suggest that they will do this? Why are his putative examples
of necessary moral truths so ill-chosen to illustrate the point that necessary moral
truths merely state the relationship between moral ideas? The answer is that
Swinburne wants some necessary moral truths to do more than we have allowed;
he wants some necessary moral truths to stretch out beyond the realm of relations
of ideas and state substantial moral facts.

If moral truths were analytic, all that would be required in order to work
out how humans ought to behave would be logical analysis. Mawson

’s

argument shows us why this is implausible. Analytic statements lack the
substantive content which moral judgements so clearly contain. So the
question resurfaces for the moral Darwinian: how do we account for
the fact humans have capacities which can track these substantive truths?

2.5.5 A hybrid theory?

A more moderate response which is open to the moral Darwinian is to
develop a hybrid theory of ethics. On such an account, moral truth might be
generated by a combination of one or two synthetic a priori principles and

T. J. Mawson,

‘God’s Creation of Morality’, Religious Studies 38 (2002): 4.

THE GAP OPENS

our normal processes of theoretical reasoning. Something like this view is
defended by Roger Crisp in Reasons and the Good. Crisp uses the process of
re

ﬂective equilibrium to defend a very limited set of synthetic a priori

principles. He argues that we have both reason to increase the pleasure and
avoid the pain of ourselves and others. In Crisp

’s view, these are the only

ultimate reasons for action.

His defence of these principles addresses (D1). His hedonistic ethic also

looks capable of answering (D2) and (D3), for there is an obvious evolu-
tionary basis for Crisp

’s ethic. From a biological point of view, pleasure and

pain are mechanisms which have developed to guide humans towards the
forms of behaviour that will maximize their chances of survival and
replication.

In defending his hedonistic ethic, Crisp argues that many of the other

things we now hold to be valuable are ultimately derivative from these
two fundamental values. An obvious challenge to his position is presented
by Robert Nozick

’s famous thought-example:

The Experience Machine. Suppose there was an experience machine that would give
you any experience you desired. Superduper neuropsychologists could stimulate
your brain so that you would think and feel you were writing a great novel, or
making a friend, or reading an interesting book. All the time you would be

ﬂoating

in a tank, with electrodes attached to your brain . . . Would you plug in? What else
can matter to us, other than how our lives feel from the inside?

The Experience Machine appears to provide a reductio ad absurdum of views
which reduce moral value to subjective states. Nozick

’s thought-example

is a contribution to the re

ﬂective equilibrium process—reminding us that

the vast majority of human beings also value things other than these
subjective states. Nozick argues that this thought-example shows three
things about that which humans

ﬁnd valuable. They are (1) accomplish-

ment

—we want to do certain things; (2) personhood—we want to be a

certain kind of person; (3) authentic understanding

—we want to be able to

make contact with a reality deeper than one that is entirely man-made.

Crisp accepts that the hedonist is committed to the counter-intuitive

view that a life lived entirely in the experience machine is a good life, and

Robert Nozick, Anarchy, State, and Utopia (Oxford: Blackwell, 1974), 42

–3.

The three terms in italics are used by Roger Crisp in his paper

‘Hedonism Reconsidered’

(delivered to the Minnesota Interdisciplinary Workshop on Well-Being, 24

–6 October

2003),15, online at <http://www.utilitarianism.com/hedonism-reconsidered.html>

FROM MORALITY TO METAPHYSICS

that accomplishment, personhood and authentic understanding are there-
fore only of derivative value. His defence is that the independent value we
ascribe to non-hedonic states helps us to maximize our pleasurable states
(he calls this

‘The Paradox of Hedonism’). His explanation of our non-

hedonistic intuitions is that we have evolved as goal-seeking beings and
enjoy the process of achieving goals. He suggests that belief in the indepen-
dent value of those goals can itself increase that enjoyment. However, this
does not mean these goals have an independent value, for

‘there is a further,

external, perspective to take on our non-hedonistic evaluative beliefs, and
this involves considering their historical origin.

’ For example, he explains

our belief in the independent value of accomplishment as follows:

It goes almost without saying that the values each of us holds are at the very least
heavily shaped by the cultural and social practices in which we found ourselves
from a very young age. The attitudes of others, especially attitudes of our parents
involving praise and blame, have a huge in

ﬂuence on what we end up valuing. At

this point, we can pull back the focus to consider the development of human values
as a whole from their origins in groups or societies very different from our own. It
would be surprising if human values had not been affected to some extent by the
attitudes of our hunter-gatherer ancestors in the Stone Age

. . . Those who

achieved more in the

ﬁeld—who brought back more meat, fungi or fruit—

would have been rewarded by their fellows, partly with a larger share of the
available goods, but also with esteem and status within the group.

Crisp

’s explanation is deliberately deﬂationary—it aims to undermine the

beliefs for which it is giving an account. He concludes by asking:

Could it not be that our valuing of accomplishment is an example of a kind of
collective bad faith, with its roots in the spontaneous and largely unre

ﬂective social

practices of our distant ancestors?

There seems an inconsistency in Crisp

’s position. The fact that there is an

evolutionary story which accounts for our valuing of accomplishment is
taken as a reason for thinking that accomplishment is not genuinely valu-
able. Could the same case not be made against our valuation of pleasure

—

where an evolutionary story is likewise available?

Crisp replies as follows:

Again, there is an evolutionary story to be told about its origins. But on re

ﬂection its

status as a value seems less contingent and dependent on the views of others. The

Roger Crisp, Reasons and the Good (Oxford: Oxford University Press, 2006), 121

–2.

THE GAP OPENS

values Nozick mentions are highly culturally relative: some cultures seem to have
emerged without the individuality they involve (and presumably one might tell an
evolutionary story about the development of non-individualist, communal values).
But all human beings have always, in different ways, sought pleasure and avoided
pain.

As an empirical generalization, this seems dubious. If universality of
acceptance is the criterion, Crisp

’s problem will be that hedonism goes

against the

ﬁrst-order ethics of all cultures. Whilst there may be doubts as

to the precise kinds of interaction with other real people and of accom-
plishments which are of value, these look like universal human values.

More pressingly, Crisp needs a more general account of the truth-

tracking nature of our moral intuitions. In the case of the intuitions to
which Nozick appeals, Crisp thinks the fact we can

‘tell an evolutionary

story

’ undermines their claim to truth. His only reason for placing greater

trust in our intuitions on pleasure and pain is that

‘on reﬂection its status as

a value seems less contingent

’. But how are we to explain the truth-

tracking capacity of this very process of

‘reﬂection’—reliant as it is on

moral intuitions as well as theoretical reason?

The question is even sharper when we come to Crisp

’s discussion of

distributive justice. Once again, he seeks to build on a limited number of
synthetic a priori principles for distributing welfare which we

ﬁnd intui-

tively plausible. In doing so, his discussion provides us with a good
example of the re

ﬂective equilibrium process.

But Crisp offers no

explanation of why natural selection would yield an accurate set of
synthetic a priori principles for distributive justice. The principles of
distributive justice which Crisp defends will not always maximize human
replication, and he makes no claim that they would. Once again we face
the dilemma outlined in 2.5.3. Either we must abandon some of our most
fundamental

ﬁrst-order moral commitments with regard to the protection

of the weak and vulnerable or we must correct the morally inadequate
principles which

ﬂow from the imperative to maximize survival and

replication, and we can only do this by using the

‘spandrel-like’ parts of

our cognitive apparatus. As we have already seen, we have no explanation
of how such

‘spandrels’ come to track the truth.

Crisp,

‘Hedonism Reconsidered’, 17.

Crisp, Reasons and the Good, ch. 6.

FROM MORALITY TO METAPHYSICS

2.5.6 Ethics as an application of theoretical reasoning?

So far, I have considered only two possibilities: (i) that moral truths are
analytic or (ii) that they are cognized via a faculty distinct from theoretical
reasoning. I have followed Mawson in arguing against (i), and I have
claimed that from an evolutionary point of view, the faculty required by
(ii) would be a

‘spandrel’—and hence we would have no explanation for

its reliability. Is there a third possibility?

Could moral reasoning just be the application of our capacities for

theoretical reasoning to a new subject matter? If so, the explanatory gap
would disappear completely. Although the capacity to reason well in
ethics might not itself be selectively advantageous, our capacity to grasp
moral truths would now be explained by the fact that our more general
capacity to reason is selectively advantageous. On this account, moral
reasoning (though not, contra Swinburne, a matter of grasping analytic
truths) would be

‘on all fours’ with complex theoretical reasoning. Both

would be activities which were not in themselves selectively advantageous,
but involved humans in deploying a capacity for reasoning well which had
evolved because it was selectively advantageous when applied to other
subject-matters.

If reasoning were simply a mechanical process, then its application to a

new subject area would simply be a case of

‘doing the same thing’—and its

reliability in the area of theoretical reasoning might be transferrable to the
domain of practical reasoning. But (as I will argue at length in my discus-
sions of Scanlon, Foot and McDowell), moral reasoning involves cognitive
capacities which are very clearly distinct from theoretical reasoning. De-
veloping a sensitivity to moral reasons is a very different process from
learning to be an excellent theoretical reasoner, and one can be an
excellent theoretical reasoner and have the moral sensibilities of a socio-
path. It is precisely because moral reasoning involves this kind of (affective,
emotionally engaged) response to the subject-matter that writers such as
Blackburn and Gibbard argue for moral subjectivism. In Chapters 4 and 6
I will argue that their conclusion is mistaken

—and that this kind of

affective engagement is part of how practical reasons are grasped. But
precisely because practical reasoning involves this kind of affective engage-
ment (and, pace McDowell, we require the virtues in order to perceive the
moral landscape) it is not plausible to characterize practical reasoning as the
application of the very powers of theoretical reasoning with which

THE GAP OPENS

evolution has endowed us. There are distinct capacities involved in prac-
tical reasoning

—and, in terms of evolutionary explanation, these are

merely

‘spandrels’. For the secular philosopher, their accuracy remains a

matter of inexplicable good fortune.

2.6 Conclusion

When asked why human beings have a capacity for tracking moral truth,
two obvious responses are available to the secular theorist. One is to deny
that an answer need be given; the other is to proffer natural selection as an
explanation. This chapter has offered a prima facie argument against both of
these responses. It has argued that the kind of explanation being demanded
is legitimate, because exactly the same kind of explanation is available with
respect to the human capacities for tracking many other kinds of truth. It
has further argued that natural selection (which explains these other capa-
cities) is unable to explain human cognition of the objective moral order.

The accuracy of the moral intuitions and commitments that lead us to

have a concern for the weak and to object (in Kitcher

’s example) to the post-

holocaust rape are not explicable with the limited resources available to
the Darwinian. Our moral intuitions do not look to be either reducible
to the bare insight that human pleasure, survival, and replication are of value
or securable by any work our faculties of theoretical reasoning can do with
that bare insight. Dawkins

’ proposal that we in some way pull ourselves up

by our Darwinian bootstraps is untenable. These arguments show that we
can only

‘pull ourselves up’ because we have truth-tracking capacities with

respect to an objective moral order which natural selection is unable to
explain.

This chapter has only made a prima facie case for the existence of this

explanatory gap. A whole range of responses remain available to secular
philosophers. Such responses will seek to do justice to the arguments
I have sketched out in Chapter 1 (for taking moral norms to be objective
and irreducible to non-moral ones) while closing the explanatory gap
identi

ﬁed in this chapter. It is to the most powerful and prominent of

these secular philosophers that we now turn.

FROM MORALITY TO METAPHYSICS

PART II

Secular Responses

As we have seen, theories of meta-ethics have to respond to competing
pulls; one to objectivism and the other to ontological parsimony. I have
plotted the theories which feature in this book by the way they balance
these forces (see Figure 2). At one end of the axis we

ﬁnd the ‘error theory’

on which the language of objectivity is thought to embody a metaphysical
mistake. At the other are positions (which I will explore in Part III) that
endow the universe as a whole with purpose.

Between these two extremes lie a wide range of positions, some more

responsive to one pull than another. What they have in common is an
attempt both to do justice to our prior moral commitments and to avoid
ascribing purpose to the universe as a whole.

As I explained in Chapter 2, the combination of full-blown moral

objectivism with a non-purposive account of the universe seems to be
problematic. Such a combination looks incapable of explaining the ability
of humans to form beliefs on moral issues which track what (on its
account) is a wholly independent truth.

Between the error theory and such objectivism lies a range of positions,

illustrated in Figure 2. In Part II of this book, I ask whether any of these
intermediate positions can avoid the problems facing their more extreme
competitors.

Among these intermediate positions, I have chosen to begin with those

that are less objectivist. Because they do not postulate an ontologically
independent moral order, their accounts are the least likely to generate an
‘explanatory gap’. In Chapters 3 and 4, I ask whether they can also do
justice to the pull of objectivism, and hence to those features of our
everyday moral discourse which are deliberatively indispensable.

NON-PURPOSIVE (‘SECULAR’) ACCOUNT OF UNIVERSE

PURPOSIVE ACCOUNT

MORAL QUASI-

REALISM

(Blackburn / Gibbard)

‘ANTI-ANTI–

REALISM’

(McDowell)

OBJECTIVISM

(Crisp / later Scanlon)

THEISM

ERROR

THEORY

(Mackie)

CONSTRUCTIVISM

(Korsgaard / early Scanlon)

NATURALISM

(Foot)

AXIARCHISM

(Leslie / Rice)

Objection:
Argument from
Deliberative
Indispensability

Objection:

Position generates an

‘Explanatory Gap’

< < Pull of reductionism

Pull of objectivism > >

Figure 2 Map of meta-ethical theories, and objections to different positions

FROM MORALITY TO METAPHYSICS

Alternatives to Realism

Simon Blackburn and Allan Gibbard

3.1 Introduction

What to do is remarkably fact-like in its behaviour

. . . This is not a

brute truth of metaphysics, but an intelligible consequence of the
nature of planning and deciding. We are agents, in the philosopher

’s

lingo, beings who can reason our ways to decisions. Anyone who
reasons what to do is committed to something very much like facts of
what to do. Reasoning commits us to thinking in terms of conclusive
to-be-doneness.

Allan Gibbard, Thinking How to Live

3.1.1 Quasi-realism in ethics

This chapter will evaluate the theory of ethics developed by Simon Black-
burn and Allan Gibbard. As the above quotation indicates, they take the
pull of moral objectivism seriously. Gibbard and Blackburn are acutely
aware of the

‘fact-like’ considerations to which our practical deliberation

makes appeal. While they respect (what Blackburn calls) the

‘realist-

seeming grammar

’ of practical deliberation, they also seek to minimize

its metaphysical implications.

Blackburn and Gibbard differ in their wider metaphysics. For Black-

burn, this theory of ethics is part of a

‘quasi-realist’ project which covers a

wider domain of human thought.

For Gibbard, by contrast, the entire

Allan Gibbard, Thinking How to Live (Cambridge, MA: Harvard University Press, 2003), 5.

Simon Blackburn,

‘Errors and the Phenomenology of Value’, in his Essays in Quasi-

Realism (Oxford: Oxford University Press, 1993), 152

–3. See also his Ruling Passions (Oxford:

Oxford University Press, 1998).

focus is on ethics.

It is to that more limited area that the focus of this

chapter is restricted.

In the overall dialectic of this book, Gibbard and Blackburn have a

particular signi

ﬁcance. Both philosophers set out to respond to both of the

con

ﬂicting pulls identiﬁed in the Introduction. In some ways, it is hard to

know exactly where to place the position (or positions) which emerge on
the diagram given above

—for the aim of these thinkers is to develop an

account that offers the best of both worlds, and avoids both of my
objections (depicted in italics).

Gibbard and Blackburn would deny that their work is merely an attempt

to broker a compromise between these competing pulls. Their view is that
moral psychology as well as metaphysical restraint lead away from a full-
blown objectivism. Precisely because moral discourse is about motivation
and action, they think it is a mistake to regard it as a descriptive enterprise.

The accounts they have developed fall within a broader tradition of

expressivism, which takes (what we have called)

‘moral truth’ to be

constituted by our sentiments and passions. This stands in contrast to the
positions developed by Tim Scanlon and Christine Korsgaard (to be
discussed in the next chapter) with their emphasis on the role of reason
in moral re

ﬂection.

Blackburn elaborates on this contrast in his New York Times review of

Scanlon

’s 1998 work, What We Owe to Each Other:

When some feature of things weighs with people in their deliberations, we can say
that they see it as a reason for or against a course of action. Which side of the
equation explains the other? Does the weight [of passion] come

ﬁrst and explain

what is meant by seeing something as a reason? On that side lie philosophers like
Hume and St. Augustine, who wrote that

‘in the pull of the will and of love appears

the worth of everything to be sought or avoided, to be thought of greater or less
value.

’ On the other side lie philosophers owing allegiance to Plato, Aristotle and

sometimes Kant. They hold that our passionate natures come entirely under the
control of truth and reason. Apollo rules Dionysus. The trouble with this sunny
picture is that Apollo

’s control is unintelligible, for beliefs that are not about the

natural world are eminently dispensable. Why should we care about anything they
supposedly represent? And if we did, wouldn

’t this care itself be an intrusion from

the dark, a present from Dionysus? On St. Augustine

’s side there is no difﬁculty: we

talk of reasons to re

ﬂect the fact that we already care.

Gibbard,

‘Critical Notice of Essays in Quasi-realism’, Mind 105(418) (1996): 331–5.

Simon Blackburn,

‘Am I Right?’, The New York Times, 21 Feb. 1999.

FROM MORALITY TO METAPHYSICS

‘Dionysian’ philosophy faces the following challenge: how is the Diony-
sian to avoid the conclusion that what we call

‘moral truth’ varies as our

attitudes and passions vary? The arguments of Chapter 1 gave us a prima
facie case for thinking Apollo must rule Dionysus. Our fundamental moral
commitments, not least our abhorrence of evils such as the Holocaust,
leave us unwilling to believe that morality varies depending on what things
we care about. We care about preventing genocide precisely because of its
evil, and were our cares to change we do not suppose the evil would
change. It is only we and our moral qualities which would have changed.
We would have become morally corrupt.

An important motivation behind quasi-realism is the need to accom-

modate such insights; to build an account which

ﬁxes ‘moral truth’ with

respect to events such as the Holocaust. While Gibbard does not use the
term

‘quasi-realism’ (except in referring to Blackburn’s work) he does

write of his

‘remarkable degree of agreement’ with quasi-realism in the

area of ethics. I will therefore use the term

‘moral quasi-realism’ to refer to

both philosophers

’ positions.

3.1.2 Aim and focus of this chapter

In this chapter, I will argue moral quasi-realism cannot vindicate the
fundamental ethical commitments which I defended in Chapter 1.
(There are of course other objections to quasi-realism

—most notably

that advanced by P. T. Geach

—but the argument of this chapter is

intended to be independent of these.

My sole concern will be with the

claim that quasi-realism leaves moral truth

ﬂuctuating with variations in

human sentiments and cultural conventions.)

3.2 Quasi-Realism and Counter-Factuals

Moral quasi-realism is designed to avoid the following kind of morally
obnoxious counter-factuals

(4)

If we approved of torturing the child it would be a good act,

This problem is expounded in P. T. Geach,

‘Assertion’, Philosophical Review 74(4) (1965):

449

–65, and John R. Searle, ‘Meaning and Speech Acts’, Philosophical Review 71(4) (1962):

423

–32. Gibbard devotes the third chapter of Thinking How to Live to meeting the objection,

and it is considered by Blackburn in Essays in Quasi-Realism, 152f.

ALTERNATIVES TO REALISM

whilst keeping the ontology of ethics to a minimum. It achieves the latter
aim by taking

‘moral truth’ to be constituted by nothing over and above

empirical facts about human beings

’ cares and passions. How and why does

it achieve the former aim?

Statements like (4) are often invoked as part of a reductio ad absurdum of

less than fully objectivist positions in ethics. The argument is made that if
such positions were correct, morality would be covertly indexical. Just as,
when I say

‘My name is “Angus”’, it would be odd to reply, ‘No, it isn’t,

my name is

“Jack”’, so (on a naïvely subjectivist analysis), my saying ‘X is

good

’ is not contradicted by Jack’s saying ‘X is evil’.

Why should Blackburn and Gibbard be worried by the charge of

indexicality? Could they not simply accept it? Such a response is ruled
out by the argument from deliberative indispensability which I outlined in
Chapter 1. Moral deliberation relies upon the assumption that there is a
fact of the matter about what kinds of behaviour are and are not accept-
able. One of the key things we all do in practical deliberation is try to get
to this fact of the matter. In such deliberation, we judge their desires and
the prevailing attitudes of our society by a standard which is (at least at

ﬁrst

blush) independent of those desires and attitudes. Any theory which entails
(4) will fatally undermine this kind of moral deliberation. For if such a
theory were correct, whatever conclusion humans arrived at in such delib-
eration would be

‘right’.

Moral quasi-realism claims that (4) should be read as a statement within

ethics. Gibbard and Blackburn deny that it need be taken as a higher-
order, meta-ethical assertion. When it is considered as a

ﬁrst-order ethical

claim, they are able to agree with the objectivist that (4) is implausible and
indeed abhorrent. Just as no remotely attractive set of moral norms is going
to claim

(5)

If the child had red hair it would be a good act to torture him,

so no such set of norms is going to make moral value relative to my
variable conviction. The norms a decent person will endorse include

(6)

Whatever we felt about torturing the child, it would be an evil act.

Cf. Nick Zangwill,

‘Quasi-Quasi-Realism’, Philosophy and Phenomenological Research 1(3)

(1990): 586.

FROM MORALITY TO METAPHYSICS

3.2.1 Dworkin and moral quasi-realism

Because Gibbard and Blackburn read (4), (5), and (6) as statements within
ﬁrst-order ethics, they are happy to accept many of the assertions made by
objectivists like Ronald Dworkin. So, for example, they accept that (in
Dworkin

’s words)

The claim that abortion is objectively wrong seems equivalent, that is, in ordinary
discourse, to

. . . [the claim] that abortion would still be wrong even if no one

thought it was

. . . [A]bortion is just plain wrong, not wrong only because people

think it is.

Dworkin thinks such assertions to commit us to the existence of an
‘objective’ moral order. This is where he and the moral quasi-realists
part company. Simon Blackburn parries Dworkin

’s argument thus:

There is

. . . less to Dworkin’s stance than meets the eye. The rhetoric of ‘realism’

has stopped marking out a particular position, and retains only a talismanic
quality

. . . The proposition that there is objective truth in ethics [once quasi-

realism is in place] can only serve as a summary for [a list of

ﬁrst-order claims] . . . It

becomes a long disjunction of claims like:

‘Slavery is wrong: you’d better believe it,

or slavery is permissible, you

’d better believe it, or genocide is bad, you’d better

believe it, or genocide is good, you

’d better believe it, or . . . ’

According to Blackburn, objectivists like Dworkin make an ultimately
futile attempt to stand outside our actual commitments and still use
evaluative language. The attempt is futile because moral vocabulary only
makes sense within those lived commitments. When I

‘spin the possible

worlds

’ and ask what would be (morally speaking) true if I had different

commitments, the only sensible answer concerns what I (now, with my
current commitments) make of the imagined situation. So, if I imagine
myself approving of torture, I will of course abhor what I am imagining:

‘If

I ever approved of that, I would be a dreadful person!

’ This, the quasi-

realist claims, is what makes the threatening counter-factuals false. Because
moral statements are always made from within our commitments, (4) can
only be read as a

ﬁrst-order moral claim, and read in this way it is a claim all

decent people will obviously reject.

Ronald Dworkin,

‘Objectivity and Truth: You’d Better Believe It’, Philosophy and Public

Affairs 25(2) (1996): 98.

Blackburn, Ruling Passions, 296.

ALTERNATIVES TO REALISM

Blackburn goes on to expound his position, and its divergence from

Dworkin

’s, in more detail. He tells us there is no substantive difference

between these three statements:

(7)

Slavery is a bad system

(8)

Our opinion is this: slavery is a bad system

(9)

Slavery is a bad system. That is just our opinion.

Objectivists such as Dworkin will evidently disagree. As Blackburn ob-
serves, they

‘recoil in horror’ from (8) as an elucidation of (7). Instead, they

claim that the right elucidation of (7) is:

(10)

Slavery is a bad system. That is an opinion re

ﬂecting an indepen-

dent order of reason.

Blackburn rejects this

ﬁnal statement, writing that it ‘verges on the inco-

herent, for we have no conception of the nature of this independent order

’.

In practice, Blackburn claims, (8) amounts to just the same as (7). It is not

‘a

piece of psychological self-description but

. . . a way of giving the same

verdict.

’ (9) is the kind of statement a relativist will want to make. Accord-

ing to moral quasi-realism, we do not need to resort to realist metaphysics
to resist the relativist. Relativism is, so Blackburn claims, defused by point-
ing out the redundancy of the word

‘just’ in the statement:

The word

. . . insinuates a contrast: it is our opinion as opposed to something as

good or better. But what as good or better is imagined? If the relativist is saying that
it is just our opinion as opposed to a good or better opinion that approves of
slavery, we are in moral dispute, certainly. But it is then up to him to locate this
good or better opinion and defend it, and this we will know he is unable to do, for
we are certain that slavery is a bad system. If he is saying that it is just our opinion as
opposed to the opinion of the gods, or the opinion gestured at in [(10)], then he is
under a misapprehension about what opinion could ever be.

The

ﬁrst-order moral judgement expressed by (7) and (8) rules out the

possibility that a change in our sentiments would make slavery a good
system or torture of a child a good act. Blackburn correctly observes that

(11)

Our opinion is this (unless we change our mind about it): slavery
is a bad system,

Blackburn, Ruling Passions, 305. I am using the same propositions as Blackburn, but

renumbering them to cohere with the rest of this chapter.

FROM MORALITY TO METAPHYSICS

is not the same as

(12)

Our opinion is this: slavery is a bad system unless we change our
mind about it.

3.2.2 Possibility of an

‘objective order of reason’

What are we to make of Blackburn

’s argument? We cannot accept Black-

burn

’s claim that the moral objectivist ‘is under a misapprehension about

what opinion could ever be.

’ The argument in Chapter 1 for the existence

of objective norms of theoretical reasoning shows that we do have a
conception of what

‘an independent order of reason’ would be. When

realists

‘recoil in horror’ from the claim that

(9)

Slavery is a bad system. That is just our opinion.

they can

ﬂesh out the contrast being insinuated between ‘just’ opinion and

something more than opinion.

In the case of theoretical reasoning

‘opinion’ can aspire to correspond to

an objective order. This was the conclusion reached at 1.3.1, where we
considered the debates within philosophy of science about the principles
of theoretical reasoning. In each of those debates, we saw that there must
be a substantive fact of the matter about what we ought to believe on the
basis of the evidence generated by empirical observation and experimen-
tation. If we are to take any of our scienti

ﬁc beliefs to be justiﬁed, then

some such principles of theory-choice must exist, and these must be both
objective and prescriptive.

It is open to the moral quasi-realist to agree that objective norms of

theoretical reasoning exist, whilst continuing to deny the existence of any
such

‘independent order’ in the moral case. This is Gibbard’s position—

and implies that Blackburn overstated his case when he claims we have no
conception of what an

‘independent order of reason’ could be.

Chapter 2 provides an important motivation for this localized quasi-

realism. The case for a purely moral quasi-realism could run as follows:

‘If

our moral views do (at least sometimes successfully) re

ﬂect an independent

order of reason, it will be legitimate to ask for an explanation of how
humans have come to such a truth-tracking cognitive capacity. As
Chapter 2 has shown, natural selection, which may be used to explain
our ability to track truth in theoretical reasoning, cannot explain our
ability to track it in the moral case. Therefore, we should adopt a quasi-

ALTERNATIVES TO REALISM

realist reading of ethics while remaining wholehearted objectivists with
respect to other areas of knowledge.

’

3.2.3 Moral objection to quasi-realism

While the quasi-realist approach may avoid speci

ﬁc counter-factuals such

as (4), Stig Alstrup Rasmussen argues it cannot avoid more general
counter-factuals, such as:

(13)

Had our sentiments been different, the moral truth would have
been different too.

If moral truth is constituted by our sentiments, Rasmussen thinks some-
thing like (13) follows on directly. He goes on to argue:

The question of the mind-dependence of moral truth cannot be treated as an
internal one. The question is whether the quasi-realist account of moral truth
entails that human attitudes are constitutive of that sort of truth; and this would
seem hard to deny in view of the fact that Blackburn has human attitudes enter the
nearest he thinks we can get to a de

ﬁnition of moral truth: a moral claim is true, just

in case the claim forms part of the set of statements expressive of the best set of
attitudes. (Blackburn, Spreading the Word, p. 198.) On this account, there can be no
denying the truth that the moral truth might very well have been different had our
(tutored) sentiments been different

. . . Stripping everything but its bare essentials,

[Blackburn

’s] argument is to the effect that it is immoral to regard moral truth as

mind-dependent

. . . The rejoinder is immediate that if so, perhaps no nice people

should endorse quasi-realism, since

. . . that doctrine does make moral truth mind-

dependent.

Rasmussen

’s argument moves rather too quickly. It is open to Blackburn to

reply that Rasmussen begs the question. For Rasmussen

’s line of argument

implicitly assumes that moral judgements are matters of fact. If moral state-
ments are not factual, then when we

‘spin the possible worlds’ regarding the

attitudes we might have, it is not obvious that

‘moral truth’ must also vary.

Rasmussen assumes that as well as a

ﬁrst-order reading of statements like

(4) and (13) (that is, one on which they record our actual commitments)
there must be an

‘external reading’ (that is, one which tells us what the

‘moral truth’ would have been if a different states of affairs had obtained).
But on the quasi-realist picture there is no such thing as a

‘moral truth’

Stig Alstrup Rasmussen,

‘Quasi-Realism and Mind-Dependence’, Philosophical Quarter-

ly 35(139) (1985): 187. The citation within the quote refers to Simon Blackburn, Spreading the
Word (Oxford: Oxford University, 1984).

FROM MORALITY TO METAPHYSICS

which is distinct from

‘the standards expressive of our best attitudes’. For

Blackburn, then, all we could sensibly mean by

‘the moral truth’ is that

which

‘our best attitudes’ commend. Ex hypothesi, this is ﬁxed by our

actual attitudes. If our attitudes had been different and we had approved of
torture, this would not change the moral truth. This is because, on the
quasi-realist account, the truth is

ﬁxed rigidly by what our actual attitudes

value and condemn.

Allan Gibbard makes the same point as Blackburn, but in rather differ-

ent terms. The central image he uses is that of

‘planning’. Gibbard suggests

that talk of

‘what we ought to do’ can be recast as talk of ‘what to do’.

When I decide what to do in any imagined situation, this is equivalent to
making a plan for the situation. And, like our attitudes, our plans are

ﬁxed

by what we currently value and condemn.

This allows Gibbard, like Blackburn, to make a distinction between

moral sentiments and mere tastes. We can see this with respect to two
contrasting cases: my dislike of strawberries (a mere taste) and my disap-
proval of kicking dogs for fun (a moral sentiment).

In Blackburn

’s terms, my dislike for strawberries need not involve the

second-order judgement

(14)

Were I to like strawberries, they would still taste awful,

whereas my abhorrence for kicking dogs does involve the second-order
judgement

(15)

Whatever we felt about kicking dogs, it would be morally
wrong.

We can make the same point in Gibbard

’s language of plans. Supposing

I know that in

ﬁve years, I will like strawberries. Then, even though I hate

them now, it is perfectly rational for me to take advantage of the futures
market and buy an option on some very cheap, high-quality strawberries.
By contrast, if we take a course of action which I abhor, such as kicking
dogs for fun, my current judgement that we ought not to kick dogs for fun
is a matter of

‘planning to avoid kicking dogs for fun, planning this even

for the contingency of being someone who approves of such fun, and who
is surrounded by people who approve.

’

Gibbard, Thinking How to Live, 186.

ALTERNATIVES TO REALISM

Scanlon criticizes this formulation as being unintelligible:

What does it mean to plan not to do x even for the contingency in which one
approves of doing x? It does not seem that one can plan to do something under
certain circumstances while knowing or believing that if one were in those
circumstances one would not do it.

As with Rasmussen, Scanlon

’s argument moves too quickly. I think we

can make sense of this kind of

‘planning’. Consider the following example:

As a committed left-wing radical student, I might anticipate the corrupting
effects of earning a wage and climbing the property ladder. To subvert this
later

‘sell-out’ I bind my future self by entering into a contract which signs

that all my future annual earnings over £30,000 to the Workers

’ Revolu-

tionary Party. There is nothing incoherent about this plan, and it seems to
ﬁt precisely Scanlon’s description of planning not to do something even
for the contingency in which I approve of doing it.

It seems less plausible to interpret my moral judgements about others as a

form of planning. As John Hawthorne observes, when I come to a conclu-
sion about what someone else should do, this looks like a straightforward
belief

—if Jim is trapped in a burning building, I may judge that ‘the thing

for Jim to do is get out of there fast

’.

Gibbard

’s only way of making sense

of this judgement that Jim ought to do x is by claiming that it expresses my
plan to do x if my situation is identical with Jim

’s in all relevant respects.

This seems an excessively convoluted reading of a very simple judge-

ment. To think that the thing for Jim to do is get out fast cannot simply be
to have a plan for that rather unlikely circumstance

—even if thinking Jim

should get out does entail that I should get out in an identical circumstance.
Rather, both thinking Jim should get out and having a plan for myself in
such unlikely circumstances are explained by the fact that I accept the norm
that in Jim-type situations, x is the thing to do. As Hawthorne observes, it is
the idea of a norm that is doing all the work here, not the idea of a plan.

Quasi-realism can in fact accept that norms and not plans are doing the

work in our moral judgements about what others should do. All this
example shows us is that Blackburn

’s account (which frames the whole

T. M. Scanlon, The John Locke Lectures 2009: Being Realistic About Reasons (online at

www.philosophy.ox.ac.uk/lectures/john_locke_lectures), Lecture III, 7.

John Hawthorne,

‘Practical Realism?’, Philosophy and Phenomenological Research 64

(2002): 171.

Cf. also Scanlon, The John Locke Lectures, Lecture III, 10.

FROM MORALITY TO METAPHYSICS

issue in terms of our approval of norms) provides a more illuminating
explanation of the position than Gibbard

’s. For this reason, I will focus on

Blackburn

’s exposition and defence of moral quasi-realism, and only turn to

Gibbard when he has a distinct line of argument which strengthens the case.

3.3 Quasi-Realism and Provisionality

We have seen that moral quasi-realism is capable of parrying the criticisms
advanced by Rasmussen, Scanlon and Hawthorne. It offers us an intelligi-
ble distinction between mere tastes and genuine moral values, and seems
to be able to hold off the morally abhorrent counter-factuals that a more
naïve subjectivism might generate when we

‘spin the possible worlds’ and

imagine ourselves to have different sentiments.

In this section, I will argue that moral quasi-realism runs into more

serious and intractable dif

ﬁculties when it has to account for the provision-

ality with which all human beings (or at least those worthy of the epithet
‘rational’) hold their ethical views. Rational agents hold such beliefs with
some degree of tentativeness. They do not regard moral truth as being

ﬁxed

completely by their current views. The challenge for Blackburn will be
whether he can make sense of this within a quasi-realist framework.

3.3.1 Blackburn

’s initial defence

Blackburn is of course well aware of this objection. In response, he writes
that correct moral judgements are expressive, not of our current attitudes,
but of our

‘best’ ones. The use of ‘best’ in a deﬁnition of ‘moral truth’ will

need careful examination, for there is an evident danger of circularity. Is
Blackburn simply smuggling a value-laden term in to what was meant to
be an explanation of evaluative terms in non-evaluative language?

On this question of provisionality, Blackburn faces two opposing perils.

On the one hand, if he ties morality too closely to our current beliefs, his
account will leave no room for progress in moral understanding. It is part
of the content of our existing moral commitments that we value openness
to alternative views, and to the possibility of our views being revised for
the better. On the other hand, Blackburn also needs to ensure morality is
not tied to whatever, on re

ﬂection, we come to believe—for if it were, we

would be unable to avoid counter-factuals such as (4) and (13).

Our willingness to revise our views comes in no small part from the

seriousness with which we take the moral views of others. In some cases,

ALTERNATIVES TO REALISM

discussion with others simply reveals ways in which our previous views
were rationally inconsistent. If I believe all human beings ought to have
equal rights but I also have discriminatory attitudes to a particular racial
group, an interlocutor may point out this contradiction, and persuade me
to revise my attitudes. In principle, I could discover this kind of inconsis-
tency without any dialogue partner by solitary re

ﬂection on the rational

coherence of my various moral attitudes.

Moral quasi-realism should not have any problem dealing with these

cases, because logically inconsistent attitudes are an unstable foundation for
action in the same way as logically inconsistent beliefs. Just as I cannot act
on the belief that London is front of me and the belief that it is behind me
when I am trying to get there on the motorway, so I cannot act on the
belief that all human beings are equal and that one race is inferior when
I am faced with the question of how to treat a human being of that race.
Once my inconsistent attitudes have been pointed out, I have no choice
but to resolve the inconsistency. I will either treat that person differently
(and show my ultimate loyalty to the discriminatory attitude) or I will treat
them equally (and show my ultimate loyalty to the principle of equality).
Both with respect to our beliefs and our attitudes, we can of course choose
to be more or less rigorous in seeking out such contradictions. However,
there are serious dif

ﬁculties involved in consciously holding contradictory

beliefs or contradictory attitudes.

Most of us take the moral arguments of other people seriously even

when they cannot point to any inconsistency in our judgements. Precisely
because of the

‘objective feel’ of ethics, there appears to be a prima facie

assumption that the moral attitudes of others, where incompatible with
our own, should have epistemic weight. The moral objectivist can explain
this in the following way: the judgements made by other people provide
us with an independent source of access to the same objective reality,
precisely because they

ﬂow from other agents’ openness to that same

reality. Our beliefs are aiming at correctness, and so are accountable to
an

‘independent order of reason’. In consequence, we take them to be

worthy of serious examination even when we know they do not expose
any inconsistency in our existing attitudes.

It is important to note that, even on the moral objectivist

’s account, we

do not have to take all such disagreement to undermine our existing
judgements. There is no need to rule out the following possibility: there
may be ongoing arguments in which each side views the other as deeply

FROM MORALITY TO METAPHYSICS

misguided and neither can offer a non-circular account of why the other
should be taken to be in an inferior epistemic position. As we saw in 1.3.4,
such disagreements can also happen over scienti

ﬁc matters, where there

may be irresoluble arguments about the right principles of theory-choice.
Consequently, there may also be irresoluble arguments about the right
scienti

ﬁc theories.

How is the moral quasi-realist to explain the regard we have for the

opinions of other people and the mores of other cultures? On Blackburn

’s

account, regard for the opinions of others is simply another of our existing
moral attitudes. At this point, his account reaches bedrock

—whereas, we

have seen, the objectivist can explain and justify our commitment to
giving regard to the views of others. It is not simply a brute fact that we
have such an attitude; it is a consequence of taking moral truth to be
objective. Thus it is not true, as Blackburn claims, that the extra ingre-
dients a moral objectivist adds to the story of moral cognition pull no
explanatory weight. He writes that objective moral truths

‘just sit on top of

a story that tells how our sentiments relate to the natural features of
things

’.

By contrast, I have shown objectivism is uniquely able to justify

our

ﬁrst-order attitude to the beliefs and attitudes of others.

3.3.2 Gibbard on the epistemic role of others

Whereas Blackburn treats it as a brute fact that

‘our’ norms include

attaching a positive value to the views of those around us, Gibbard claims
to have an argument for that valuation. He argues that, in formulating our
own plan we should trust the judgements of competent judges

‘whose

planning issues from a deep and vivid awareness of available facts in an alert
and dispassionate frame of mind.

’ (He refers to these conditions on

competent judging as

‘conditions K’.) Gibbard continues:

We don

’t plan to defer to the judgments that any conceivable being that met

conditions K would make. Nor do we embrace much by way of speci

ﬁcations of

the conditions that qualify a planner as reliable. Rather, our plan, for any situation
S, is to defer to the judgments that people

—or most people, perhaps—who inhabit

S would make if they met conditions K. This is an impure planning matter, one
that mixes plan with fact. Specify exactly what a constitution C consists in

—

dopamine levels and all

—and I have no idea whether to defer in my planning

judgments to a person with such a constitution. For I don

’t know whether

Blackburn,

‘Errors and the Phenomenology of Value’, 153, 155.

ALTERNATIVES TO REALISM

constitution C is typical of human beings or something bizarre and alien to us. My
plan, rather, is to trust us. With many quali

ﬁcations, I plan to trust us as we are,

whatever our range of constitutions may be. I treat the typical human constitution
as trustworthy given certain vague conditions, and know that this commits me to
taking certain other kinds of constitutions as untrustworthy. I don

’t judge which

constitutions are which, except that the ones we in fact have are at least to some
degree to be relied on.

This seems to me the only coherent stance we could take in our planning, the

only stance which would not undermine its own acceptance. Otherwise, how
could I trust my own planning from one minute to the next, without reviewing all
its grounds again? If we plan at all, we are committed to attributing to ourselves
some knowledge of what to do. This is knowledge that goes beyond just happen-
ing to track the right property in our planning, just happening to track the property
that constitutes being the right thing to do. We fasten on to the right things to live
for because we are in a condition to be trusted on such matters

—so we judge.

This is a puzzling line of thought. Blackburn

’s argument was as follows: our

trust in the judgement of others is simply one of our fundamental moral
attitudes, on the same level as our abhorrence of cruelty or our valuing of
integrity. As we saw, this position had a disadvantage when compared with
objectivism (for the objectivist can explain why we have this pro-attitude to
respecting others

’ attitudes), but at least it was consistent.

Gibbard

’s thought seems to be that if I plan, I must ‘attribute’ to myself

‘some knowledge of what to do.’ In attributing to myself that kind of
knowledge, I cannot, without losing

‘coherence’, avoid attributing the

same kind of knowledge to other human beings. The reason Blackburn
has shied away from this kind of defence is presumably that such talk of
‘knowledge’ implies there is an ‘independent order of reason’ to which my
plans are seeking to correspond. And, indeed, this problem seems to arise
for Gibbard

’s account. Talk of ‘planning’ was supposed to explicate our

moral vocabulary, but here Gibbard seems to presuppose some kind of
evaluative vocabulary standing independent of my current plans. For he
does not merely say that I should have a plan to trust others, but he goes on
to argue that my trust of others is justi

ﬁed by the likelihood that they, like

me, are capable of

‘tracking the property that constitutes being the right

thing to do

’. This seems to be incompatible with the wider moral quasi-

realist account, on which

‘what to do’ just is constituted by my attitudes.

For moral quasi-realists, the

ﬂow of explanation is surely the opposite of

Gibbard, Thinking How to Live, 234

–5.

FROM MORALITY TO METAPHYSICS

that suggested in the passage of Gibbard just cited. For according to quasi-
realism, it is solely because I plan to take (or have a pro-attitude to taking)
the opinions of others into account that these attitudes are indicative of
‘what to do’. Only the moral objectivist is entitled to make the (more
plausible) claim that the explanation runs in the opposite direction.

3.3.3 Emancipatory moral changes

The openness humans have to the revision and correction of their moral
attitudes extends further than a concern for the opinions and plans of their
contemporaries. It is not merely that we are committed to allowing our
views to be corrected by the current opinions of other human beings. We
are open to the possibility of revisions in our communal moral outlook
which go beyond making them maximally consistent. We admit that we
could come to recognize that the prevailing consensus is wrong, and in
consequence support more radical changes. For the objectivist and the
quasi-realist, the epistemology of evaluating proposals for such radical
change is the same. Both can appeal to the re

ﬂective equilibrium process

described in 1.3.2. But, as we shall see, serious ontological problems arise
for the quasi-realist that have no parallel for the objectivist.

Consider the following three judgements:

(16)

Inter-racial marriage is wrong.

(17)

Homosexual relationships are wrong.

(18)

Killing animals to feed human beings is morally acceptable.

They exemplify signi

ﬁcant changes in sensibility that have either happened

in the last century, or may take place in the current one. A hundred years
ago, many otherwise morally sensitive human beings would have assented
to (16). Had they been told that the position would be widely rejected in
2013, they would probably have judged this to be due to our moral
degeneracy, not theirs. (17) is more contentious

—although outside of

religious circles, there has been a fairly sharp decline in the number of
‘otherwise morally sensitive human beings’ with substantive objections to
homosexual practice. Many animal rights advocates argue that in due
course (18) will seem as obviously wrong as (16) and (17)

—these changes

all being part of an

‘expanding circle’ of imagination and concern which

has yet to be completed.

ALTERNATIVES TO REALISM

It is not clear what account the quasi-realist can give of an agent who is

unsure as to whether to assent to (17) or (18). When such an agent weighs
up the case for (17), she is trying to work out which one is right. While

(19)

She has as yet no clear view on whether homosexual relationships
are wrong,

will describe her position, this is not equivalent to

(20)

Her view is this: homosexual relationships are neither right nor
wrong.

As we saw in Chapter 1, there is an

‘about-ness’ to such deliberation here

that seems quite different from deliberation about whether something
tastes nice or is entertaining. In the latter kinds of deliberation, once an
agent

’s sensibilities reach a new equilibrium there is no further question

about whether they are right. The claim that I could be

‘wrong’ about

whether something tastes nice is merely shorthand for the claim that I have
a different preference from the majority of people, or perhaps from those
considered sophisticated. Of course, I might greatly care about acquiring
tastes that matched those of people I considered sophisticated, but then
that would provide me with a deliberative standard (that is to say, it would
give me something to try and get right). In such a case, I would be seeking to
forecast the judgements of that group, rather than simply to discern my
own culinary preference. Parallel remarks might be made about delibera-
tion on whether something is

‘really’ funny, a case I considered in more

detail in Chapter 1. By contrast, when my moral sensibilities reach a new
equilibrium, it is still meaningful to ask whether they are right. If it were
not meaningful to ask whether our sensibilities were right, we would face
the very predicament moral quasi-realism was designed to avoid. For then
moral truth would be dependent on our changing sensibilities, and hence
we would be committed to appalling counter-factuals such as (4).

Moral quasi-realism faces a dilemma. On the one hand, a key part of its

attraction is supposed to be its avoidance of unattractive counter-factuals.
The

‘realist-seeming grammar’ of ethics is such that I do not suppose that

whatever answer I come to on (16), (17), or (18) thereby becomes the right
answer. Rather, the very fact of my deliberation about (18) suggests
I believe there is, independently of my opinions, already a right answer.
On the other hand, if Blackburn concedes that

‘the right answer’ is something

FROM MORALITY TO METAPHYSICS

in principle distinct from

‘whatever the answer I come to’ it is then difﬁcult

to see how the notion of an

‘independent order of reason’ is to be avoided.

When Blackburn writes that

If the relativist is saying that it is just our opinion as opposed to a good or better
opinion that approves of slavery, we are in moral dispute, certainly. But it is then
up to him to locate this good or better opinion and defend it, and this we will
know he is unable to do, for we are certain that slavery is a bad system.

It is unclear how we

‘locate . . . and defend’ a ‘better opinion’. How does the

quasi-realist make sense of the rational revisability of moral views? What is it
for one opinion to be

‘better’ than another? Is it just that it is the opinion ‘we’

eventually come to hold? What would it be for

‘us’ to make a mistake? Surely

it cannot be that a radical change in our opinions is always a mistake

— for that

would make moral progress impossible, and all change from our current
views would, by de

ﬁnition, be movement away from the moral truth.

It is built into the

‘realist-seeming grammar’ of ethics that, while we take

our current commitments, when considered singly, to each be correct

—

for otherwise they would not be our commitments

—we also think that,

considered as a whole, they may contain signi

ﬁcant errors. This is what

creates the possibility of future emancipatory changes which parallel the
widespread rejection of (16) in the past century. It is surely reasonable to
think both that our current moral opinions may have oppressive blind
spots and to be unsure what those blind spots are.

3.3.4 A possible response

At this point, Blackburn might feel entitled to deploy some of the argu-
ments I made at 1.3.4. In defence of objectivism, I argued that the scope of
moral disagreement might be narrower than it initially seemed. I suggested
that, as communities come to a deeper understanding of one another

’s

identity and culture they generally

ﬁnd it harder to deny the other’s full

humanity. I also claimed that

‘dehumanizing’ another community usually

involves empirical as well as moral falsehoods. Can Blackburn not claim
that it is this very process by which signi

ﬁcant moral changes such as (16),

and perhaps also (17), come about?

Up to a point, such a response holds water. A moral quasi-realist has no

more epistemological dif

ﬁculty than an objectivist with the cases under

discussion. Whatever meta-ethical theory one holds, there is a problem in
resolving moral disagreements when both sides can present a coherent and

ALTERNATIVES TO REALISM

consistent argument. However, the challenge to quasi-realism is not purely
epistemological. It is ontological. The question is not just how agents might
decide between competing views. It is also what we take agents to be doing
when they consider radical (and purportedly emancipatory) changes to the
consensus in their culture.

Any adequate account of radical moral change must both allow for the

possibility of such change being warranted (and so not tie moral truth to
‘what we currently think’) and allow for the possibility that such change is
mistaken (and so not collapse moral truth into

‘what we come to think’).

Moral objectivism clearly meets both of these criteria. The fact that

moral truth is independent of our sentiments and attitudes would explain
both why we would consider the possibility of radical moral changes and
would allow for the possibility that change might be mistaken, thus
avoiding counter-factuals such as (4).

By contrast, moral quasi-realism faces a dilemma. Blackburn

’s strategy

for avoiding (4) was to root moral truth in our current convictions. Black-
burn wants to af

ﬁrm that

(6)

Whatever we felt about torturing the child, it would be an evil act,

and he secures this by pointing out that (6) is a statement made with our
current attitudes. If Blackburn sticks to this line rigidly, so that

‘moral truth’

ﬁxed by the maximally consistent extrapolation of our current opinions,

then emancipatory moral change is ruled out. By contrast, if Blackburn ties
moral truth, not to a maximally consistent extrapolation of our current
opinions, but to whatever the re

ﬂective equilibrium process might yield at

some future date, he allows back in counter-factuals such as (4).

3.3.5 Indexicality of moral quasi-realism

It would seem that quasi-realism is unable to avoid making moral judge-
ments covertly indexical. Our loyalty to the moral views we have is not in
the end because they are better than any others

—but because they are

ours. More precisely, the very notion of

‘better’ turns out to be indexical.

A similar set of issues arise for Frank Jackson

’s moral functionalism. Like Blackburn,

Jackson wants to avoid the metaphysical commitments of objectivism, and seeks to construct
morality from a

‘fully mature’ version of our existing ‘folk theories’. See his From Metaphysics

to Ethics: A Defence of Conceptual Analysis (Oxford: Oxford University Press, 2000) and
Stephen Yablo

’s review, ‘Red, Bitter, Best’, Philosophical Books 41 (2000):13–23.

FROM MORALITY TO METAPHYSICS

What it is for an attitude to be

‘better’ is nothing more than for it to be

congruent with a maximally consistent version of our current views.
Indeed, we can go further: the only reason that even my contemporaries
have a role in shaping what it is for an attitude to be

‘better’ is I already

have a pro-attitude to respecting their attitudes.

Any account which makes moral judgements indexical will fall foul of

the argument from deliberative indispensability which I advanced in
Chapter 1. For, as I argued there, when we are deliberating on the
stance we should take to take towards our existing attitudes, and how
radically we should consider altering them, we need to understand our-
selves as trying to get something right. Precisely because, in the cases I am
now considering, we are deliberating over whether to modify attitudes
which are consistent with one another and based on accurate empirical
judgements, we cannot understand what we are doing as merely rendering
of our existing attitudes more coherent.

The moral quasi-realist may reply that this is indeed so: that although

much of the

‘objective feel’ of ﬁrst-order ethics can be salvaged, the

position does indeed require some modi

ﬁcation of our pre-philosophical

views. Perhaps all we can do is make our communal attitudes maximally
consistent, in the light of the best possible empirical evidence. And
perhaps, once we have achieved this, we should not be troubled by the
possibility that they are in error.

The price of making this reply would be very high indeed. For then moral

quasi-realism would become nothing more than a sophisticated version of
Mackie

’s error theory. On such an account, there would be no ‘right’ answer

to the question of what our moral attitudes should be. In consequence,
much of our pre-philosophical practice of moral deliberation would turn
out to be otiose. Moreover, it is not only our deliberation which would be
threatened by such a development of the quasi-realist position. Many of our
moral sentiments are inextricably bound up with truth-claims that it would
then be unable to accommodate. When an agent, or a community, abhors a
practice such as slavery or torture, it is not merely that they

ﬁnd it unpleasant

or that they wish others did not do it. It appears to them that the practice is
wrong. It is not clear that we can explicate the difference between (on the
one hand) dislike of an action, and a wish that humans should not commit it
and (on the other hand) disapproval and abhorrence without reference to
some independent standard of right and wrong.

ALTERNATIVES TO REALISM

When we say

(7)

Slavery is a bad system,

or, indeed,

(8)

Our opinion is this: slavery is a bad system,

we are saying more than that we have a negative attitude towards it, and
also wish that others would desist from it. We are acknowledging that,
unlike even the tastes about which we have the strongest views and the
deepest inclination to proselytize, we believe there to be some imperative
to desist from it which is not merely a re

ﬂection of our own preferences.

Of course, it is precisely in the cases where we abhor, rather than cases
where we merely dislike something, that we take the disagreement of
others to be a compelling reason to consider revising our attitudes.

To conclude this discussion, I think that Blackburn succeeds in his

initial objective

—that of avoiding counter-factuals such as (4) and (13).

However, this success comes at a very heavy price. Moral quasi-realism
cannot make sense of the deliberation which humans engage in to deter-
mine what attitudes and norms they should adopt. For such deliberation
goes far beyond that of systematizing the attitudes and norms people have
already adopted. Humans are open to the possibility of much more radical
re

ﬂection and revision of their beliefs.

For reasons outlined in Chapter 1, I take this kind of re

ﬂection on our

attitudes to be deliberatively indispensable. Therefore, I must conclude
that Blackburn

’s theory gets the relationship between attitudes and beliefs

the wrong way round. Our pre-philosophical moral attitudes only make
sense as a response to claims we take an independent reality to have upon
us. To regard these claims as constituted by our attitudes is in fact to
contradict the content of those very attitudes.

3.4 Conclusion

Earlier in this chapter (at 3.2.2), I quoted Stig Alstrup Rasmussen

’s critique

of quasi-realism in ethics:

Stripping everything but its bare essentials, [Blackburn

’s] argument is to the effect

that it is immoral to regard moral truth as mind-dependent

. . . The rejoinder is

immediate that if so, perhaps no nice people should endorse quasi-realism, since

. . .

that doctrine does make moral truth mind-dependent.

FROM MORALITY TO METAPHYSICS

Although his argument moved too quickly, we can now see that Ras-
mussen put his

ﬁnger on the central weakness of moral quasi-realism. The

paradox facing Blackburn and Gibbard is that the very sentiments out of
which they seek to build their (limited) moral ontology presuppose
something beyond themselves. To abhor torture inextricably involves taking
its evil to consist in something other than one

’s distaste for it—or indeed,

the community

’s distaste for it. Likewise, moral reasoners take an interest

in the views of others, not merely because they have a brute preference for
open-mindedness, but because they think others have access to the same
moral reality. Our moral sentiments direct us not to take them to be the
last word on things. The reason for this is simple. Those very sentiments
take themselves to be responses to, and articulations of, a reality beyond
them.

Quasi-realism is unable to answer a crucial question. From which senti-

ments is

‘moral truth’ constructed? As we have seen, Blackburn and

Gibbard have only two options, and neither is satisfactory. Either they
must say that our current sentiments

ﬁx moral truth (whoever ‘we’ are

deemed to be) or they must allow that the truth could be different from
what we now think. The former position rules out future emancipatory
changes, whereas the latter (in the absence of any notion of a more
objective order of reason) re-opens the door to the kind of morally
obnoxious counter-factuals their position was devised to avoid.

Given the critique which has now been offered of the error theory and of

quasi-realism, two choices remain for secular moral philosophers. One
option is to hope that our fundamental moral commitments can still be
accommodated within a non-objectivist framework

—that is, to hope that

they can succeed in the task at which Blackburn and Gibbard have failed.
The other option is to develop a form of objectivism which avoids the
problems described in Chapter 2. In the chapter which follows, we will
discuss two philosophers who attempt the

ﬁrst of these tasks, before moving

on (in Chapters 5 and 6) to consider some more objectivist accounts.

ALTERNATIVES TO REALISM

Procedures and Reasons

Tim Scanlon and Christine Korsgaard

4.1 Introduction

I have long been sceptical about desire theories as an account of well-
being appropriate for moral theory, but I have supposed that there is
more to be said for them as an account appropriate for individual
decision-making. This seems to me to be a mistake, and I now
believe that desire theories should also be rejected as accounts of
well-being appropriate from the

ﬁrst-person point of view.

T. M. Scanlon, The Dif

ﬁculty of Tolerance

4.1.1 Aim of the chapter

As we saw in Chapter 3, Simon Blackburn contrasts

‘Dionysian’ and

‘Apollonian’ accounts of ethics. Against him and Gibbard, I have argued
that our fundamental moral commitments require us to reject their Dio-
nysian account. We do not

‘talk of reasons to reﬂect the fact that we

already care.

’ In fact, we care (or at least, we ought to care) because there are

reasons to care.

In this chapter we will examine the Apollonian alternative

—in the

versions advanced by Tim Scanlon and Christine Korsgaard. Scanlon is
very clear that our desires and cares are explained by reasons and not vice
versa. But, like Blackburn, he is keen to avoid any metaphysical hostages to
fortune. In What We Owe to Each Other, Scanlon seeks to accommodate
the pull of reductionism by stressing rational procedures rather than an
ontologically distinct moral reality.

By holding that what is morally right

T. M. Scanlon, The Dif

ﬁculty of Tolerance: Essays in Political Philosophy (Cambridge:

Cambridge University Press, 2003), 176.

Scanlon, What We Owe to Each Other (Cambridge, MA: Belknap Harvard, 1998).

is determined by a procedure, it looks as if he can avoid the counter-
factuals that so discom

ﬁt Blackburn.

Scanlon

’s early position therefore promises the best of both worlds:

while remaining metaphysically inoffensive, it gives moral truth suf

ﬁcient

independence from our sentiments and cultural conventions to meet the
fundamental ethical commitments defended in Chapter 1. In developing
this position, Scanlon makes an explicit contrast with Blackburn
and Gibbard. He seeks to retain the metaphysically appealing features of
their position whilst responding more effectively to objectivist concerns.
However, his view has since changed in some quite fundamental respects.
Scanlon

’s revised position is elaborated in his John Locke Lectures (given

in Trinity Term 2009). He now embraces a fuller moral objectivism

—

while still claiming to avoid any uncomfortable metaphysical implications.

Christine Korsgaard

’s ethical constructivism provided the meta-ethical

foundations for What We Owe to Each Other

—and he advances his revised

position by contrasting it with hers. In this chapter, I will explain
why I agree with Scanlon

’s change of mind—and in doing so, will

offer my own critique of both Scanlon

’s early position and Korsgaard’s

constructivism. I will argue that Scanlon

’s new position (unlike these other

two) meets the demands of Chapter 1. However, in doing so, it generates
the explanatory gap described in Chapter 2.

I will argue that Scanlon

’s change of heart, and the weaknesses of his

earlier and later positions, provide evidence for my more general claim.
The dilemma outlined in Part I (insuf

ﬁcient objectivity to underwrite our

moral commitments or generation of explanatory gap) looks unavoidable.

4.2 Procedural Realism

In What We Owe to Each Other, Scanlon offers the following account of
moral truth:

[an] act is wrong if its performance under the circumstances would be disallowed
by any set of principles for the general regulation of behaviour that no one could
reasonably reject as a basis for informed, unforced, general agreement.

Scanlon, What We Owe to Each Other, 153.

PROCEDURES AND REASONS

Conversely, the rightness of an act consists in it being one which could
sustain such an agreement. On this account, the realities at the heart of
morality are procedural rather than substantive.

Early on, Scanlon tells us that he is going to use the notion of a

‘reason’

as the most basic and abstract element of normative thought, to provide a general
characterisation of a slightly more speci

ﬁc normative notion, the idea of value.

He goes on to explain that

being valuable is not a property that provides us with reasons. Rather, to call
something valuable is to say that it has other properties that provide reasons for
behaving in certain ways with regard to it.

This is (in his words) a

‘buck-passing’ account of goodness and value. The

requirements of Chapter 1 are met because

‘good,’ ‘valuable’, and ‘evil’ are all

on this account objective, non-natural properties. However, the position
remains metaphysically inoffensive because it holds that these properties are
the

‘purely formal, higher-order properties of having some lower-order

properties that provide reasons [for action].

’

For the purposes of this section, I am going to assume that such a

‘buck-

passing

’ account is viable, although I shall raise some questions about

it later (at 4.5.1). I will restrict my argument here to showing that,
even on a

‘buck-passing’ account, it is impossible both to be sufﬁciently

objectivist to meet the requirements of Chapter 1 and to avoid generating
an explanatory gap.

Why should it should be less problematic to take reasons rather than

values as ontologically fundamental? Even on a

‘buck-passing’ account,

there will need to be a fact of the matter about whether a given set of
lower-order properties gives an agent appropriate reasons for action. If
Scanlon

’s account is to avoid the objections made in Chapter 1, then he

must surely endorse the following claim:

The judgements we make about whether a given state of affairs provides an agent
with good reason to act in a given way are themselves claims that can be true or
false, independently of human conventions or attitudes.

Scanlon writes that:

Scanlon, What We Owe to Each Other, 78.

Scanlon, What We Owe to Each Other, 96.

FROM MORALITY TO METAPHYSICS

Understood in the right way, this is quite correct: but when understood in that
way, the independence does not raise metaphysical and epistemological problems.
The example of mathematical judgements may be helpful here

. . . The thinking of

a good mathematical reasoner

‘represents’ or ‘tracks’ the truth about arithmetic

insofar as it takes into account the right considerations in the right way. This
need not be construed as a matter of being in touch, through some mechanism
analogous to sense-perception, with mathematical objects which exist apart from
me. Similarly, in order for judgements about reasons to be taken to be about some
subject matter independent of us

. . . what is necessary is for there to be standards for

arriving at conclusions about reasons.

Crucial to this response is the assumption that

‘nothing metaphysical’ is

going on when a good mathematical reasoner takes into account

‘the right

considerations.

’ Scanlon refers readers on to Christine Korsgaard’s more

detailed treatment of these issues:

My thinking about this question has been helped by Christine Korsgaard

’s discus-

sion of what she calls

‘substantive moral realism’ in The Sources of Normativity . . .

I believe that my own view

. . . amounts to what she calls procedural rather than

substantive realism.

In What We Owe to Each Other, Scanlon says no more about meta-ethical
theory. So we can only evaluate his position by way of a more detailed
discussion of the procedural realism Korsgaard presents and defends in The
Sources of Normativity.

4.3 Christine Korsgaard on Procedural Realism

Korsgaard describes of the difference between procedural realism and
substantive moral realism (that is, what I have termed

‘moral objectivism’)

in the following terms:

The procedural moral realist thinks there are correct answers to moral questions
because there are correct procedures for arriving at them [whereas]

. . . the substan-

tive moral realist thinks that there are correct procedures for answering moral

Scanlon, What We Owe to Each Other, 380 (footnote to 63).

Christine M. Korsgaard, The Sources of Normativity, ed. Onora O

’Neill, with responses by

G. A. Cohen, Raymond Geuss, Thomas Nagel, and Bernard Williams (Cambridge: Cam-
bridge University Press, 1996). Scanlon points us to the discussion on pp. 32

–7.

PROCEDURES AND REASONS

questions because there are moral truths or facts which exist independently of those
procedures, and which those procedures track.

The attraction of this position is clear: the existence of

‘correct procedures’

is supposed to give the position its objectivity, while escaping (what to her
are) unacceptable ontological commitments.

4.3.1 The categorical imperative

ﬂeshing out her position, Korsgaard offers a twofold account of moral

deliberation. Firstly, she argues that in all intentional action, my choices
commit me to some kind of more general truth about the value of that
which I pursue.

[S]ince the will is practical reason, it cannot be conceived as acting and choosing for
no reason. Since reasons are derived from principles, the free will must have a
principle

. . .

Korsgaard

’s argument is similar to that advanced in Chapter 1 of this

book

—that it is impossible to reason practically without such a deliberative

standard. However, she goes on to tell us we should not see this standard as
something external which is imposed upon us. Rather, we should regard it
as something which has issued from our own will. To engage in practical
reasoning just is freely to choose the maxims upon which we will act, and in
choosing a maxim we commit ourselves to the notion that it is (in some
sense) right. In consequence, we are committed to the notion that the
maxims we choose should also be willable as a universal law. As Korsgaard
writes:

[T]he categorical imperative [i.e. acting only according to maxims we could will to
be universal law] does not impose any external constraint on the free will

’s

activities, but simply arises from the nature of the will. It describes what a free
will must do in order to be a will.

Korsgaard thinks this account can avoid the pitfalls of both reductionism
and of objectivism. It seems to avoid the pitfalls of reductionism because it
acknowledges (in contrast to Mackie, Blackburn, and Gibbard) that the
moral law is not a mere re

ﬂection of the agent’s sentiments. There is a fact

of the matter about what maxims agents can will as universal laws. It seems

Korsgaard, The Sources of Normativity, 36.

Korsgaard, The Sources of Normativity, 98.

FROM MORALITY TO METAPHYSICS

to avoid the pitfalls of objectivism because in speaking of such

‘facts of the

matter

’ there is no commitment to any realm of ‘moral facts’, access to

which might be problematic. We can determine through our faculties of
deductive reasoning what maxims can be willed consistently as universal
laws. Therefore, it would seem, there is no explanatory gap.

How far does this take us towards a substantive set of moral truths? On

Korsgaard

’s own estimation, it only gets us part of the way. After all, an

agent could choose as his maxim for action

‘the law of acting on the desire

of the moment

’. In such a case, his conduct would be ‘that of a wanton’—

but as yet nothing has been said which rules this out as a maxim which
could be chosen by him in practical deliberation.

While some courses of

action may be ruled out by the categorical imperative, it will not yield all
the substantive moral principles which humans need in order to deliberate.

4.3.2 Self-identity

The second part of Korsgaard

’s account is supposed to yield these substan-

tive principles. Her claim is that the substance of the maxims we will
choose

ﬂows from our sense of identity:

When you deliberate, it is as if there were something over and against all your
desires, something which is you, and which chooses which desire to act on. This
means that the principle or law by which you determine your actions is one that
you regard as being expressive of yourself.

Korsgaard goes on to give examples of such principles: being someone

’s

friend or lover, or a member of a family or nation, or

‘a Citizen of the

Kingdom of Ends

’. On her account, when you deliberate concerning your

identity, you are not engaged in a theoretical or factual enquiry, but are
seeking

‘a description under which you value yourself, a description under

which you

ﬁnd your life to be worth living and your actions to be worth

undertaking.

’ The conceptions of ourselves which are most important to

us give rise to the obligations which are

‘unconditional’. When we violate

these, we lose our integrity and our identity

—for it is the ability to live up

to standards and commitments, and not merely respond to the urges of
the moment, that

‘makes [someone] a person at all’. In violating your self-

authored maxims, you are

Korsgaard, The Sources of Normativity, 99.

PROCEDURES AND REASONS

no longer able to think of yourself under the description under which you value
yourself and

ﬁnd your life to be worth living and your actions to be worth

undertaking. It is to be for all practical purposes dead or worse than dead.

4.4 Critique of Korsgaard

’s Account

Although Korsgaard

’s account differs from Blackburn’s in some key respects,

they share a fundamental weakness. Both philosophers locate moral value in
a feature of the agent

’s attitudes. For Blackburn, moral value is located in the

agent

’s sentiments—while for Korsgaard it is located in the agent’s self-

conception. I will argue that the agent

’s self-conceptions (like her senti-

ments) only make sense as responses to an external order of value.

4.4.1 Possibility of immoral self-identity

The role of the categorical imperative in Korsgaard

’s account is to place

limits on what agents can reasonably will and do. She agrees that a wide
ﬁeld of incompatible potential maxims remains. How do we choose
between the different candidates for universalizable moral norms? As
we have seen, Korsgaard

’s (un-Kantian) answer is that this is a matter of

the agent

’s existential commitments.

Does this mean that any existential commitments would generate

equally valid obligations? Let us suppose (following G. A. Cohen) that
I am a Ma

ﬁoso, and that it is central to my self-identity that I am strong

in obeying my organization

’s honour code. Would that generate an

‘obligation’ to do so?

Korsgaard replies that in one sense it would, for

‘there is a real sense in which you are bound by a law you make
for yourself until you make another

’. However, she believes that

the Ma

ﬁoso should, given sufﬁcient reﬂection, come to see that obedience

to the honour code is the wrong law to make for himself:

What I would like to claim about a person

’s relation to an immoral form of self-

identi

ﬁcation . . . [is] that there is no coherent point of view from which it can be

endorsed in the full light of re

ﬂection. If Cohen’s Maﬁoso attempted to answer the

Korsgaard, The Sources of Normativity, 102.

Thomas Nagel uses this term in his response to Korsgaard, Nagel,

‘Universality and the

ﬂective Sense’, in Christine M. Korsgaard (ed.), The Sources of Normativity, 203, and she

endorses it in her

‘Reply’, 237.

G. A. Cohen,

‘Reason, Humanity and the Moral Law’, in Korsgaard (ed.), The Sources

of Normativity, 183

–4.

FROM MORALITY TO METAPHYSICS

question why it matters that he should be strong and in his sense honour-bound
even when he was tempted not to, he would

ﬁnd that its mattering depended on

the value of his humanity, and if my other arguments go through, he would

ﬁnd

that this commits him to the value of humanity in general, and so to giving up his
role as a Ma

ﬁoso.

Korsgaard

’s view is that, ontologically speaking, morality is underwritten

by (i) the logical relations between the maxims an agent can consistently
will for himself (which provide us with the framework of the

‘categorical

imperative

’) and (ii) facts about the self-conceptions with which human

beings can live (which provide us with substantive moral attitudes). For
Korsgaard, there is no

‘objective moral order’ which makes it true that the

ﬁoso’s honour code is inferior to the value system of the ‘Citizen of the

Kingdom of Ends

’ (who values all human beings, and indeed animals).

ﬂesh out (ii), Korsgaard offers the following argument. It is meant to

show that the contingent, partial obligations and affections of the Ma

ﬁoso

or of an agent who initially cares only for those of within his own tribe
ultimately only make sense within a framework which ascribes value to all
human beings:

We endorse or reject our impulses by determining whether they are consistent
with the ways in which the ways in which we identify ourselves. Yet most of the
self-conceptions which govern us are contingent

. . . Because these conceptions

are contingent, one or another of them may be shed

. . .

What is not contingent is that you must be governed by some conception of your

practical identity. For unless you are committed to some conception of your
practical identity, you will lose your grip on yourself as having any reason to do
one thing rather than another

—and with it your grip on yourself as having any

reason to live and act at all. But this reason for conforming to your particular
practical identities is not a reason that springs from one of these particular practical
identities. It is a reason that springs from your humanity itself, from your identity
simply as a human being, a re

ﬂective animal who needs reasons to act and to live.

And so it is a reason only if you treat your humanity as a practical, normative, form
of identity, that is, if you value yourself as a human being.

Once again, there are some parallels between the argument which
Korsgaard makes and the argument for deliberative indispensability

Korsgaard,

‘Reply’, in her The Sources of Normativity, 256–7.

Cf. Korsgaard

’s discussion of the moral claims of animals in her The Sources of Normativity,

Chapter 4.

Korsgaard, The Sources of Normativity, 120

–1.

PROCEDURES AND REASONS

advanced in Chapter 1. There is one crucial difference: namely, Kors-
gaard

’s account of the ontological basis for arguments against the Maﬁoso’s

moral stance. Korsgaard claims that, for all human beings, it is their identity
as humans which is ultimately

‘essential’. Those valuations which ﬂow from

more parochial self-conceptions turn out, on closer examination, to lead
on by an inexorable logic to the valuing of humanity in general. This
is because the more parochial identities are contingent and can be shed,
whereas one

’s identiﬁcation as a human is not contingent and cannot

be shed.

I will argue below (at 4.4.3) that agents

’ valuations only have the wider

implications Korsgaard

’s argument requires if they are understood as

responses to an objective order of value. If the Ma

ﬁoso’s valuations are

not incipiently objectivist, it is hard to see why they should commit him to
any less parochial valuations. The Ma

ﬁoso’s need for ‘some conception

of his practical identity

’ can surely be secured by his stable commitment to

ﬁa values. The fact that humans can slough off different contingent

identities does not undermine the stability of the Ma

ﬁoso’s own identity,

provided he has made a settled decision to identify with this particular set of
attitudes and commitments. Korsgaard may disapprove of his existential
choice, but it is hard to see why (on her account) the Ma

ﬁoso’s settled

choice threatens his

‘grip on [himself] as having any reason to do one thing

rather than another, and with it [his] grip on [himself] as having any reason
to live and act at all

’.

By contrast, the objectivist has rather more to say to the Ma

ﬁoso.

Objectivism does not generate a deductively valid argument from the
Ma

ﬁoso’s existing commitments to a wider moral concern. However,

the objectivist can advance arguments which will put the Ma

ﬁoso’s

position under some pressure. If (as I argued in Chapter 3) our moral
sentiments only make sense as responses to an objective order, then the
Ma

ﬁoso can be shown that his valuations make implicit claims about

objective goodness. It can then be pointed out that he has no reason to
suppose that the objective moral order uniquely favours him or his
associates. His valuation of them does then provide a reason for valuing
others: he has no non-indexical ground for thinking his associates uniquely
worthy of respect. Ironically, the very feature of our pre-philosophical
moral commitments that Korsgaard denies (namely their incipient objec-
tivism) turns out to be the only one with the traction to argue the Ma

ﬁoso

from his contingent valuations to a positive valuation of all human beings.

FROM MORALITY TO METAPHYSICS

4.4.2 Deliberative indispensability and moral ontology

Korsgaard offers a response to this argument: even if objectivism is
indispensable in our dialogue with the Ma

ﬁosi, she thinks this would

provide a dubious basis for its acceptance. We cannot simply postulate
an entity because our argument suggests it is necessary:

Having discovered that obligation cannot exist unless there are actions which it is
necessary to do, the realist concludes that there are such actions, and that they are
the very ones we have always thought were necessary, the traditional moral
duties

. . . The realist . . . places the necessity where he wanted to ﬁnd it. And that

argument cannot even get started, unless you assume that there are some actions
which it is necessary to do.

The

ﬂaw in Korsgaard’s argument here is revealed by her ﬁnal sentence—

‘the argument cannot even get started, unless you assume that there are
some actions which it is necessary to do.

’ For her position and mine both

assert that all decent human beings are committed to the notion that
some actions are necessary. The circularity in the objectivist

’s argument

is virtuous and not vicious, because the assumption that there are moral
obligations has already been shown to be deliberatively indispensable. As
we saw at 1.3.1, the inference from deliberative indispensability to truth
can be a legitimate one.

The question, then, is not whether the moral objectivist is entitled to

make the move he does, but whether Korsgaard can get by without making
the same move. I will argue that we cannot avoid drawing the inference
from our fundamental moral commitments to an objective order of reason.
Without that inference, morality becomes unacceptably self-referential.

4.4.3 Priority of value over self-conception

Korsgaard repeatedly argues that if an agent makes a certain choice, he
implicitly values certain things, and this in turn commits him to valuing the
choices made by others. But she gives no account of what it is to value
something. For the reasons I advanced against Blackburn and Gibbard,
what it is to value something cannot be elucidated in terms of liking it or
having a pro-attitude towards it. Rather, to value something is to ascribe a
property to it.

Korsgaard, The Sources of Normativity, 33

–4. Cf. David Hume, The Dialogues Concerning

Natural Religion, part IX.

PROCEDURES AND REASONS

Korsgaard

’s arguments seem parasitic on the unspoken notion of value

as a property of the thing valued. According to Korsgaard, when I choose
something, I cannot without inconsistency deny (i) that it has value, (ii)
that I have value, and (iii) that the choices of other human beings have
value. What does it mean to say that I

‘must accept’ that the thing I choose

‘has value’ unless there is more (ontologically speaking) to its value than
the bare fact that I choose it? I can

‘like’ something without committing

myself to the notion that it is (and hence the things others like are)
‘likeable’. Why should the word ‘valuable’ have a different grammar?
Korsgaard seems to be making an illicit move from verbs to adjectives.
In choosing something, perhaps I commit myself to

‘valuing’ it, but it is

not obvious that this commits me to the notion that it is

‘valuable,’ let

alone that anything else is.

If valuing was not a response to a property in the thing or action chosen,

but merely an expression of my identity, morality would also become self-
referential, and therefore intolerably narcissistic. It is one thing (and a
perfectly fair thing) to remind me that I ought to do the right thing
because otherwise I would violate my deepest commitments, and in a
sense become a different person. But the force of that reminder lies in the
fact that I take those commitments to be correct. They are not tastes or
lifestyle choices. I am committed to them precisely because I think them
worth committing to. Once again, it seems that the features of agency and
choice on which Korsgaard wishes to found her account of value are
themselves parasitic on prior judgements of value.

Thomas Nagel advances this same argument, by way of the following

thought-example:

If someone accepts death rather than betraying a number of other people to the
killers, it might be unappreciative to explain this in terms of the conception he had
of himself. Of course if he cares about the survival of others, and is unwilling to save
his own life by betraying those others, then that is an important fact about his
conception of himself. But to explain the grip on him of those reasons would be to
get things backwards, and incidentally to cheapen the motive

. . . Even if he can get

motivational help from thinking that he couldn

’t live with himself if he saved his

life by this method, that is not the

ﬁnal explanation—indeed, it couldn’t be. The

real explanation is the non-

ﬁrst-personal reason against the betrayal.

Thomas Nagel,

‘Universality and the Reﬂective Self ’, in Korsgaard, The Sources of

Normativity, 206.

100

FROM MORALITY TO METAPHYSICS

The non-

ﬁrst-personal reason that counts against the betrayal is the intrin-

sic value of human beings. What it is to value other human beings is surely
not to be construed as saying

‘I couldn’t live without valuing them’? That

seems to get things the wrong way round. Korsgaard responds to Nagel

’s

thought-example by conceding that his criticism is accurate:

I think it be

ﬁts an adult rational being to question the necessity of extreme sacriﬁce, and

it is to be hoped that if she does not it is because she already has a grip on the answer.
But the answer certainly is, on my view, that she is unprepared to give up a certain way
she looks at herself, a description under which she

ﬁnds her life worth living and her

action worth undertaking. And this is what bothers Nagel

—the idea that, in this way,

normativity derives from our self-conceptions. And I do think this.

I argued at 3.3.5 that to say

‘slavery is abhorrent’ is not in the end to make an

assertion about my attitudes, but rather to assert something about slavery which
I take to call forth my abhorrence. The argument against Korsgaard parallels
this: to say

‘slavery is abhorrent’ is not to make an assertion about the commit-

ments essential to my self-conception, but to assert something about the value
of other persons which makes certain attitudes essential to that self-conception.

4.5 Scanlon on

‘Reasons Fundamentalism’

Since writing What We Owe to Each Other, Scanlon has moved towards a
more wholeheartedly objectivist position. His 2009 John Locke Lectures
explicitly repudiate Korsgaard

’s attempt to ground normative reasons

in facts about rational agency and self-identity. Scanlon now describes
himself as a

‘Reasons Fundamentalist’, contrasting the position with

Korsgaard

’s own as follows:

According to a Reasons Fundamentalist, the relation that holds between an agent
and a consideration X [in a situation where X is a reason for the agent to do A] just
is the relation of X

’s being a reason for the agent to do A. The ‘grip’ that X has on the

agent just is this relation: being a reason for him or her (or, in the strongest case, a
decisive reason). As Korsgaard puts it, quite correctly, a Reasons Fundamentalist
‘insists on the irreducible character of normativity.’ The fundamental disagreement
is whether there is some further explanation that can and should be given of why
the agent on this situation must treat X as a reason.

Korsgaard,

‘Replies’, in her The Sources of Normativity, 249.

Scanlon, Being Realistic About Reasons, Lecture I, 15. The quotation from Korsgaard is

from her The Sources of Normativity, 32.

PROCEDURES AND REASONS

101

Korsgaard rejects

‘Reasons Fundamentalism’ because it simply postulates

objective moral realities. As we have seen, she takes this to be an illicit
move. Korsgaard also claims that, if moral norms are objective (in the way
Scanlon and I both take them to be) they will not be able to

‘get a grip’ on

the agent. This objection echoes Mackie

’s ‘argument from queerness’.

Korsgaard and Mackie assume that anything which will motivate an
agent must make reference to existing desires and commitments of the
agent. My contention is that both philosophers have got this back to front.
It is only possible to make sense of practical deliberation on the assumption
that, in re

ﬂecting on how to act, agents are trying to get something right.

That is to say, they are trying to discover what is truly valuable, and thus
what reasons there are for action.

4.5.1 Scanlon on reason and desire

Even though he was not fully objectivist at that point, Scanlon

’s discussion

of reason and desire in What We Owe to Each Other helps us to see why
reasons are logically prior. In doing so, it provides further grounds for
dismissing Korsgaard

’s worry about the way in which objective reasons

‘get a grip’ on agents.

Scanlon

’s account has two components. Firstly, he claims that some

objective notion of

‘reason’ is written into our understanding of a desire:

‘Having what is generally called a desire involves having a tendency to
see something as a reason.

’ Secondly, he advances an account of ‘attention-

directed desire

’ to explain how reasons inﬂuence action.

In defence of the

ﬁrst claim, Scanlon invites us to consider an agent who

consumes a cool drink because of their thirst. This scenario obviously
involves a sensation (dryness in the throat) and a belief (that a more
pleasant state will be secured by having a drink). (It also involves the
agent in question taking the belief to be a good reason for having a
drink.) Does this not leave out the crucial feature

—namely the urge to

drink? Scanlon thinks not:

when we focus on this idea of a mere urge to act, separated from any evaluative
element, it does not in fact

ﬁt in very well with what we ordinarily mean by desire.

He goes on to offer a thought-example (borrowed from Warren Quinn

’s

excellent essay on

‘Putting Rationality in its Place’). Scanlon tells us:

Scanlon, Being Realistic About Reasons, Lecture IV, 25.

102

FROM MORALITY TO METAPHYSICS

[W]e may consider

. . . a man who feels an urge to turn on every radio he sees. It is

not that he sees anything good about radios

’ being turned on; he does not want to

hear music or news or even just to avoid silence; he is simply moved to turn on any
radio that he sees to be off

. . . [S]uch a purely functional state fails to capture

something essential in the most common cases of desire: desiring something
involves having a tendency to see something good or desirable about it.

Quinn

’s example provides a powerful further argument against the

‘Dionysian’ account of morality. His point is that bare wanting is not
enough to make something qualify as a desire, let alone as a basis for
intelligible action: It is only because we

ﬁnd the object of the wanting an

intelligible object of desire that the action quali

ﬁes as intelligible action.

If this is correct, and so desiring and valuing only make sense against a

prior backdrop of reasons, how do such reasons gain traction on our
motivation? To answer this, we must turn to the second part of Scanlon

’s

account, his conception of

‘attention-directed desire’. Neither Scanlon

nor Quinn wants to equate a desire for X with the judgement that X is
some way valuable. As Scanlon explains:

Having a desire to do something (such as to drink a glass of water) is not just a
matter of seeing something good about it. I might see something good about
drinking a glass of foul-tasting medicine, but would not therefore be said to have a
desire to do so, and I can even see that something would be pleasant without, in the
normal sense, feeling a desire to do it.

Clearly, there must be an additional phenomenological component in
desiring:

A person has a desire in the directed-attention sense that P if the thought of P keeps
occurring to him or her in a favourable light, that is to say, if the person

’s attention

is directed insistently toward considerations that present themselves as counting in
favour of P.

Desire, then, has an element of judgement. What distinguishes it from a
more straightforward belief is that the judgement is presented to us
through our sensibilities.

Scanlon, What We Owe to Each Other, 38

–9. The example is drawn from Warren

Quinn. See Quinn

’s ‘Putting Rationality in Its Place’, in Rosalind Hursthouse, Gavin

Lawrence, and Warren Quinn (eds), Virtues and Reasons: Philippa Foot and Moral Theory
(Oxford: Clarendon Press, 1996), 181

–208.

Scanlon, What We Owe to Each Other, 39.

PROCEDURES AND REASONS

103

Scanlon does not develop this line of thought any further. However,

there is much more to be said about the

‘favourable light’ in which P

presents itself

—and indeed about the cases of aversion, in which P presents

itself to us in an unfavourable light. Our experience indicates this may have
a variety of emotional tones. The

‘light’ cast by reverence, eros, compas-

sion, and sympathy are all different, and each can present to us speci

ﬁc

kinds of consideration which count in favour of an action. Likewise, the
‘light’ cast by abhorrence, fear, and disgust present different considerations
which count against certain actions.

The need for a more detailed account of the role of sensibility in

judgement is reinforced by an example Scanlon gives in What We Owe
to Each Other. In this passage, Scanlon is making the (I think correct) point
that we not only need to know what to value, but that we also need to
know how to value it. He writes that

Having recordings of Beethoven

’s late quartets played in the elevators, hallways,

and restrooms of an of

ﬁce building, for example, would show a failure to under-

stand the value of music of this kind. What I am suggesting is not that this would
show a lack of respect for this music, but rather that it shows a lack of understanding
of what one should expect from it, and in what way it is worth attending to.
The question of what music, if any, to play in such a setting may not be a weighty
one. But it illustrates a point of more general importance: that understanding the
value of something often involves not merely knowing that it is valuable or how
valuable it is, but also how it is to be valued.

This seems right, and although the example is an aesthetic one, similar
examples clearly occur when we come to consider the value of human
beings. Human sexuality is a case in point. Those who argue against a
sexual ethic grounded entirely in utilitarian notions of

‘avoiding harm’

might well say that certain bodily practices show a lack of understanding
of why other persons are worth attending to, and reduce them to
mere objects.

It can only be through our affective engagement with Beethoven

’s

quartets, and our apprehension of them as beautiful, haunting, and deep,
that we come to know that it would be inappropriate to play them as the
background music in a public toilet. Thus, our normative responses seem
to make most sense when they are taken to be responses to a value that
inheres in persons and objects. The explanation of why certain kinds

Scanlon, What We Owe to Each Other, 100.

104

FROM MORALITY TO METAPHYSICS

of response are or are not morally appropriate is derivative from the
distinctive kinds of value possessed by music, persons, and nature.

It is hard to see how this insight is reconcilable with a

‘buck-passing’

account of value. It seems more plausible to say that the reasons the music
gives us for treating it with respect (e.g. not playing it in public toilets)

ﬂow

from its intrinsic value (as beautiful, haunting, and deep).

I will not dwell

on this issue at length, for the more fundamental objection I want to press
(at 4.6) against Scanlon

’s new position will still stand even if he can answer

these doubts about his

‘buck-passing’ account of value.

4.5.2 Reason and emotion

The suggestion that affective responses can disclose to us reasons for action
(and, perhaps also, the value which grounds those reasons) is explored in
a growing literature on philosophy and the emotions

—most notably,

Martha Nussbaum

’s recent work.

We might express the arguments of

this developing school of thought by distinguishing between two senses
of the word

‘reason’. These are:

Reason

a ground for a belief or action that is the result of ratiocination,

independently of any affective engagement

Reason

a genuine ground for a belief or action

Reasons

are things an agent could know without any affective engage-

ment with the subject-matter: to use Aneurin Bevan

’s famous phrase, a

human being who resembled a

‘desiccated calculating machine’ could

discern all reasons

for action.

But it is not obvious

—and certainly not

an analytic truth

—that all grounds for action are accessible without affec-

tive engagement. There may be genuine grounds for action which are not

This is reinforced by Roger Crisp

’s argument in ‘Value, Reasons and the Structure of

Justi

ﬁcation: How to Stop Passing the Buck’, Analysis 65(285) (2005): 80–5.

For a survey of other objections to the buck-passing account, see Jussi Suikkanen,

‘Buck-Passing Accounts of Value’, Philosophy Compass 4(5) (2009): 768–79.

Martha Nussbaum, Love

’s Knowledge: Essays on Philosophy and Literature (Oxford: Oxford

University Press, 1992) and Upheavals of Thought: The Intelligence of the Emotions (Cambridge:
Cambridge University Press, 2001).

Cited in Geoffrey Foote, The Labour Party

’s Political Thought: A History (London:

Routledge, 1986), 273.

PROCEDURES AND REASONS

105

known by ratiocination alone. In other words, there may be reasons

which are not reasons

If we can distinguish between two senses of

‘reason’, we can also distin-

guish two corresponding senses of the word

‘subjectivism’. These are:

Subjectivism

Reasons

do not provide the only grounds for belief or

action. Some such grounds are provided by the engagement of our
sentiments.

and

Subjectivism

There are no objective reasons

for action, merely subjective

attitudes and preferences.

Subjectivism

takes affective engagement to be a mode of cognition. On

this account, sentiments can disclose objective grounds (reasons

) for

action. By contrast, subjectivism

denies that there are objective grounds

any (reasons

) for action. On this account, sentiments, attitudes, and

preferences are all that there is to the ontology of ethics.

Although he does not spell this out, Scanlon relies on such a distinction

in the fourth of his John Locke Lectures. He tells us that:

Substantive conclusions about reasons that are not formally based seem

. . . ‘subjec-

tive

’ only if they are assumed to be isolated individual responses, like the occur-

rences of a desire. But the judgements about reasons that survive the kind of
equilibrium process I have described are not like this. They, too, have undergone
careful re

ﬂection and reexamination. Perhaps it will be said that the process of

ﬂection through which we arrive at an overall view of reasons for action is not an

intellectual process in the relevant sense. But this seems to me a mere prejudice.

Scanlon

’s point might equally well be expressed by saying that the accep-

tance of subjectivism

as an epistemological theory does not in any

way commit us to subjectivism

, and indeed is fully compatible with his

new-found moral objectivism. The germs of such an account are to be
found in the early Scanlon

’s example of the ‘reasons’ for not playing

Beethoven in lifts and public toilets. These reasons are reasons

that are

not reasons

—that is to say, they are grounds for action but not ones we

could know by ratiocination alone. Therefore, moral epistemology must
rely on sentiments and sensibility as well as ratiocination.

Scanlon, Being Realistic About Reasons, Lecture IV, 26.

106

FROM MORALITY TO METAPHYSICS

Philosophers such as Nussbaum provide a more developed account of

the ways in which the sentiments can provide us with reasons

for action.

In her recent work on the subject, Nussbaum writes of an

‘affective

equilibrium

’ being reached by agents as they develop a more and nuanced

affective response to events, in dialogue with both rational arguments
and the educated sensibilities of other agents.

This account is entirely

compatible with the theory of re

ﬂective equilibrium Scanlon has

been developing. Indeed, Scanlon

’s account needs to develop in just

such a direction if his rather sketchy remarks are to be worked up into a
comprehensive theory.

Such development is a vital task for moral philosophy. Although I have

explained why I take Blackburn

’s moral theory to be ﬂawed, I think its

continuing appeal comes in part from the thing it gets right

—namely, the

crucial role of the sentiments in moral reasoning. The account Scanlon has
begun to develop does more than merely make room for Blackburn

’s

insight (which in itself helps to defuse the appeal of the quasi-realist
position). It goes further, showing that Blackburn

’s subjectivism

offers

us too shallow an account of what an affective response must be. To make
sense of a response like abhorrence, we need to understand it as something
which embodies a judgement

—that is, as a nascent truth-claim. Thus, an

appropriate subjectivism

is incompatible with subjectivism

. By contrast,

subjectivism

is not merely compatible with moral objectivism. It requires

such objectivism to make it complete.

4.6 The Explanatory Gap Re-opens

Scanlon

’s new position is completely in harmony with the secular moral

objectivism described in Chapter 2. His account may be

‘buck-passing’ (in

that he thinks facts about value reduce to facts about what we have reason
to think, feel, and do) but he holds that our reasons cannot be encapsulated
in procedures statable in advance of our singular moral judgements.
Scanlon suggests we come to know the reasons for action through the
re

ﬂective equilibrium process sketched out in Chapter 1. He combines this

with a very robust objectivism:

Nussbaum, Love

’s Knowledge, 173, 182–3.

PROCEDURES AND REASONS

107

The Method of Re

ﬂective Equilibrium is not a constructivist procedure which

constitutes the subject at hand. Even if re

ﬂective equilibrium is the method we

should use to describe what sets there are or what reasons we have, this procedure
itself is not an account of how the domain of

. . . reasons is constructed.

It is by this method of re

ﬂective equilibrium that human beings apprehend

this domain of objective reasons. The domain, however, is not only
independent of the method but is more generally independent of rational
agents:

The kind of independence of us that normative facts are supposed to have, and do
have on my view, comes to at least this: they are facts that we, individually, could
be mistaken about. Could we all, collectively, be mistaken about certain reasons?
We might all fail to notice some relevant feature of a situation, and all die before
learning any better. If this could happen through descriptive error or lack of
information, why not through normative error as well?

With its combination of objectivism and the method of re

ﬂective equilib-

rium, Scanlon

’s account is vulnerable to the following objection: What

grounds are there for holding that the results of the method of re

ﬂective

equilibrium (whether or not it includes our affective responses) tracks an
objective truth? Why should we suppose the two to be aligned?

4.6.1 Scanlon

’s response to objections

In Chapter 2, I listed three questions which can legitimately be asked
about all human cognitive capacities:

(D1)

What is the justi

ﬁcation for our faith in their reliability?

(D2)

What is the historical explanation of their development?

and

(D3)

What is the explanation for their capacity for tracking truth?

Scanlon

’s John Locke Lectures provide a convincing response to the (D1);

indeed, this is the primary purpose of the series. What, then of (D3)

—the

request for an explanation of this reliability?

Scanlon addresses something close to (D3) in his fourth lecture, and

repeats the analogy (made in What We Owe to Each Other) between ethical

Scanlon, Being Realistic About Reasons, Lecture V, 9.

Scanlon, Being Realistic About Reasons, Lecture V, 14.

108

FROM MORALITY TO METAPHYSICS

and mathematical knowledge. He now explores it at greater length,
writing that

The fact that mathematical facts have no spatial location may be taken as a ground
for thinking that there is, after all, a special problem about mathematical knowl-
edge. For if these objects exist

‘outside of space and time’ the problem of explaining

how information could get from them to us seems even greater than in the case of
empirical truths.

As I argued against Dworkin (at 2.3), (D3) does not require there to be
some kind of quasi-physical contact between mathematical or moral truths
and human beings

’ cognitive capacities. Consequently, I can concede for

the purposes of this argument that

‘the idea of a region of existence

“outside of space and time” and hence inaccessible to us, is one we need
not accept.

’

However, Scanlon

’s next move in the dialectic needs to be challenged. For

having dismissed this idea of an

‘inaccessible region of existence,’ he claims the

only remaining epistemological issue is one of

‘how we can characterise these

structures [that is, of mathematical and also ethical truth] in a way that makes
clear which principles and modes of reasoning are valid.

’

The arguments of Chapter 2 showed that there is another set of

questions we may legitimately ask. These do not concern what modes
of reasoning are objectively valid but how humans have developed belief-
generating capacities that are valid. Scanlon assumes that once he has
shown the

‘Method of Reﬂective Equilibrium’ to be a legitimate one

for us to use in mathematical and moral reasoning he has met all
the demands which can legitimately made of an objectivist account. But
this only answers (D1)

—and leaves (D3) hanging in the air.

In the argument I am pressing against Scanlon, nothing depends

on whether he includes affective responses in his method of re

ﬂective

equilibrium. I have made the case for including these sentiments in our
moral epistemology, because I think it helps us to understand and defuse
some of the remaining appeal of Blackburn

’s position. But even if Scanlon

could

ﬂesh out his account of reﬂective equilibrium in a different way, he

would not be able to meet (D3). As we saw in Chapter 2, natural selection
can account for the validity of our theoretical reasoning, including the

Scanlon, Being Realistic About Reasons, Lecture IV, 3.

Scanlon, Being Realistic About Reasons, Lecture IV, 4.

PROCEDURES AND REASONS

109

truths of mathematics, because we have reason to think that in these

ﬁelds,

true beliefs will be selectively advantageous. However Scanlon develops
his account of re

ﬂective equilibrium, there is no reason to suppose that in

the

ﬁeld of morality truth and selective advantage will coincide. Scanlon

has done nothing to address or defuse the arguments made against secular
objectivism in 2.5.

4.7 Conclusion

Examining Tim Scanlon

’s account of practical reason has been instructive

in several ways. Firstly, it has developed and deepened the case I made
against Blackburn

’s ‘Dionysian’ account. In consequence, I have not only

been able to advance objections to the

‘Dionysian’ position but to identify

ways of developing the

‘Apollonian’ account which learn from and defuse

the appeal of quasi-realism. Secondly, our examination has traced the way
in which Scanlon

’s position has evolved in response to the kinds of

consideration advanced in Chapter 1. His move from a procedural to a
substantive realism occurs for reasons that reinforce those arguments.
Thirdly, and most signi

ﬁcantly for the central thesis of this book, Scanlon’s

philosophical journey provides us with evidence of each of the perils
which I claim the secular moral philosopher must face.

My contention is that secular moral philosophy faces an invidious

choice. All secular accounts either fall foul of the Scylla of insuf

ﬁcient

objectivity or the Charybdis of the explanatory gap (that is to say, they are
unable to account for the human capacity to track moral truth in our
practical reasoning). Scanlon

’s early work makes the ﬁrst of these mistakes.

This is a criticism he has clearly taken on board, which is why his John
Locke Lectures repudiate the procedural realism of What We Owe to Each
Other. I have argued that in correcting this

ﬁrst mistake, Scanlon develops

an account that falls foul of the second of our objections. That Scanlon

’s

account (so elegant and persuasive in other respects) cannot avoid
these perils provides us with powerful evidence that they constitute a
systemic problem for secular moral philosophy, rather than an isolated
and idiosyncratic weakness in a few positions. Either secular accounts fall
foul of our most fundamental moral commitments, or in vindicating them,
they generate an explanatory gap. There seems to be no space for them to
steer between Scylla and Charybdis.

110

FROM MORALITY TO METAPHYSICS

Natural Goodness

Philippa Foot

’s Moral Objectivism

5.1 Introduction

’m not clever at all. I’m a dreadfully slow thinker, really. But I do

have a good nose for what is important. And though the best
philosophers combine cleverness and depth, I prefer a good nose
over cleverness any day.

Philippa Foot, The Grammar of Goodness

This self-deprecating account of Philippa Foot

’s strengths and weaknesses

is at least half correct. Philippa Foot

’s ‘nose’ for what is important in moral

philosophy has led her to a position that seeks to combine the components
we have argued are necessary for a convincing secular account.

5.1.1 Necessary components of secular objectivist account

Firstly, a convincing account needs to vindicate our fundamental moral com-
mitments. (I argued this in Chapter 1.) As we saw earlier, Foot

’s interest in

moral philosophy was stirred by the news coming out of Germany after
the liberation of the concentration camps in 1945. She recognized that the
Nazi terror presented a real challenge to the prevalent moral subjectivism,
for

‘there is no way, if one takes this line, that one could imagine oneself

saying to a Nazi,

“but we are right, and you are wrong” with there being any

substance to the statement.

’

Alex Voorhoeve,

‘The Grammar of Goodness: An Interview with Philippa Foot’,

Harvard Review of Philosophy 11 (2003): 32.

Voorhoeve,

‘The Grammar of Goodness’, 34. A fuller quotation from this passage is

given above at 1.4.6.

Secondly, such an account needs to explain the epistemic access humans have

to moral truths within a secular framework. (I argued this in Chapter 2.) Foot
has always understood the dif

ﬁculty of this task. This is why, in the 1970s,

she explored the possibility that morality might be a set of objective but
hypothetical imperatives (in that sense, rather like the rules of etiquette).

The arguments I have advanced against Korsgaard and Blackburn apply
with even greater force to this phase in Foot

’s thinking, and she had

abandoned the position by 1980.

At that point, she faced a predicament with which readers will (by now)

be familiar. Because of her pre-philosophical moral commitments, she was
unable to accept that morality was hypothetical. However, until she
developed a new theory, she was unable to give a plausible account of
the nature and knowability of moral truth. In her most recent work, Foot
has sought to build a satisfying theory of moral goodness from a wider
account of goodness in nature.

Thirdly, a convincing secular account must provide an adequate explana-

tion of situations where human

ﬂourishing is not to be identiﬁed with maximizing

survival

—for the individual or the collective. (Again, this was argued in

Chapter 2, in the discussion of

‘Darwinian ethics’.) Foot’s treatment of

this issue again reveals the formative in

ﬂuence of the post-war years. In her

most recent book, she discusses the case of the

‘Letter Writers’—prisoners

about to be executed after trial in Nazi Germany for their resistance to the
regime, and assistance they offered to those it sought to exterminate.

When Foot considers the moral choice these writers made, she remarks

that:

One may think there was a sense in which the Letter Writers did, but also a sense
in which they did not, sacri

ﬁce their happiness in refusing to go along with the

Nazis

. . . It may seem that one can get to the bottom of this matter simply by

thinking about the shame that men of the Letter-Writers

’ calibre would no doubt

have felt, in later life, had they gone along with the Nazis

. . . Yet this is not the

heart of the matter. For supposing they had been offered a

‘Lethe-drug’ that would

have taken from them all knowledge of the action? They would not have accepted.
And there would have been a way in which they would not have felt that happiness lay
in acceptance. It is the latter dif

ﬁcult thought I want to hold on to and understand. At

Philippa Foot,

‘Morality as a System of Hypothetical Imperatives’, repr. in her Virtues and

Vices (Oxford: Oxford University Press, 2002).

112

FROM MORALITY OF METAPHYSICS

the very least its presence gives us a clear indication that happiness isolated from
virtue is not the only way in which the concept is to be found in our thoughts.

5.1.2 Aim and structure of the chapter

Foot

’s ‘nose’ seems to have identiﬁed the central components that a

convincing moral theory will need to contain. I will argue that her attempt
to meet these criteria has only limited success, but that the line of thought
she has opened up bears further exploration and development (which it
receives in the work of John McDowell). For Foot

’s attempt to root moral

goodness in natural goodness looks like the only plausible way of avoiding
the predicament faced by the later, more objectivist, approach of Tim
Scanlon. As we have seen, although Scanlon accounts for the epistemolo-
gy of

‘moral reasons,’ he lacks any credible explanation for the way in

which humans came to acquire truth-tracking capacities with respect to
these reasons. What seems to be needed is exactly what Foot is seeking to
provide

—an account on which moral goodness is somehow emergent

from natural goodness, but one which nonetheless avoids collapsing into
Darwinian ethic we rejected in Chapter 2.

This chapter has a threefold structure, and it will consider in turn:

(i) Foot

’s account of the objective qualities and defects of living things

in general;

(ii) her application of this to human beings and their wills; and

(iii) her account of why such facts ought to motivate agents to behave

well.

I will argue that her account only has the

ﬁrst of the three components she

correctly identi

ﬁes as necessary for a successful account of moral truth.

5.2 The Qualities and Defects of Living Things

Since 1980, Foot has been building an account of natural goodness upon
foundations laid by her then graduate student, Michael Thompson.
Thompson argues that there is a category of statement (which he describes
variously as

‘vital descriptions’ or ‘Aristotelian categoricals’) overlooked in

Foot, Natural Goodness (Oxford: Oxford University Press, 2003), 95

–6. Cf.

H. Gollwitzer, K. Kuhn, and R. Schneider (eds), Dying We Live: The Final Messages and
Records of Some Germans who De

ﬁed Hitler (London: Collins, 1960).

NATURAL GOODNESS

113

‘empiricist science’.

Such statements do not simply record the statistical

frequency of particular traits in a species, but specify what traits would be
present in a non-defective member of the said species. The signi

ﬁcance of

this category of statement is that it combines biological description with
normativity. According to Foot, it is from such non-moral normativity in
the natural world that moral norms emerge for natural creatures with both
will and reason.

In this section, I will give a brief exposition of Thompson

’s account of

‘Aristotelian categoricals’ and their role in biological description. I am not
going to evaluate the claims he makes about the logical distinctiveness of these
statements. For the sake of argument, I will assume that he and Foot have
indeed identi

ﬁed a distinctive kind of normativity. My critical remarks will

focus on the transition Foot makes from this natural normativity to moral
objectivity. I will argue that, even on the most generous assessment of
Thompson

’s work, it cannot be the basis for a viable form of moral objectivism.

For Thompson and Foot, the paradigm of an Aristotelian categorical is

given by G. E. M. Anscombe

—the statement ‘Men have thirty-two

teeth

’.

As they observe, the statement does not give the average number

of teeth actual men have. Rather, it is

‘the number of teeth for the species’,

and a man with fewer teeth is, to that extent, physically

‘lacking’. While this

example refers to humans, Thompson argues such categoricals are an
essential part of all biological description. He develops his case with refer-
ence to the way one might describe biological research into a particular
species of jelly

ﬁsh. One might be observing a reef, and come across a

particular jelly, with novel features: an unusually large number of secondary
mouths for its size; disproportionately short tentacles; a very thin and
spread-out

‘bell’. At ﬁrst blush, this would be thought to be a defective

example of a familiar species

—perhaps due to some local pollution. But if

one observed many of these in different places, one would come to think of
them as

‘a novel species . . . a new way for physical particles to be trapped in

the vortex of life-processes.

’ In consequence, one would begin to make

vital descriptions, not based on a statistical survey, but on what turns out be

These terms are used respectively in Michael Thompson,

‘Apprehending Human Form’,

in Anthony O

’Hear (ed.), Modern Moral Philosophy: Royal Institute of Philosophy Supplement 54

(Cambridge: Cambridge University Press, 2004), 51, and Thompson,

‘The Representation

of Life

’ in R. Hursthouse, G. Lawrence, and W. Quinn (eds), Virtues and Reasons: Philippa

Foot and Moral Theory (Oxford: Clarendon Press, 1996), 283.

G. E. M. Anscombe,

‘Modern Moral Philosophy’, Philosophy 33 (1958): 14.

114

FROM MORALITY OF METAPHYSICS

characteristic of the life-form. Statements such as

‘The medusa of the um-

brella jelly has 144 tentacles

’ might be true of this new species—true in the

same sense as the statement

‘Men have thirty-two teeth’.

How salient a defect is (indeed, whether it counts as a defect at all) will

depend on an interlocking set of concepts. Foot lists these as

‘life, death,

reproduction and nourishment, together with less general

—we might say

local

—ideas such as fruiting, eating or ﬂeeing’.

A man with thirty-one

teeth lacks a tooth, but to say he is physically defective would be an
exaggeration. If he had no upper teeth at all, the word

‘defective’ would

be more applicable.

Foot offers a further example:

Compare the colour in the tail of a peacock with that of a bird I have in my garden,
the blue tit. Now, the colour in the tail of the male peacock is necessary: it needs it
to get its mate. On the other hand, as far as I know, the little patch of blue in the
blue tit has no such role in the life of that bird. So if it lacked it, it would be an
oddity, a rarity, but not a defect.

Thompson and Foot claim this collection of statements (

‘Men have thirty-

two teeth

’, ‘A male peacock’s tail is brightly coloured’, and ‘The medusa of

the umbrella jelly has 144 tentacles

’) are examples of a distinctive logical

category

—neither purely empirical nor involving any moral or aesthetic

value-judgement. They exemplify a distinctive, but not usually action-
guiding, kind of natural normativity.

Can these statements not be recast as more straightforward empirical

assertions

—perhaps by adding a ceteris paribus clause? Why can we not say

‘other things being equal, men have thirty-two teeth?’ Would this not
count as a straightforward empirical statement? As Thompson observes, it
is very hard to see how we are going to know what counts as

‘other things

being equal

’ without some judgements about what a ‘normal’ life-cycle

for a creature is.

‘Normal’ is itself irreducibly normative—for in this

context, the word does not mean

‘statistically most prevalent’.

This point is made by Thompson as he develops his thought-example (of

the way one might describe biological research into a particular species of jelly
ﬁsh). He considers the story one would tell of ‘the eggs’ which the jelly ﬁsh
will lay, and how they turn into

‘the polyp’ and, eventually, ‘the medusa’. One

Thompson,

‘Apprehending Human Form’, 47–51.

Foot, Natural Goodness, 36.

Voorhoeve,

‘The Grammar of Goodness’, 38.

NATURAL GOODNESS

115

must write this as a normative story, and in doing so one is not making a claim
that the developmental process thus described happens to most of the eggs:

The population must for generations have remained more or less stable, at least
within a few orders of magnitude. It is clear, then, that only a tiny fraction of the
umbrella jelly eggs have ever realised the story you told about how

‘the egg’

develops into

‘the polyp’ and then ‘the medusa’—a narrative which might seem a

bit Pollyannaish from a certain point of view, but which is forced on you by the
form of representation in which you are engaged. If even a sizeable proportion of
them had for some time followed, in reality, your account of what

‘happens’ to

such a thing, in the natural history, then in a few generations the seas would have
become completely clogged with gelatinous goo.

The notion of what are

‘normal’ circumstances, of what we take to meet the

ceteris paribus condition, is itself parasitic on the Aristotelian categoricals.
Sometimes ceteris paribus implies active intervention. It counts as

‘normal’ for

a bobcat to be nurtured by its mother,

‘abnormal’ to have to fend for itself.

Which interventions count as normal in the representation of life and which
do not can only be expressed in terms of such categoricals. Thompson claims
that statements such as

‘ceteris paribus, the S is F ’ are irreducible to empirical

statements (of the form

‘Some S is F ’, ‘All Ss are F ’, or ‘Most Ss are F ’).

If categoricals are not reducible to empirical statements, might they consti-

tute an ultimately detachable combination of empirical statements and value-
judgements? Again, Thompson argues not, for biological normativity cannot
be understood as a combination of empirical observation and evaluation.
Aristotelian categoricals do not commit us to thinking there is anything
morally good about the

ﬂourishing of cockroaches or cancer cells. Facts

about the life-cycle, without abnormal intervention, of a cancer cell are not
plausibly constituted of (i)

‘Some/All/Most’ observations about the cells and

(ii) value-judgements about what ought to happen. Rather, they tell us what
is true of

‘normal’ cancer cells, and in what their biological ﬂourishing consists.

As indicated above, I am going to assume for the purposes of argument

that Thompson

’s account succeeds. For my claim is that, even with this

concession, Foot cannot build a viable theory of moral goodness on these
foundations of natural normativity.

Thompson,

‘Apprehending Human Form’, 50.

Thompson,

‘The Representation of Life’, 286.

116

FROM MORALITY OF METAPHYSICS

5.3 Goodness and the Human Will

5.3.1 From natural to moral goodness

Foot claims that from a certain (formal) point of view, there is no change in
meaning between uses of the word

‘good’ when referring to an umbrella

jelly, human body, cancer cell, or cockroach, where it does not guide
rational human action in any direct way, and when referring to the human
will. Goodness in human action relates to the same cluster of concepts (life,
death, reproduction, nourishment), and, she claims, the virtues are needed
by humans in the same way the jelly needs a considerable number of its
tentacles, or as bees need stings.

Like bees and ants, humans

ﬁnd a great deal of their good in cooperative

activities. Unlike bees and ants, they have a will which can be consciously
guided by that which enables the species to

ﬂourish. As Anscombe’s

discussion of promising makes clear, the virtue of honesty makes a huge
contribution to human

ﬂourishing. Promising is

a language-game

. . . whose use is part and parcel of an enormous amount of human

activity and hence of human good; of the supplying both of human needs and
human wants so far as the satisfactions of these are compossible. It is scarcely
possible to live in a society without encountering it and even actually being
involved in it.

Each of the virtues will have an analogous justi

ﬁcation: an explanation for

why the dispositions and temperaments involved in them enable human
beings to ful

ﬁl the potential of the species.

5.3.2 Limitations of the biological account

In my critique of moral Darwinism (2.5) I argued that the term

‘ﬂourish-

ing

’ was highly ambiguous. This is because of the unbridgeable gap

between anything that might be justi

ﬁed with reference to our biological

ﬂourishing and our pre-philosophical moral commitments. While Foot’s
position is clearly very different from moral Darwinism, I want to suggest it

An image used by Peter Geach in his The Virtues (Cambridge: Cambridge University

Press, 1977), 17.

G. E. M. Anscombe,

‘On Promising and its Justice, and Whether it Need be Respected

in Foro Interno

’, in her Ethics, Religion and Politics: The Collected Philosophical Papers of

G. E. M. Anscombe (Oxford: Blackwell, 1981), 18.

NATURAL GOODNESS

117

faces a similar problem in bridging the gap between biological

ﬂourishing

and moral value.

On Foot

’s account, a good person will not break a promise simply

because the person to whom it is made will never know and they judge no
harm will be done. While the practice of promising emerges because
human beings need voluntary coordination of action in order to

ﬂourish,

the good human being cares about promising not simply because others
ﬂourish when she keeps her promises, but because it is honest and hon-
ourable to do so.

The problem here for Foot is that, as yet, the only criterion she has

provided for determining which character traits are to be valued is that
they lead to certain states of affairs (namely the

ﬂourishing of the species).

The virtues, as we have seen, are precisely those dispositions and tempera-
ments which enable human beings to ful

ﬁl the potential of the species. It

is hard to see this is to avoid collapsing into what we might call trait
consequentialism

—a position something rather like motive utilitarianism.

How is Foot to rule out the possibility that the pragmatist (who decides

whether to keep promises on a case-by-case basis) has a

‘better’ character

than the principled promise-keeper? It looks as if the only ground for
valuing promise-keeping is the empirical assertion that the species

’ ﬂour-

ishing will be promoted more by the wider character traits of the principled
promise-keeper than by the traits of an agent who opts for case-by-case
pragmatism.

Valuing character traits on the basis of the states of affairs they promote

will lead to some unappealing conclusions. Let us consider again Philip
Kitcher

’s thought-example (discussed at 2.5.2):

Imagine a stereotypical post-holocaust situation in which the survival of the human
gene pool depends on copulation between two people. Suppose, for whatever
reason, that one of the parties is unwilling to copulate with the other. (This might
result from resentment at past cruel treatment, from recognition of the miserable
lives that offspring would have to lead, from sickness, or whatever.) Under these
circumstances [the principle of maximizing the survival of the collective] requires
the willing person to coerce the unwilling person, using whatever extremes of
force are necessary

—perhaps even allowing for the murder of those who attempt

to defend the reluctant one.

See Robert M. Adams,

‘Motive Utilitarianism’, Journal of Philosophy 73 (1976): 467–81.

Philip Kitcher, In Mendel

’s Mirror: Philosophical Reﬂections on Biology (Oxford: Oxford

University Press, 2003), 328.

118

FROM MORALITY OF METAPHYSICS

For a trait consequentialist whose sole concern is species-

ﬂourishing, the

only basis for disapproving of such a post-holocaust rape is the empirical fact
that someone who would commit rape in that situation would be likely
in most other situations to do many things which would be harmful to
the species

’ ﬂourishing. The more ﬁne-grained and nuanced an agent’s

character-traits can become, the more he will be able to act in every
situation in ways that maximize species-

ﬂourishing. So, if traits are to be

valued solely in terms of their contribution to such

ﬂourishing, it is hard to

see why we should prefer traits which lead us to keep our promises and
refrain from sexual assault at all times over a more

ﬁne-grained set of

character traits which (while in all other respects the same as our current
ones) would enable us to get over our abhorrence of promise-breaking
and assault in all and only those cases where they would be optimal for
species-

ﬂourishing.

Of course, Foot will reply that abhorrent acts are, by their very nature,

incompatible with genuine

ﬂourishing.

It is constitutive of a human

being

’s ﬂourishing that he or she would never perform such abhorrent

acts. Therefore human

ﬂourishing is not, as the moral Darwinian claimed,

the same thing as survival, replication and pleasure. Such a reply, while
entirely correct, reveals the one who makes it to already have a moralized
conception of

ﬂourishing. The same moralized conception of ﬂourishing is

behind her perceptive remarks on the Letter Writers. As she observers, the
Letter Writers

’ actions were emblematic of that which is good and noble

in the human character. That would in no way be undermined if it would
be shown that a more pragmatic path would have been better for their,
and indeed the species

’ replication.

Far from meeting the challenge, this simply reinforces the question. For

now there is no longer any sign of a purely biological story of natural
normativity from which morality might emerge. Our judgements as to
what a (morally) good human is, as opposed to say a (biologically) good
cancer cell, are not reducible to facts about biological

ﬂourishing. Foot’s

account of

‘natural goodness’ fails to provide us with the additional

resources we need to make such judgements.

See Foot

’s critique of consequentialism in ‘Utilitarianism and the Virtues’, Mind 94

(374) (1985): 196

–209.

NATURAL GOODNESS

119

5.4. Natural Goodness and Human Motivation

Having argued that Foot

’s account of objective value is deﬁcient, I will

now turn to her account of human motivation. Why does she think
human beings ought to care about promoting the

‘goods’ which ﬂow

from biological

ﬂourishing (that is, the norms deducible from ‘Aristotelian

categoricals

’ about that which makes a member of one’s species good)?

What reason have I for being, in this sense, a

‘good’ human being?

In replying to this query, Foot invokes the work of Warren Quinn:

He made what I think was an absolutely brilliant suggestion, though perhaps neither
he nor others at the time saw the force of it. His move was to ask:

‘What would be so

important about practical rationality if it could be rational to do despicable actions?

’

Now, this thought was extraordinarily original. For it has been more or less taken
for granted in modern moral philosophy that one must

ﬁrst develop a theory of

practical rationality, in terms, say, of the maximum satisfaction of desires, and then
somehow show that even the greatest self-sacri

ﬁce could be rational. And no one,

not even the cleverest, could do it. But Quinn

’s remark suggested that one shouldn’t

tackle it like that at all. One shouldn

’t think that morality must pass the test of

rationality, but rather that rationality must pass the test of morality.

I am happy to concede (both for the purposes of this argument, and because
I think it is actually correct) that Quinn is right about this. Once we have
established a course of action as morally good, there is no need for a further
explanation of why it is

‘rational’ for an agent to do it. As Quinn writes:

A virtue isn

’t a virtue because it’s rational to have it. A good action isn’t good

because it

’s rational to do. On my view, the only proper ground for claiming that a

quality is rational to have or an action rational to do is that the quality or action is,
on the whole, good. It is human good and bad that stand at the centre of rational
thought and not any independent ideas of rationality or reasons for action.

There is something inappropriate about asking for reasons we have for
doing the morally right thing (as opposed to asking what reason we have
to think something is the morally right thing). Either the answer will
offer non-moral reasons for right action, of a kind in principle detachable
from the action

’s rightness (suggesting that if the two came apart, there

would be no reason to act well), or it will offer moral reasons, and be

Voorhoeve,

‘The Grammar of Goodness’, 40–1. Foot also makes clear her debt to

Quinn in the introduction to Natural Goodness and indeed its dedication.

Warren Quinn,

‘Putting Rationality in its Place’, in Hursthouse, Lawrence, and Quinn

(eds), Virtues and Reasons, 207.

120

FROM MORALITY OF METAPHYSICS

question-begging to someone who has not already acknowledged the
compelling nature of moral considerations.

However, even if this is all granted, it does not answer the question

which was posed. Quinn

’s insight could only help Foot at this point in the

dialectic if she had already shown that the natural normative facts about the
survival and

ﬂourishing of the human race could yield facts about moral

goodness. As we have seen, we have no (moral) reason to want (biologi-
cally) good cancer cells and cockroaches to

ﬂourish. What is the precise

connection between facts about the biological

ﬂourishing of our species

and moral goodness? Foot seems to have moved rather too quickly from
one to the other.

Let us suppose cancer cells were agents with choices

—or, perhaps slightly

less fantastically, that humans were agents whose growth and

ﬂourishing

caused immense pain to another species with greater rational and emotional
capacities than our own. Surely the very moral considerations which make
me choose not to maximize my own welfare but to sacri

ﬁce it on occasion

for the wider good of humanity might, in this circumstance, make me
sacri

ﬁce the good of humanity for the good of this other species.

The essence of the problem is that, in moving from Aristotelian cate-

goricals about what constitute different species

’ ﬂourishing to moral im-

peratives about what agents ought to do, judgements are required about
which kinds of

ﬂourishing are of true worth, and how the ﬂourishing of

different species ought to balance against one another. It is no more
reasonable to suppose that I ought only to value the

ﬂourishing of my

own species because I am a member of it, than that I ought only to value
the

ﬂourishing of my nation or family—or indeed, of plain old me. It isn’t

enough to say that because we are humans, we are concerned with human
as opposed to cockroach-

ﬂourishing, any more than it is enough to say

that because I am A. W. Ritchie I am concerned with A. W. Ritchie-
ﬂourishing as opposed to human ﬂourishing.

It is right to reject the question

‘Why should I promote the common

good of humanity?

’ if it is an attempt to secure a non-moral reason for

moral action. The signi

ﬁcant question is this: ‘Why would the correct

objective be the maximizing of the human good, as opposed to (more
narrowly) my own good, the good of my nation, or (more widely) the
goods of other life-forms?

’ When it is put that way, it becomes clear that

Quinn

’s important insight cannot rescue Foot’s account. The norms

which have a claim on us, those which tell us what is objectively good

NATURAL GOODNESS

121

action, are of a kind which cannot be provided with the limited materials
Foot has at her disposal.

Foot

’s account promised a great deal, but without further development

it cannot avoid the inadequacies we identi

ﬁed in Darwinian accounts of

ethics. As it stands, her theory is trapped on the horns of an insoluble
dilemma. On one horn, we can de

ﬁne ‘good’ naturalistically, in which

case it is reduced to that which enables the species to replicate and perhaps
increase in complexity. If we do this, then what ends up being called
‘good’ is not something we take ourselves to have reason to promote. The
alternative horn requires us to de

ﬁne ‘good’ unashamedly to include

evaluative judgements

—including the excellent insights Foot offers in

Natural Goodness about the things that give human life depth, subtlety
and beauty. If we adopt this latter alternative, we have gone far beyond
anything Thompson

’s ‘Aristotelian categoricals’ could justify. To make

this choice is to concede that the idea of

‘ﬂourishing’ is itself heavily

moralized. There is no longer any sign of a purely biological story of
natural normativity from which morality might emerge. Rather, our
judgements as to what a good human is, as opposed to say a good cancer
cell, emerge from previous evaluative commitments.

5.5 Excursus: Varieties of Ethical Naturalism

As I explained in the Introduction, this book does not offer a comprehen-
sive treatment of secular accounts of morality. Philippa Foot

’s later work

offers us only one of a range of theories which we might collectively label
‘realist naturalism’. ‘Cornell realism’—the account developed by Richard
Boyd, Nicholas Sturgeon (who both teach at Cornell), and David Brink
(who was taught there)

—represents another variant on the position.

Like Foot, the Cornell realists argue that

‘goodness’ simply is a natural

property. Richard Boyd describes it as

‘a property quite similar to the other

properties studied by psychologists, historians and social sciences.

’ What is

‘good’ for human beings relates to their biological and social needs and
ﬂourishing.

See Richard Boyd,

‘How to be a Moral Realist’, and Nicholas L. Sturgeon, ‘Moral

Explanations

’, in Geofrey Sayre-McCord (ed.), Essays on Moral Realism (Ithaca, NY: Cornell

University Press, 1988); and David O. Brink, Moral Realism and the Foundations of Ethics
(Cambridge: Cambridge University Press, 1989).

122

FROM MORALITY OF METAPHYSICS

Much of the work done by Boyd, Sturgeon, and Brink is devoted to

combating moral anti-realists

—arguing that the move away from positiv-

ism and towards

‘conﬁrmational holism’ in the philosophy of science

creates a supportive climate for an objectivist understanding of moral
truth. The

‘reﬂective equilibrium’ humans arrive at about what constitutes

our

ﬂourishing and our good is something we are entitled to take to be as

genuine a

‘truth’ about as genuine a ‘reality’ as any other. Our ability to

come to such truth is no great mystery: Boyd argues we are

‘ﬁtted by

nature for moral knowledge

. . . . in a way we are not so ﬁtted for scientiﬁc

knowledge of other sorts

’ precisely because we need to be able to recog-

nise our own and others

’ fundamental needs in order to survive and

replicate as a species.

Much of the argument made by Boyd, Sturgeon, and Brink is echoed in

Part I of this book

—most notably, their defence of ‘reﬂective equilibrium’

thinking as something with a title to truth; the parallels they press between
the nature of theoretical and practical reasoning, and the role of evaluative
concepts (even if supervening on non-evaluative ones) in deliberation and
the description and explanation of action. While in these respects, Cornell
realism is an attractive and powerful contribution to contemporary secular
moral philosophy, with respect to the dialectic of this book, it does not
take us much beyond the dilemmas pressed against Phillipa Foot

’s version

of realist naturalism. By their invocation of re

ﬂective equilibrium, the

Cornell realists do have an answer to the question I pressed (at 5.3) as to
how we know which natural properties are

‘good’ (for they are happy to

concede, in a way Foot cannot, that this is already a moralized process).
But the price of this

—predictably—is that they cannot explain our capaci-

ty for cognition of this kind of moral truth. Boyd

’s invocation of natural

selection to explain these capacities faces the familiar objections I

ﬁrst

advanced in Chapter 2: namely, that our sensitivity to moral reality in-
volves capacities that exceed and indeed contradict that which would be of
selective advantage.

5.6 Conclusion

I began this chapter by praising Philippa Foot

’s ‘nose’ for philosophical

issues. Her theory of ethics has been measured against three criteria which

Richard Boyd,

‘How to be a Moral Realist’, in Sayre-McCord (ed.), Essays on Moral

Realism, 208

–9.

NATURAL GOODNESS

123

have been argued for throughout this book, and are also featured in her
own work

—vindication of our moral commitments; an account of the

human capacity for accurate moral knowledge; and a distinction between
genuine

ﬂourishing (that is ‘happiness’, in the sense Foot herself identiﬁes

with respect to the Letter Writers) and survival and replication of the
species.

Whatever the undoubted merits of Foot

’s instincts on these issues,

I have argued that the account she has developed is unable to meet our
three criteria. In a recent festschrift for Foot, John McDowell hints at the
same verdict:

Philippa Foot has long urged the attractions of ethical naturalism. I applaud the
negative part of her point, which is to reject various sorts of subjectivism and
supernaturalist rationalism. But I doubt whether we can understand a positive
naturalism in the right way without

ﬁrst rectifying a constriction that the concept

of nature is likely to undergo in our thinking. Without such preliminaries, what we
make of ethical naturalism will not be the radical and satisfying alternative to Mrs
Foot

’s targets that naturalism can be. Mrs Foot’s writings do not pay much

attention to the concept of nature in its own right, and this leaves a risk that her
naturalism may seem to belong to [a] less satisfying variety.

If McDowell is correct, Foot

’s account of ‘Aristotelian categoricals’ will

not provide us with all the materials needed for a convincing account of
ethics. However, her

‘nose’ has once again been vindicated, not only in

identifying the components a secular account of ethics needs to contain,
but also indicating the direction in which they are most likely to be found.
As we have seen, the weakness of Scanlon

’s position lay in the lack of a

convincing account of how embodied natural agents acquire a truth-
tracking capacity for moral reasoning. Foot has correctly seen that any
plausible secular account must be rooted in facts about nature, and what it
is for embodied creatures like us to

ﬂourish as a species.

For her account to succeed, however, it requires additional materials

that neither she nor the Cornell realists have provided. To develop a more
satisfying variety of ethical naturalism requires, as McDowell recognizes, a
transformation in our concept of nature.

‘Two Sorts of Naturalism’ invites

us to consider the possibility of a

‘re-enchanted’ naturalism, which does

John McDowell,

‘Two Sorts of Naturalism’, in his Mind, Value and Reality (Cambridge,

MA: Harvard University Press, 1998), 167. (The article

ﬁrst appeared in Hursthouse, Law-

rence, and Quinn (eds), Virtues and Reasons, 149

–79.)

124

FROM MORALITY OF METAPHYSICS

not seek to account for concepts such as

ﬂourishing with the materials of

science alone

—neither the materials used by Boyd, Sturgeon, and Brink,

nor even those of Aristotelian biology. Rather, it urges an expansion in
our understanding of nature to include other kinds of knowledge. In the
ﬁnal chapter of Part II, we will turn to examine the account McDowell
is offering. Given the failures of the various accounts considered in
the preceding chapters, his re-enchanted naturalism provides the secular
objectivist with one remaining hope for success.

NATURAL GOODNESS

125

Natural Goodness and

‘Second Nature’

John McDowell and David Wiggins

6.1 Introduction

[My work can] be taken to defend a version of what has been called
‘moral realism’. But that label would risk obscuring the fact that what
I urge is more negative than positive; my stance in these essays is
better described as

‘anti-anti-realism’ than as ‘realism’. What I urge is

that anti-realist positions such as emotivism and its sophisticated
descendants, all the way down to Simon Blackburn

’s projectivist

quasi-realism, are responses to a misconception of the signi

ﬁcance

of the obvious fact that ethical, and more generally evaluative,
thinking is not science.

John McDowell, Mind, Value, and Reality

6.1.1 Summary of

ﬁndings so far

This is the last of the chapters on secular attempts to close the explanatory
gap. Taken together, they have discussed a wide range of ways in which
this gap might be avoided. We began with the less objectivist accounts of
Simon Blackburn, Allan Gibbard, Christine Korsgaard, and the early Tim
Scanlon, which (I have argued) fail to vindicate our fundamental moral
commitments. As she herself acknowledges, the same criticism applies
to Philippa Foot

’s early attempt to construe morality as a system of

hypothetical imperatives.

John McDowell, Mind, Value, and Reality (Cambridge, MA: Harvard University Press,

1998), viii.

We then turned to the more full-blooded objectivism presented by the

most recent work of Scanlon and of Foot. Scanlon

’s account was immedi-

ately vulnerable to the explanatory gap described in Chapter 2, because it
could not account for the human acquisition of truth-tracking capacities
with respect to the objective moral order. Foot

’s account, with its invoca-

tion of

‘Aristotelian categoricals’, seemed initially more promising, for

it seemed to offer an account of how knowledge of moral goodness
emerged from knowledge of the natural world. However, as Chapter 5
has argued, the kind of

‘ﬂourishing’ we judge valuable cannot be deduced

from (non-moralized) judgements about purely biological

ﬂourishing.

Therefore, there remains an unanswered question as to how humans
acquire the capacity to make accurate judgements of value.

‘Two Sorts of Naturalism’, McDowell advances a similar line of criticism

against Foot. He goes on to argue that the attempt to ground judgements
about (morally valuable)

ﬂourishing in a scientiﬁc account of natural ﬂourish-

ing is a mistake. Instead of seeking to ground ethics in a non-moralized
account of the natural world, McDowell urges Foot to acknowledge that
ethical reasons are themselves part of any adequate account of nature. Once her
conception of nature has been thus

‘re-enchanted’, he claims that Foot’s

account can be developed into a

‘radical and satisfying alternative’ to both

subjectivism and

‘supernaturalist rationalism’.

McDowell argues that once Foot adopts this

‘re-enchanted’ naturalism,

there will no longer be any pressure to offer a non-circular defence of virtue:

Of course decent people (like us) think acting in accordance with the virtues is, as
such, a good. So we can say we need the virtues, since without them we cannot
attain that good

. . . To say ‘We need the virtues’ is just to say that ethical con-

siderations constitute genuine reasons for acting, not to give the outline of a
grounding for that claim.

6.1.2 Structure and aim of the chapter

The aims of this chapter are:

ﬁrstly, to introduce McDowell’s concept of ‘second nature’, and

explain the role he takes it to play in solving Foot

’s difﬁculties;

McDowell,

‘Two Sorts of Naturalism’, in his Mind, Value, and Reality, 168.

NATURAL GOODNESS AND

‘

SECOND NATURE

’ 127

secondly, to locate McDowell’s position in a wider philosophical

context

—in particular, with respect to the thought of Kant and

Wittgenstein;

thirdly, to argue that, even on McDowell’s own terms, there is a

legitimate demand for explanation, and,

fourthly, to argue his account of ‘second nature’ does not give him

the resources to answer the demand.

After this primary task is completed, I will go on to discuss whether the
problems which McDowell faces might be avoided on a slightly weaker
conception of the objectivity of moral truth. That discussion will draw
heavily on the work of David Wiggins. I will argue that Wiggins

’ position

is vulnerable to the arguments given in Chapters 3 and 4 against Blackburn
and Korsgaard.

6.2 Introducing

‘Second Nature’

6.2.1 Contrast with alternative positions

McDowell introduces his

‘re-enchanted’ account of nature by drawing a

contrast between it and what he calls

‘bald naturalism’. By this latter phrase

he means a reductionist position which takes

‘nature’ to be synonymous

with

‘that which science investigates and explains’. When this reductionist

position has to account for

‘the space of reasons’—that is, human belief,

reasoning and action

—it must attempt to reconstruct the ‘space’ using only

the conceptual materials that already belong in a scienti

ﬁc depiction of

the natural world.

As McDowell observes, the prospects for such a

reconstruction do not look promising.

If bald naturalism has such problems, what is its attraction? Its advocates

tend to assume that the only other materials which might be available to
reconstruct the space of reasons would be supernatural. In McDowell

’s words,

they assume that, unless we can reconstruct the space of reasons from the
materials of the natural sciences, we must instead picture that space as

an autonomous structure

—autonomous in that it is constituted independently of

anything speci

ﬁcally human, since what is speciﬁcally human is surely natural (the

John McDowell, Mind and World (Cambridge, MA: Harvard University Press, 1994), 73.

128

FROM MORALITY TO METAPHYSICS

idea of the human is the idea of what pertains to a certain species of animals), and
we are refusing to naturalize the requirements of reason.

McDowell terms this latter position (to which bald naturalism is a reaction)
‘rampant Platonism’. One of many reasons rampant Platonism is unattrac-
tive to secular thinkers is that it looks likely to generate precisely
the explanatory gap that I described in Chapter 2. For, if the space of
reasons is wholly independent of the natural world, then

‘[o]ur capacity to

resonate with [its autonomous structure] has to be mysterious; it is as if we
had a foothold outside the animal kingdom, in a splendidly non-human
realm of ideality.

’

McDowell

’s own conception of second nature has, he assures his reader,

‘no whiff ’ of rampant Platonism. If there are features of the natural world
which are not captured by the concepts of the sciences, then these features
may both explain the independence of norms and account for the ability
of human beings (as rational animals) to resonate with them.

McDowell

’s

re-enchanted naturalism therefore looks as if it may provide the solution that
we have been looking for throughout Part II

—namely, a secular account

which is both suf

ﬁciently objective to vindicate our fundamental moral

commitments and which none the less avoids generating an explanatory gap.

If re-enchanted naturalism is to provide a via media between bald natural-

ism and rampant Platonism, it must also provide an account of reason, belief,
intention, and value which makes them both entirely natural and irreducible
to the concepts of natural science. In this chapter, my main concern will be
whether re-enchanted naturalism can offer a tenable account of the reasons
such value gives humans for action, and of the ways beliefs about value are
generated within, and weighed up by, human beings. For it is one thing for
McDowell to state what the proposed

‘re-enchantment’ of nature would

need to do in order to succeed, and quite another to show that it can be
done. In the section which follows, we begin to evaluate his account, and in
particular the role of

‘second nature’ in practical reasoning.

6.2.2 A thought-example: wolves with logos

McDowell introduces his conception of

‘second nature’ by way of a

thought-example. We are invited to imagine some wolves to acquire

McDowell, Mind and World, 77.

McDowell, Mind and World, 94.

McDowell, Mind and World, 88.

NATURAL GOODNESS AND

‘

SECOND NATURE

’ 129

the ability to reason about their action.

The point of the example is to

show how moral reasons might or might not emerge from a wholly
naturalized account. (Of course, this thought example has some immediate
problems. Worries about a private language argument are bypassed
by McDowell

’s stipulation that it is a group of wolves who acquire this

new capacity, not a solitary beast. As he acknowledges, there is also the
further problem of how we would understand such wolves, and if we
could not, whether we would have reason to take them to be talking.
McDowell

’s response to these worries is that ‘for my purposes it does not

matter whether it is genuinely intelligible

. . . . it will be enough if we can

get as far as pretending that it is.

’)

The thought-example is designed to help us see how

‘second’ nature

emerges from

‘ﬁrst’ nature—that is, from that which is investigated and

described by science, and which controls the behaviour of non-rational
animals. The hypothetical wolf is one of a small sub-group who are not
trapped in

ﬁrst nature, in which actions are constrained solely by law-like

regularities. They, by contrast, are able to stand back from habitual wolf-
like behaviour, and consider what kind of action they have reason to
perform. Given their re

ﬂective ability regarding what was previously

habitual behaviour (such as participation in the cooperative hunting of
the pack), any individual rational wolf can now ponder why he should not
become a free-rider. McDowell argues that to think of any such wolf as
being in the

‘space of reasons’ (or a ‘possessor of logos’) involves viewing

him from a different perspective to the law-governed one taken by the
natural sciences:

A possessor of logos cannot be just a knower, but must be an agent too; and we
cannot make sense of logos as manifesting itself in agency without seeing it as
selecting between options, rather than simply going along with what is going to
happen anyway

. . . Freedom of action [is] inextricably connected with a freedom

that is essential to conceptual thought.

(In McDowell

’s usage, ‘logos’ articulates a concept which in English

requires a range of words

—reason, rationality and discursiveness. The

wolf

’s possession of logos enables him to reﬂect on how to act, and so to

McDowell,

‘Two Sorts of Naturalism’, 169.

McDowell,

‘Two Sorts of Naturalism’, 170.

130

FROM MORALITY TO METAPHYSICS

see this re

ﬂective, discursive process as having a causal impact through his

actions on the wider world.)

What reasons does the wolf have for acting virtuously? McDowell

argues, as I did at 5.4, that the

‘Aristotelian categoricals’ advanced by

Thompson and Foot will not help. The statement that

‘Human beings

have thirty-two teeth

’ (a classic example of such a categorical),

does not

entail, when combined with

‘I am a human being’, that ‘I have thirty-two

teeth

’. As McDowell observes, the same is true in the case of needs: from

‘Wolves need such-and-such’ and ‘I am a wolf ’ our hypothetical wolf
cannot conclude that he needs such-and-such:

The deductive impotence of the

‘Aristotelian categorical’ brings out that what

wolves need is not guaranteed a rational bearing on [the question

‘what should

I do

’, asked by such a wolf]; and this even if he never forgets that he is a wolf.

Reason does not just open our eyes to nature, as members of the animal species we
belong to; it also enables and even obliges us to step back from it, in a way that puts
its bearing on our practical problems into question

. . . Even if we grant that human

beings have a naturally based need for the virtues in a sense parallel to the sense in
which wolves have a naturally based need for cooperativeness in their hunting, that
need not cut any ice with someone who questions whether virtuous behaviour is
genuinely required by reason.

Having argued that

‘Aristotelian categoricals’ are not enough to get the

wolf to the reasons for virtuous action, McDowell goes on to argue that
the justi

ﬁcatory project in which Thompson and Foot deploy them is in

fact unnecessary. It is only if we make the mistaken assumption that the
impersonal empirical method of the sciences provides our sole access to
objective reality that we will also assume (as a consequence) that any
objective reasons the wolf has to act must be discernible from a perspective
which

ﬂoats free from the wolves’ common life.

McDowell writes:

Consider a rational wolf whose acquisition of practical reason [logos] included being
initiated into a tradition in which co-operative behaviour in the hunt is regarded as
admirable, and so as worth going for in its own right. What wolves need might
ﬁgure in a bit of reﬂection that might help reassure him that, when he acquired a
second nature with that shape, his eyes were opened to real reasons for acting.

See G. E. M. Anscombe,

‘Modern Moral Philosophy’, Philosophy 33 (1958).

McDowell,

‘Two Sorts of Naturalism’, 172–3.

McDowell,

‘Two Sorts of Naturalism’, 190–1.

NATURAL GOODNESS AND

‘

SECOND NATURE

’ 131

The re

ﬂection which McDowell urges on this wolf would, he says,

‘Neurathian’ (this is a reference to Otto Neurath’s famous image of

humans and their knowledge of the world as being

‘like sailors who on the

open sea must reconstruct their ship but are never able to start afresh from
the bottom

’).

Such a re

ﬂection would not offer the wolf grounds for

moral action wholly independent of his sensibilities, for no such dry land
is available. Rather, it would reassure the wolf that his (ex hypothesi well-
formed) sensibilities were to be trusted in the absence of speci

ﬁc reasons for

doubt. Such reassurance would not cut any mustard with a wolf who had
not acquired a sense that cooperation for its own sake was valuable. But
this, we are told, should not worry our wolf, or indeed us.

‘[T]he point is

harmless to the genuine rationality of virtue, which is visible (of course!)
only from a standpoint from which it is open to view.

’

Thus far, this account is in harmony with the position I defended in

Chapter 1. Neither I nor McDowell expects to be able, nor feels the need to
ﬁnd, anything persuasive to say to a human or an otherwise rational wolf
whose upbringing has left them insensitive to the demands of morality.
(That is to say, we do not think such an agent can be reasoned into morality by
deductive argument from premisses he will already accept.) Because

‘the

genuine rationality of virtue

’ is visible only from the stand point of virtue,

we both take a key task of moral philosophy to be the reassurance of morally
sensitive agent that their epistemic position is not undermined by an inability
to force these sensibilities on someone with a perfect grasp of theoretical
reasoning, but without the necessary sensibilities.

On McDowell

’s account, the fact that morality has no external

guarantees (that is, the fact that the grounds for accepting it are internal
to the practice) does not make it uniquely problematic. He argues that, in
the natural sciences as in ethics, there is no dry land. Our justi

ﬁcation for

the ways in which we reason in science are also necessarily

‘Neurathian’. In

each domain we must presuppose some trust in the veracity of the ways
in which well-formed human reasoners proceed:

Otto Neurath, Empiricism and Sociology, ed. Marie Neurath and Robert S. Cohen

(Dordrecht: Reidel, 1973), 198.

McDowell,

‘Two Sorts of Naturalism’, 186. McDowell develops his conception of

moral knowledge via a more general discussion of the way in which we come to know the
world. He wants to show that moral anti-realists (such as Mackie) cannot help themselves to
the use of logos in our knowledge of the natural world while simultaneously denying the same
logos gives us knowledge of an objective world of values.

132

FROM MORALITY TO METAPHYSICS

Whatever the details, an acceptable world-picture consists of articulable, concep-
tually structured representations

. . . We have to suppose that the world has an

intelligible structure, matching the structure in the space of logos possessed by
accurate representations of it.

To pose a general question about the reliability of logos in science or in
ethics (as opposed to making speci

ﬁc proposals for constructive revision)

is to raise a doubt to which there cannot be a non-question-begging
answer. While we can use logos to stand back from and evaluate the
deliverances of habit, impulse and inclination, we have no apparatus by
which to evaluate logos itself. To attempt such an evaluation would be a
category mistake. Logos just is that by which good reasoners reason.

It is because of this wider philosophical picture, that McDowell is resistant

to accepting

‘moral realism’ as a label for his meta-ethical position. His

hesitancy does not

ﬂow from any denial that moral truth is ‘real’ or ‘objective’.

ﬂows from the fact he takes many who call themselves moral realists to have

gone beyond

‘relaxed common sense’—a mistake which mirrors that of their

anti-realist opponents.

His charge is that such

‘realists’ have given the reality

of moral truth a metaphysically problematic foundation which it does not
need, and in doing so have provoked an unnecessary anti-realist response. To
contrast his position with either of these mistaken views, McDowell prefers to
call his position

‘anti-anti-realism’.

6.2.3 McDowell on Aristotle

Anti-anti-realism is conceived as a return to an ethical outlook which
predates the distinctively modern preoccupations of philosophy. For
McDowell, Aristotle is the prime example of a

‘Neurathian’ moral philos-

opher, who understood that only those with virtuous sentiments could
comprehend the demands of goodness.

It is anachronistic to think Aristotle would (along with today

’s

moral realists) seek to give morality some kind of external validation. As
McDowell writes:

[i]t is so tempting for modern readers to credit Aristotle with a different picture of
the sort of validation which an ethical outlook needs: a picture in which

. . . the

[Neurathian] boat is put ashore for a certi

ﬁcation of its seaworthiness.

McDowell,

‘Two Sorts of Naturalism’, 178.

He uses this term in his debate with Crispin Wright on normativity and meaning in

McDowell, Mind, Value, and Reality, 321.

NATURAL GOODNESS AND

‘

SECOND NATURE

’ 133

It never occurs to Aristotle that such a thing might be needed, for he
assumes the basic reliability of our faculties without even noticing he is
doing so. The question of global justi

ﬁcation (as opposed to speciﬁc

proposals for improvement) is foreign to him.

McDowell

’s contention is that it is we moderns, rather than Aristotle,

who have made the mistake:

Making historical and cultural speci

ﬁcity into a metaphysical issue is distinctively

modern

. . . [In fact,] organising our metaphysics around the idea of transcending

historicity is profoundly suspect. Its true effect is to undermine the very idea of
getting things right. We conceal that from ourselves only if we think we can make
sense of the idea of a mode of enquiry that transcends historicity. In our modern
culture, natural science tends, quite intelligibly, to be cast in that role, but any such
conception of science is an illusion.

For McDowell, all of the aspects of our reasoning (that is, all that is part of
logos) are the product of historically and culturally contingent commu-
nities. He assures us that this should not call into question our capacity to
get things right, because although we have to proceed by our own lights,
those lights are

‘self-correcting’ and ‘stand up to reﬂective scrutiny’.

Where does this leave us? In this book, I have been advocating an

account of moral epistemology which is very similar to McDowell

’s. We

both take it that, for all our deliberatively indispensable practices, it is
legitimate to trust our faculties and to assume that the best efforts we have
at

‘Neurathian’ reﬂection (in other words, what I have called ‘reﬂective

equilibrium

’) move us towards the truth. It has been part of my argument,

just as it has been part of McDowell

’s, that the epistemology of moral

reasoning is not in this respect much different from that of the norms of
theoretical reasoning on which the sciences rely. Much that McDowell has
to say on the role of well-formed sensibilities is also complementary to the
case I advanced in Chapter 4 for an account of the cognitive role of the
moral sentiments. And yet there remains an explanatory question which
McDowell is unwilling to answer, namely the one spelt out in Chapter 2.
This question is distinct from the justi

ﬁcatory issues on which he and I agree.

It is not obvious that we can simply wish ourselves back to a kind of

Aristotelian innocence. For if Aristotle is right to be untroubled by our
modern epistemological concerns, he is certainly not a

‘quietist’ on the

McDowell,

‘Eudaimonism and Realism in Aristotle’s Ethics’, in his The Engaged Intellect:

Philosophical Essays (Cambridge, MA: Harvard University Press, 2009), 37

–8.

134

FROM MORALITY TO METAPHYSICS

explanatory issues that this book has been pressing. If McDowell is to
develop a

‘re-enchanted naturalism’ with Aristotelian ethics at its heart, he

must recognize the signi

ﬁcance of the fact Aristotle has a purposive view of

the universe. Aristotelian biology is fundamentally different from modern
science, in that Aristotle takes it for granted (in a way that we shall see
echoed in the next chapter by Hugh Rice) that the ultimate explanation
for all that occurs is teleological. An excellent example of this is his
discussion of the lung, larynx, and oesophagus in On the Parts of Animals:

Now the larynx is present by nature for the sake of breathing; for it is through this
part that animals draw in and expel air when they inhale and exhale. This is why
those without a lung have no neck, e.g. the kind consisting of the

ﬁsh. The

esophagus is the part through which nourishment proceeds to the gut; so that
animals without necks manifestly do not have an oesophagus. But it is not necessary
to have the oesophagus for the sake of nutrition; for it concocts nothing. And
further, it is possible for the gut to be placed right next to the position of the mouth,
while for the lung this is impossible. For there needs

ﬁrst to be something common

like a conduit, which then divides in two and through which the air is separated into
passages

—in this way the lung may best accomplish inhalation and exhalation.

The assumption at the heart of Aristotle

’s discussion is that states of affairs

can be explained by reference to their function. He is therefore able to
offer a parallel explanation with respect to the truth-tracking quality of
humans

’ belief-generating and belief-correcting mechanisms. These too

will have a teleological explanation; namely that, for humans to function as
rational animals, it will be necessary for both their capacities for theoretical
and practical reasoning to track the truth.

One does not have to be in thrall to scientism to notice that the

mainstream understanding of explanation has changed. For McDowell
to sustain something like an Aristotelian account of ethics, he will have
to address metaphysical issues that Aristotle did not face. To say

‘Aristotle

did not face them

’ is neither to say that he was a quietist nor that he

overlooked them through negligence. Rather, Aristotle

’s conception of

science, unlike that of most moderns, meant they did not arise. Unless
McDowell is urging a return to a fundamentally purposive account of the
universe, the question of how we explain (rather than justify) the reliability
of our belief-generating and belief-correcting processes will arise for him

Aristotle, On the Parts of Animals, 664a14

–34, trans. J. G. Lennox (Oxford: Oxford

University Press, 2001).

NATURAL GOODNESS AND

‘

SECOND NATURE

’ 135

in a way it did not for Aristotle.

To see how McDowell answers this new

question, we must turn to his wider account of the relationship between
the human mind and the world it seeks to comprehend.

6.3 The Wider Philosophical Framework

Whereas Aristotle

’s ‘lack of anxiety’ about our ability to cognize

an objective world is supported by his purposive account of nature,
McDowell

’s equally sanguine attitude has a different basis—namely, the

account he gives of the relationship between mind and world. It is to Kant
and Wittgenstein that McDowell turns to develop an account of
the relationship between these which might enable him to reject the
explanatory demand I am seeking to press. In this section I will give an
outline of McDowell

’s wider philosophical framework. I will make it a

brief sketch, for I do not wish to contest the framework. Rather, having
explained it in outline, I will go on to argue that, even if we assume it to be
correct, my demand for explanation is legitimate.

6.3.1 Kant and the

‘myth of the given’

McDowell pays homage to Immanuel Kant for demonstrating that
rational agents

‘have to suppose that the world has an intelligible structure,

matching the structure in the space of logos possessed by accurate
representations of it.

’

However, Kant

’s demonstration generates one of

the central problems with which he and his successors have to grapple. The
problem is this: without Aristotelian biology to hand, the correlation
between the structure of the world and the structure of our cognitive
capacities requires an explanation. McDowell

’s response to this is to

deploy another Kantian insight, that our concepts structure the world. If
the world is not

‘constitutively independent’ of logos then, he argues, the

question of how mind and world come to be in harmony does not arise.

As McDowell acknowledges, the way he combines these two insights

differs in some fundamental respects from Kant

’s own. On Kant’s account,

the world which our concepts structure is itself undergirded by the
noumenal world. It is only because there is a

‘Given’ (wholly independent

Thomas Nagel

’s forthcoming book argues for something much closer to Aristotelian

teleology

– and has been much inﬂuenced by Sharon Street’s arguments for the incompatibility

of evolutionary biology and moral realism. See Thomas Nagel, Mind and Cosmos: Why the
Materialist Neo-Darwinian Conception of Nature Is Almost Certainly False (Oxford: Oxford University
Press, forthcoming) and Sharon Street,

‘A Darwinian Dilemma for Realist Theories of Value’.

McDowell,

‘Two Sorts of Naturalism’, 178.

136

FROM MORALITY TO METAPHYSICS

of our conceptualization) that human concepts have anything to work on
in the

ﬁrst place. Here McDowell parts company with Kant, rejecting

what he calls the

‘myth of the Given’. All there is, on McDowell’s account,

is the world which our concepts structure.

McDowell goes on to argue that Kant

’s position is inconsistent, for if he

is to talk of the noumenal world he will require human concepts.
One cannot talk of the thing-in-itself existing without recourse to
our conceptual scheme. McDowell argues that this leaves Kant with an
invidious choice. One option is that Kant can stop short of claiming that
the thing-in-itself exists, in which case he will end up with a coherentism
which cannot distinguish between our beliefs and the world. The other
alternative is that he say the thing-in-itself does exist

—in which case, on

pain of anti-realism, he must accept that at least some of our concepts
‘reach out’ to how things actually are. Once he has accepted that some of
our concepts can reach out to the world, there is no motivation for
postulating an

‘ontological gap between the sort of thing one can mean,

or generally the sort of thing one can think, and the sort of thing that can
be the case

’. In consequence, McDowell asserts, we should replace Kant’s

transcendental idealism with a straightforward af

ﬁrmation of our ability to

know the world:

‘When one thinks truly, what one thinks is the case.’

On McDowell

’s account, Kant’s Idealist successors were therefore right

to grasp, as Kant himself did not, that

the fundamental thesis, that the world cannot be constitutively independent of the
space of concepts, does not require this residual recognition of an

‘in itself ’ . . . It

was only because of the picture of co-workers in the transcendental constitution of
the world that it seemed necessary to gloss the idea that world and thought are
constitutively made for each other by saying, what indeed sounds idealistic, that
mind makes a contribution to the world.

To see that world and thought are constitutively made for each other is not
yet to see how this is so. That we must take our thoughts to succeed
in reaching right out to the world does not explain how they succeed in
making contact with it, for our very concept of a

‘world’ is precisely of

something which is not dependent on humans for its existence. This is not
to say we have to conceive of the world being

‘brutely alien to the space of

McDowell, Mind and World, 27.

McDowell,

‘Intentionality and Interiority in Wittgenstein’, in his Mind, Value and

Reality, 306

–7.

NATURAL GOODNESS AND

‘

SECOND NATURE

’ 137

logos

’ but to say that speaking (as McDowell himself often does) of this

in terms of a

‘made-for-each-other-ness’ and a ‘matching’ invites a further

question: why do the world and our concepts

ﬁt so snugly together?

Once again, there are two opposing answers which McDowell

wishes to avoid. At one extreme is anti-realism, which explains the
‘made-for-each-other-ness’ of the world and mind by dependence of
the former on the latter. At the other extreme, the

‘rampant Platonist’

assumes that the only alternative is to conceive of our concepts locking
onto a completely independent reality; taking there to be a mysterious
correlation between the way we deploy words and concepts and a set of

‘rails

of meaning

’ stretching out to inﬁnity.

6.3.2 Against Platonism: Wittgenstein on

‘rule-following’

McDowell

’s objection to ‘rampant Platonism’ is twofold. Firstly, as we

have already seen, it generates an explanatory gap. In its secular version,
Platonism cannot answer the question of how humans manage to lock on
to the non-natural

‘rails’ it postulates. Secondly (and more fundamentally)

McDowell argues that if such

‘rails of meaning’ existed they would be of

no practical use. Here his argument depends heavily on the work
of Ludwig Wittgenstein on rule-following. As Wittgenstein seeks to
demonstrate, the evidence we have at any point for having

‘locked on’

to a set of rails at a given time is compatible with divergence in the future.

Wittgenstein goes on to rule out a range of candidates for the

item which secures our continuation along these

‘super-rigid’ rails.

This function could not be ful

ﬁlled by an interpretation of the rule (for

reasons familiar to readers of Lewis Carroll)

nor by a mental token

(for the question of its interpretation would arise).

So the bulwark

against such divergence is not a formula but a practice. On this account
of what it is to follow a rule, there is a point at which the chain of
justi

ﬁcations comes to an end. It must come to an end in action:

When I obey a rule, I do not choose. I obey the rule blindly.

Lewis Carroll,

‘What the Tortoise said to Achilles’, Mind 4 (1895): 278–80.

Cf. David Pears, The False Prison (Oxford: Clarendon Press, 1988), esp. chs 13

–17, and

William Child

’s ‘Pears’ Wittgenstein: Rule-Following, Platonism and Naturalism’, in David

Charles and William Child (eds), Wittgensteinian Themes: Essays in Honour of David Pears
(Oxford: Oxford University Press, 2001), 81

–114.

Ludwig Wittgenstein, Philosophical Investigations (Oxford: Blackwell, 1953),

}219.

138

FROM MORALITY TO METAPHYSICS

‘So you are saying that human agreement decides what is true and what is

false?

’—It is what humans say that is true and false; and they agree in the language

they use. That is not agreement in opinions but in form of life.

Wittgenstein

’s (and McDowell’s) position requires a certain cordoning

off of epistemology from anthropology. Within our linguistic practice,
communal agreement is no part of what we take to justify a particular
piece of concept-application. Noticing that communal agreement is
not what justi

ﬁes concept-application, the rampant Platonist (and the

anti-realist) go off and look for something else to play this role. The ram-
pant Platonist thinks his rails provide the missing thing, whereas the idealist
rejects realism on the basis that no such rails are to be found. McDowell
and Wittgenstein reject both these options as mistaken pieces of philoso-
phy. As scientists we can study what kinds of anthropological commonal-
ities prevent agreement from breaking down and the practice from
fragmenting

—but as philosophers we should eschew the ‘sideways-on’

view. To use a favourite image of Wittgenstein:

The dif

ﬁcult thing here is not, to dig down to the ground; no, it is to recognise the

ground that lies before us as the ground.

The temptation to dig below bedrock comes when we forget that
objectivity can only be spoken of from within our practices. McDowell
claims that, if we are feeling dissatis

ﬁed by the lack of platonic rails, the

very last thing we need is yet more philosophy:

What is needed is not so much reassurance

—the thought that after all there is solid

ground under us

—as not to have felt the vertigo in the ﬁrst place. Now if we are

simply immersed in our practices, we do not wonder how their relation to the
world would look from outside them

. . . [W]e would be protected against the

vertigo if we could stop supposing that the relation to reality of some area of our
thought and language needs to be contemplated from a standpoint independent of
that anchoring in our human life that makes the thoughts what they are for us.

There is a great deal more to be said here

—both about McDowell’s

account of Kant and about the validity of Wittgenstein

’s account of

rule-following. But, for the purposes of this book, I propose to sidestep
these disagreements and assume for the sake of argument that McDowell

Wittgenstein, Philosophical Investigations,

}241.

Wittgenstein, Remarks on the Foundations of Mathematics (Oxford: Blackwell, 1978), VI-31.

McDowell,

‘Non-Cognitivism and Rule-Following’, in his Mind, Value, and Reality, 211.

NATURAL GOODNESS AND

‘

SECOND NATURE

’ 139

is correct on these matters. That is to say, I will accept for the sake
of argument that any demand for explanation that involves a

‘sideways-

’ comparison between an unconceptualised ‘Given’ and our cognitive

apparatus is illegitimate.

In the sections which follow, I will argue that even with this concession,

the demand for explanation which I advanced in Chapter 2 remains legiti-
mate. Having argued that the demand is legitimate, I will turn (in 6.4.4)
to consider whether McDowell

’s account has the resources to meet it.

It is important to see that these are potentially separable questions.

McDowell

’s own reasons for refusing to answer the demand for explanation

ﬂow from his quietism. If the quietism is rejected, we can still ask whether
the demand for explanation stands any chance of being met, given the
account of

‘second nature’ which McDowell has been advancing. I will

address these two questions in turn in the sections which follow.

6.4 The Demand for Explanation

6.4.1 McDowell

’s rejection of the demand

I argued in Chapter 2 that we need to distinguish three demands that
might be pressed with respect to our belief-generating faculties. They are:

(D1)

What is the justi

ﬁcation for our faith in their reliability?

(D2)

What is the historical explanation of their development?

(D3)

What is the explanation for their capacity for tracking truth?

McDowell clearly allows (D1) and (D2) as legitimate demands. Regarding
(D1), we have seen that his account is very similar to that which I have
been developing throughout this book. But in terms of an explanation for
the development of those capacities, McDowell allows only an historical
account of the ways in which we are formed by those around us as virtuous
and reasonable members of a culture.

One of the clearest and most succinct examples of such an explanation is

given in Sabina Lovibond

’s essay on the emergence of ‘second nature’

from its animal precursors. (In a response to the essay, McDowell explicitly
endorses this as a fair representation of his position.) Lovibond writes:

[A] human being remains an animal

. . . but humans alone develop beyond the stage of

‘enslavement to immediate biological imperatives’, since they alone acquire the capaci-
ty for a continual active adjustment of their beliefs and intentions to the

‘deliverances of

140

FROM MORALITY TO METAPHYSICS

experience

’—a process of adjustment that engages powers of self-criticism (‘Perhaps

p isn

’t true after all . . . Perhaps a would not be the best thing to do . . . ’ etc.), and so

exhibits our

. . . ‘free, distanced orientation’ to the input we receive from our senses

from moment to moment, or to our emotional responses to that input. And the
sensitivity to reasons that is expressed in the workings of our

‘spontaneity’, or in the

continual active correction of our thinking, can be seen as providing the content of
the

‘second nature’ that we acquire through ordinary human upbringing. For the

distinctively human thing that we get from our upbringing, the thing that goes
beyond mere habituation to various routines, is an ability to

ﬁnd our way around

within the

‘space of reasons’ established by the culture into which we are being

initiated

—an ability, then, to treat that culture as a ‘repository of tradition, a store of

historically accumulated wisdom about what is a reason for what

’.

This is a good example of an account that meets (D2)

—an historical

explanation of the development and transmission of our cognitive capa-
cities through culture and tradition. Yet the account does not explain how
our capacities come to track truth, which (McDowell and Lovibond
would both agree) is more than simply communal agreement.

How then does McDowell respond to (D3)? Can he not refer us back to

the justi

ﬁcation for relying on logos? An account like Lovibond’s explains

how we have acquired our set of cognitive capacities. McDowell

’s episte-

mological arguments explain why we are entitled to trust them. Surely to
demand more than this is to lapse into seeking a

‘sideways-on’ comparison

between our capacities and the world they represent to us? In the section
which follows, I will argue it is not.

6.4.2 Response to McDowell

I agree with McDowell that we are entitled, in the absence of speci

ﬁc

reasons for scepticism, to rely upon the operations of our belief-generating
and belief-revising faculties. Thus, we can assume that, in addition to their
many distinctive properties, our perceptual capacities {a, b, c

. . . }, our

capacities for theoretical reasoning {g, h, i

. . . }, and our capacities

for practical reasoning {m, n, o

. . . } all share the property T (that is, the

property of tracking truth).

Accounts like Sabina Lovibond

’s explain the development and transmis-

sion amongst humans of the speci

ﬁc capacities we have—{a, b, c . . . g, h,

Sabina Lovibond,

‘Practical Reason and its Animal Precursors’, in Jakob Lindgaard (ed.),

John McDowell: Experience, Norm and Nature (Oxford: Blackwell, 2008), 115. McDowell
endorses her account of his position in

‘Responses’, 234–8.

NATURAL GOODNESS AND

‘

SECOND NATURE

’ 141

. . . m, n, o . . . }. But there is nothing as yet to explain why they all possess

this further property T. Given the arguments made in Chapter 2, we are
surely entitled to ask: why is it that we arrived at {g, h, i

. . . } which have

property T, rather than, say, {g

’

, h

, i

} which are selectively advantageous

simulacra? That is not to raise a sceptical doubt: the question here is not
whether we have capacities with property T, for that question has been
answered in the way McDowell and I both meet (D1). What we want to
know is why the capacities we have acquired track the truth.

McDowell

’s view is that once (D1) and (D2) have been met, any further

explanatory demand must be seeking a

‘sideways-on comparison’. I have

accepted (for the sake of argument) that

‘sideways-on’ comparisons are to be

rejected. So I will now argue that (D3) is indeed compatible with his
strictures.

McDowell gives a clear statement of his position in

‘Wittgenstein on

Following a Rule

’, and to ensure a fair representation of his account, I will

quote it at some length:

Understanding is a grasp of patterns that extend to new cases independently of our
rati

ﬁcation, as required for meaning to be other than an illusion (and—not

incidentally

—for the intuitive notion of objectivity to have a use); but the con-

straints imposed by our concepts do not have the platonistic autonomy with which
they are credited in the picture of the super-rigid machinery. As before, what
obscures the possibility of this position is the

‘anti-realist’ attempt to get below

‘bedrock’. [Crispin] Wright suggests that the emergence of a consensus on wheth-
er, say, to call some newly encountered object

‘yellow’ is subject to no norms. That

is indeed how it seems if we allow ourselves to picture the communal language in
terms of sub-

‘bedrock’ resemblances in behaviour and phenomenology. But if we

respect Wittgenstein

’s injunction not to dig below the ground, we must say that

the community

‘goes right or wrong’ according to whether the object in question

is, or is not, yellow; and nothing can make its being yellow, or not, dependent on
our rati

ﬁcation of the judgment that that is how things are. In Wittgenstein’s eyes,

as I read him, Wright

’s claim that ‘for the community itself there is no authority, so

no standard to meet

’ can be, at very best, an attempt to say something that cannot

be said but only shown.

There is nothing in demand (D3) that need contradict McDowell, for the
request for explanation is not coming from

‘sideways on’, but from within

the community

’s existing language. In asking how our cognitive capacities

come to have property T, we need not presuppose that any Platonistic

McDowell,

‘Wittgenstein on Following a Rule’, Synthese 54 (1984): 353.

142

FROM MORALITY TO METAPHYSICS

‘rails of meaning’ are being ‘tracked’. All that is being presupposed is
that alongside human beings

—that is, the entities who perceive physical

objects, form beliefs, value things and act

—there exist physical objects and

norms of theoretical and practical (including moral) reasoning. These
objects and norms are

‘independent’ in precisely the sense that their being

such as they are is not constituted by the communal rati

ﬁcation of the judgement that

that is how they are. The question therefore arises of how human beings

’

cognitive capacities come to be sensitive to the reality and characteristics of
these norms and objects.

As I argued in Chapter 2, it is only through a teleological explanation

or something which explains the appearance of teleology (as natural selec-
tion does) that (D3) can be met. We saw in 6.2.3 that Aristotle

’s account

does this, but that it succeeds only because it embodies a worldview very
different from that of modern science. Aristotle is able to account for
the truth-tracking abilities of all these diverse human faculties precisely
because he takes nature itself to be expressive of a purpose.

With respect to perceptual capacities {a, b, c

. . . } and capacities

for theoretical reasoning {g, h, i

. . . } the explanation we are requesting can

be given without appeal to a purposive account of either nature
or the universe, through the quasi-teleological account offered by natural
selection. By contrast, it is clear that no such explanation is available for
the fact our moral faculties {m, n, o

. . . } also possess T. The explanation with

respect to the

ﬁrst two sets of capacities does not involve any attempt to make

‘sideways-on’ comparison between our cognitive capacities and an un-

synthesized

‘Given’. Rather, it arises from within evolutionary biology.

In the case of theoretical reasoning, the explanation runs as follows:

(P1)

The fact that {g, h, i

. . . } are the ‘bedrock’ of justiﬁcation for

theoretical reasoning does not itself guarantee that {g, h, i

. . . }

have property T.

However,

(P2)

Animals whose cognitive capacities did not have property T
would be less able to anticipate and negotiate the natural world
in which they are embedded.

To deny (P1) would commit us to anti-realism of the kind rejected by McDowell in his

discussion of Wright.

NATURAL GOODNESS AND

‘

SECOND NATURE

’ 143

Therefore,

(P3)

Possessing cognitive capacities with property T will confer a
selective advantage,

which explains why

(P4)

Human beings have developed cognitive capacities with pro-
perty T.

Our discussion in Chapter 2 made clear that natural selection never
provides the only level of explanation for a phenomenon. The process
of genetic mutation is random and blind. It is from amongst the range of
mutations generated by the underlying natural mechanisms that natural
selection favours those which maximize species survival and replication.

As (P1) to (P4) demonstrate, natural selection explains why humans

would evolve a capacity for basic kinds of theoretical reasoning. As we saw
in 2.4, a parallel story can be told of the development of our perceptual
systems, and why they would track the independent reality that is the
physical world. When William Paley invoked a divine designer to explain
the development of the eye, and when Dan-Erik Nilsson and Susanne
Pelger instead offered an evolutionary account, they were all responding
to a legitimate demand for an explanation of the reliability of human
perceptual capacities.

Nilsson and Pelger

’s work clearly satisﬁes (D3). The issue they are

addressing is neither how we justify faith in our perceptual capacities (as
D1 requests), nor simply how we give a natural history of the particular
stages of the development of homo sapiens (which would meet D2). In its
explanation of the possession by human perceptual capacities of the prop-
erty T, the evolutionary account, like Paley

’s before it, meets (D3). What

would otherwise be deeply puzzling as the product of a set of random
mutations is rendered intelligible as the product of a larger evolutionary
history in which the ability to track is selectively advantageous.

In this argument, I have been assuming that the presence of an answer to

(D3) with respect to our ability to detect physical objects and norms of
theoretical reasoning provides a prima facie reason for wanting (D3) to be

Dan-Erik Nilsson and Susanne Pelger,

‘A Pessimistic Estimate of the Time Required for

an Eye to Evolve

’, Proceedings of the Royal Society of London, Biological Sciences 256 (1994).

144

FROM MORALITY TO METAPHYSICS

answered in the moral case. Could McDowell not reject the parallel?
Why should there be a

‘one-size-ﬁts-all’ model for the explanation of

our cognitive capacities? What if McDowell were to reply as follows:

‘If

science in general, and evolutionary theory in particular, cannot explain
our capacity to detect moral norms, then so much the worse for its
pretensions to explain everything?

’

6.4.3 Where should explanation cease?

Where exactly we should cease asking for explanations, and accept that we
have reached

‘bedrock’, is indeed a difﬁcult question. But the impulse to

seek explanations is quite fundamental to human thought, and it seems
right to argue (as I did in Chapter 1) that Inference to the Best Explanation
is an indispensable practice. I now want to make the further argument that
Inference to the Best Explanation must have as one of its fundamental
principles the notion that there is a burden on anyone who refuses to offer
an explanation for any given explanandum to justify that refusal.

Unless the norms of practical reasoning are in some way less objective

than those of theoretical reasoning (something which McDowell is at pains
to deny), why should we take our acquisition of truth-tracking capacities
with respect to practical reasoning to be a brute datum (that is, an
explanandum which does not require a further explanation) when our
acquisition of truth-tracking capacities with respect to theoretical
reasoning is not?

Clearly, Inference to the Best Explanation cannot demand that all

explanatia that are advanced immediately become further explananda.
The chain of explanation must stop somewhere. However, it seems a
plausible principle that all phenomena either require an explanation or an
account of why it would be a mistake to seek a further level of explanation.
At no point is the bald response

‘Well why wouldn’t it be a mistake to seek

a further level of explanation?

’ going to be enough of an answer. When we

refuse further explanation, and assert that

‘we have reached bedrock’,

we need to be able to offer a justi

ﬁcation of that claim. That is, we need

to show why the request for a further level of explanation is misguided.

This is not a hugely onerous requirement. If McDowell is to refuse (D3)

in the moral case, he simply requires a plausible argument for the claim that

This objection has been pressed against earlier versions of this chapter by Stephen

Mulhall and Philip Krinks.

NATURAL GOODNESS AND

‘

SECOND NATURE

’ 145

‘here we have reached bedrock and no further explanation is appropriate’.
But this is precisely what he lacks. The only argument he had for thinking
(D3) inappropriate was his general claim that it represented an illegitimate
demand for a

‘sideways-on’ comparison between mind and world. I have

already dealt with that argument above, showing that McDowell is wrong
to think (D3) is incompatible with acceptance of his attack on the

‘myth

of the Given

’. More decisively, I have shown that (D3) can be met

with respect to theoretical reasoning. The claim that our demand for an
explanation for the truth-tracking properties of our belief-generating and
belief-evaluating capacities is illegitimate

—that it seeks an impossible ‘side-

ways-on

’ comparison—is surely decisively refuted by the fact the demand

can be met in other cases, namely those of physical perception and theoreti-
cal reasoning.

If McDowell were to concede this, could he not postulate a disanalogy

between our capacities for accurate theoretical reasoning and accurate
practical reasoning? Might this not provide a basis for resisting (D3) in
the one case and not the other? It is hard to see how such a dialectical
move could succeed, for there seem to be no good grounds for claiming
that there is such a disanalogy. In each case, the epistemology is the same;
namely that of re

ﬂective equilibrium and not of foundationalism. (This

was argued at 1.3.2.) Just as we only know what is practically reasonable
from the standpoint of virtue, so our knowledge of what is theoretically
reasonable comes from a standpoint located within the practice itself.
Indeed, McDowell relies on this very point in his argument against
moral anti-realists:

The rationality of virtue

. . . is not demonstrable from an external standpoint. But to

suppose that it ought to be is only a version of the prejudice [of foundationalism]. It
is only an illusion that our paradigm of reason, deductive argument, has its
rationality discernible from a standpoint that is not necessarily located within the
practice itself.

It is very hard to see any basis for arguing that (D3) is uniquely unnecessary
in the case of moral knowledge unless one thinks that

‘objectivity’ in the

moral case amounts to something less than it does in the case of theoretical
reasoning.

McDowell,

‘Virtue and Reason’, in his Mind, Value, and Reality, 71.

146

FROM MORALITY TO METAPHYSICS

6.4.4 Inability of McDowell

’s account to meet demand

I have now shown why I take (D3) to be a fair demand to make of our
capacity for moral knowledge, given McDowell

’s conception of morality

and its objectivity. Since McDowell agrees with my assertion (in
Chapter 2) that

‘that which is morally good and right’ is not identical to

‘that which is conducive to the survival and replication of the species’,
natural selection will not provide an adequate answer to the question.

We saw at 6.4.1 that McDowell seeks to explain the presence in humans

of cognitive capacities with respect to moral reasoning which have the
property T by reference to moral education. His claim is that it is only
from the perspective yielded by such an education that we will be able to
discern the requirements of virtue, and thus know where they diverge
from the biological imperative to survival and replication. This answers
(D2) but evades the prior question, essential to addressing (D3), of how
beings

ﬁrst managed to discern the moral values which through deliberate

processes of nurture and formation they now seek to pass on across the
generations. McDowell cannot answer that our moral sensibilities are
simply extensions of our ability to follow non-moral norms of reasoning.
For, as we saw earlier, he thinks the moral norms are not accessible to
someone with a perfect grasp of theoretical reasoning, but the sensibilities
of a sociopath. While we can say of particular agents that their moral
sensitivity is explained by a good upbringing, such explanations only meet
(D2) and not (D3).

Where does this leave us? My contention is that McDowell

’s position

fails to resolve the central dilemma identi

ﬁed by this book. Either it

generates the same

‘explanatory gap’ as full-blown objectivism in the

ways outlined above or it must water down the sense in which moral
truth is

‘objective’ to a point where ‘detection’ ceases to be an appropriate

image for what our cognitive capacities are doing in moral reasoning.

It is of course open to McDowell and those who share his approach to

reject the kind of objectivity I am taking moral statements to have. Might
there be an intermediate position vis-à-vis the objectivity of moral truth
which avoids both the objections I have made against Blackburn, Gibbard,
the earlier Scanlon, and Korsgaard in Chapters 3 and 4 and the

‘explana-

tory gap

’ identiﬁed in this chapter?

NATURAL GOODNESS AND

‘

SECOND NATURE

’ 147

Such a position is developed, along lines that are compatible with much

that McDowell writes, in the work of David Wiggins.

In the

ﬁnal section

of this chapter, I will consider whether his work develops McDowell

’s

account in a way that would enable McDowell to avoids my dilemma.

6.5 Wiggins: A Sensible Subjectivism?

On the face of it, David Wiggins

’ account of moral obligation seems to

meet the requirements of objectivity argued for in Chapter 1. For he
writes that morality presents us with

a requirement that applies regardless of inclination

. . . [w]hat we are apt to think is

that categorical requirements apply to you

. . . even if you ignore them and try to

renounce every concern whatever.

The reason Wiggins

’ position is of such interest at this point in the

discussion is that he is seeking the same reconciliation of apparently
incompatible claims as McDowell. That is to say, he too wants to combine
a commitment to some kind of objectivity in ethics with an ontology that
will not generate an explanatory gap. Where he differs from McDowell is
in his preparedness to give a fuller account of the distinctive nature of the
‘truth’ moral judgements can achieve.

It is not clear to me whether Wiggins

’ account is less objectivist than

McDowell

’s, or whether Wiggins is simply more willing to articulate these

issues. Either way, his position is important, for it offers someone convinced
by McDowell

’s moral epistemology and his attack on anti-realism a possible

defence against the demand for explanation which I have been pressing.
However, as I will argue, the price of escaping the explanatory gap yet again
is the inability to vindicate our fundamental moral commitments.

6.5.1 Morality as emergent

In the passage just cited, Wiggins is exploring whether it is possible to
reconcile a Humean account of moral sensibilities with a quasi-Kantian

David Wiggins,

‘Categorical Requirements: Kant and Hume on the Idea of Duty’, in

R. Hursthourse, G. Lawrence, and W. Quinn (eds), Virtues and Reasons: Philippa Foot and
Moral Theory (Oxford: Clarendon Press, 1996); and

‘A Sensible Subjectivism?’, ‘Truth,

Invention and the Meaning of Life,

’ ‘Truth as Predicated of Moral Judgments’ his and

‘Postscript III’, in Needs, Values, Truth (Oxford: Blackwell, 1987).

Wiggins,

‘Categorical Requirements’, 298.

148

FROM MORALITY TO METAPHYSICS

notion of the categorical nature of moral imperatives. Given the argu-
ments of previous chapters, this is a promising starting point, for we saw
(at 4.5.2) that a satisfactory meta-ethical theory needs both to make a place
for sensibility in its epistemology and to vindicate our pre-philosophical
sense of the absoluteness of the claim moral truths have upon us.

Crucially for our discussion, the picture Wiggins paints is one where

morality is emergent from the practices of the community:

Genealogically or aetiologically speaking, the public standard that informs our
evaluation of characters and sustains our understanding of the distinction of vice
and virtue is still an elaboration of natural benevolence. But it is crucially important
that that to which it has given rise reaches far beyond that original sentiment.

What claim do these moral standards and concerns have upon us? How
does the fact that, say, cruelty is condemned by these standards generate
categorical requirements upon us, irrespective of our feelings? Wiggins,
like Blackburn, notes that the fact that the genesis of a set of standards lies in
human feelings does not stop the standards applying to a given human
whatever he feels. He invites us to imagine an agent, John, who
is unmoved by the obligation to be kind to his son, and gives him a
stone when he asks for bread. How, on his account, can the bare fact that a
public standard condemns cruelty, generate a duty of kindness for John?

Why should it not? The standard is after all our standard. Does John want to escape
from our standard? Is this not the standard by which John effects most of the moral
distinctions he does want to effect?

. . . If he really does want to escape from its

judgments, then he must follow through. Does John want to say he does not care
what the standard says? If that is all that he says, why should moral philosophy
regard that as a problem? He does not care. (And, as we shall see, not even the
Kantian doctrine can do anything about that.)

. . . Or does John want to propose an

alternative standard that is not so soft on children? If so, let him propose it.

For reasons I have already given, I think Wiggins is right to warn us against
expecting philosophy to move agents who simply do not care about moral
norms. The purpose of meta-normative discussion is to establish what
reasons there are for action, not to guarantee that a sociopath will feel
compelled to act upon those reasons. So the signi

ﬁcant threat to his

account comes, not from the agent who would ignore normative reasons

Wiggins,

‘Categorical Requirements’, 303–4.

Wiggins,

‘Categorical Requirements’, 307.

NATURAL GOODNESS AND

‘

SECOND NATURE

’ 149

for action, but from the philosopher who queries what exactly such
reasons are.

6.5.2 The claim of morality

Wiggins describes the reasons why John must give his son bread as follows:
‘Even where [he] does not care, or he says he does not care, the actual
features of the situation that render his actions open to criticism are features
that engage with the attitudes (however

ﬁtful) that make him party to the

point of view that shall be in common to him with others.

’

If John is not otherwise a decent human being, it may be that our points

of view lack suf

ﬁcient commonality for the conversation to continue. The

natural way to take this is surely to say that if he lacks the relevant
sensibilities there is a genuine duty on him which he cannot be made to
see (in which case, we are back to the sociopath, whom we discussed
above). However, if the sentiments have the constitutive role in morals
which Wiggins seeks to give them, then problems arise not only at the
level of argument but of ontology. What is it that John fails to see, other
than that the (idealized) consensus of the community is against him? Surely
the force of that consideration only gives him a normative reason if there is
something independent about which the consensus of the community is
correct?

The outright objectivist will want to say that John is reason-bound to

act kindly. Wiggins wonders why:

‘We feel bound. But why reason-bound?

Why not say we feel bound by our moral nature, i.e. bound by those
moral sentiments without which

. . . we should not recognize ourselves?’

This response raises a number of issues. One thing Wiggins may be

resisting is an excessively self-conscious notion of moral reasons. The
classic example of this is given by Bernard Williams, when he invites us
to consider the case of a man who, on seeing two drowning people

—one

his wife and one a stranger. Williams argues that if the man, in deciding
what to do, justi

ﬁes his partiality by reference to a higher order moral

principle (rather than simply

‘because she is my wife’) he has had ‘one

thought too many

’.

Raimond Gaita gives a parallel case when he claims

that the wrongness of torture lies in its cruelty and barbarism, and

Wiggins,

‘Categorical Requirements’, 310.

Bernard Williams,

‘Persons, Character and Morality’, in his Moral Luck (Cambridge:

Cambridge University Press, 1981), 18.

150

FROM MORALITY TO METAPHYSICS

that philosophers who take the wrongness to lie in

‘the violation of an

objective rational order

’ have, at a deep level, missed the point.

It is tempting to use such arguments to drive a wedge between the claim

our

‘moral sentiments’ have upon us and the (‘Kantian’) notion that we are

‘reason-bound’ to save our spouses, abstain from torture and so on. But I want
to suggest the wedge being driven here between

‘moral sentiments’ or ‘moral

nature

’ and ‘reason’ depends on an ambiguity in the notion of a ‘reason’.

As we saw in 4.5.2, among the meanings of the word

‘reason’ are:

Reason

: a ground for a belief or action that is knowable by ratiocination

alone

—that is, knowable without any affective engagement

Reason

: a genuine ground for a belief or action.

Thus, one could oppose excessive reliance on reason

in discerning moral

truth and yet be wholly objectivist in thinking one

’s sentiments offer better

modes of access to reasons

for ethical belief.

Even if we agree with Wiggins and Gaita (against the moral rigorism of

Kant) that the morally sensitive agent responds in love to the need of his
wife, is repelled by the violence and pain of torture, and is better for acting
on these sentiments than merely upon reverence for the moral law, we
may also hold that in responding thus to his sentiments the agent is
endorsing their implicit claim to correctness. That the agent is

‘bound’

by something beyond the mere sentiment is revealed by the fact that his
sentiments are open to re

ﬂective modiﬁcation. Our moral sentiments

represent themselves as revealing to us the reasons

for action. They

contain an implicit claim to this kind of warrant, and so are answerable
to a standard of correctness beyond themselves.

An agent may come to see that some kinds of partiality to his wife are

inappropriate, or that under certain circumstances the resistance to causing
pain to another human needs to be overcome. Changing views on racial
segregation and slavery provide more fundamental examples of the ways
our views may be revised over time. The only way of making sense of such

‘Raymond Gaita on Torture’, webcast at <http://philosophybites.com/applied_ethics/

page/2/>

NATURAL GOODNESS AND

‘

SECOND NATURE

’ 151

ﬂective modiﬁcation, as Wiggins himself acknowledges, involves seeing

our moral sentiments as aiming at some kind of truth.

Our moral convictions

aspire to be correct (as contextually interpreted) while holding themselves answer-
able to any case anywhere. On these terms, we become open to the further re

ﬂection

that, by their nature, ethical systems have within them powers of regeneration,
reparation and renewal that will always invite the efforts of moralists, satirists and
other analogisers who strive to make their participants follow their ethical commit-
ments through. The objectivist

’s faith is this: that when, or if, participants do try to

follow through

. . . in this process, disagreement and conﬂict can diminish.

Returning to Wiggins

’ thought-example, it is hard to see how a wedge

can be driven between, on the one hand, an appeal to John

’s ‘moral

nature

’ and ‘moral sentiments’ and, on the other, the idea that he is ‘reason

bound

’ to give his son bread rather than a stone. What is meant by saying

he is

‘reason bound’ is not that he ought to be motivated out of reverence

for the moral law, nor that he can be compelled to act in the right way by
some watertight syllogism. Rather, to say that John is

‘reason bound’ to

give his son bread and not a stone is to say that the sentiments that move
him are appropriate and correct. They disclose genuine reasons for action,
and if they did not he would not be bound by them.

6.5.3 Objectivity and truth in ethics

Wiggins is happy to talk about

‘objectivity’ and ‘truth’ in ethics, but when

he does so his primary focus is on intersubjective agreement. Truth, as
predicated of moral judgements is, he writes

an objective property of a special subset of the sentences that are thrown up by, and
then survive in, certain processes of thinking and talking that can stand up against
criticism

—human criticism, that is, not criticism by a race of creatures whose focus

on the world was so different from ours that, even if each race became convinced
the other had a language, there would still be no clear prospect of mutually
transparent interpretation.

It is revealing that Wiggins regards the comprehensiveness of moral truth
as the key issue between him and the

‘outright moral realist’.

Wiggins

Wiggins, Ethics: Twelve Lectures on the Philosophy of Morality (Harmondsworth: Penguin,

2006), 353.

‘Postscript III’, in Needs, Values, Truth, 349.

‘Truth as Predicated of Moral Judgments’,in his Needs, Values, Truth, 163f.

152

FROM MORALITY TO METAPHYSICS

suggests that if an agent is faced with a dilemma and there is more than one
possible

‘thing to think’, this represents a problem for the moral realist. It is

not clear why this should be so: one could surely be an

‘outright moral

realist

’ and think that competing moral values lead to tragic choices, where

agents face irreconcilable claims of, say loyalty and benevolence and will
have some reasons for regret whatever choice they make. In some such
cases, it may be true that one choice is better than another. In other cases
the considerations on each side might objectively have equal amounts to
be said for them.

‘The world’ might then be experienced as ‘a void’ in that

there would be nothing more to be said on the matter, and yet a choice of
great moment might need to be made.

Where the

‘outright realist’ and Wiggins do part company is in thinking

that having certain

‘marks’ such as those enumerated in ‘Truth as

Predicated of Moral Judgments

’ could ever capture the heart of what

truth is. The marks of truth, the outright realist will wish to say, are rather
signs of our judgements matching a reality independent of them. They
cannot be constitutive of what truth is.

The problem for Wiggins is clearest in the case of counterfactuals.

Wiggins appears to endorse the Humean claim that

‘If human nature

had been different, then what we actually mean by morality

—that is,

morality

—might at many points not be fully intelligible to anyone.’

That claim might be read in a way that made it unexceptionable, for in
such an eventuality there would be some realm of reality which would
then (but not now) be missed by agents. We can imagine a form of life,
similar in many other respects to ours, but which considered causing
otherwise unjusti

ﬁed pain to weaker beings to be, all things considered,

good because it is entertaining. Such creatures would have missed an
important part of the moral landscape. However, it seems that by the
Humean claim, Wiggins means something more contestable.

The crucial question is whether, when Wiggins writes of seeing morality

as a

‘new creation’, he simply means that our civilization is a new creation,

arising as part of our response to the goodness in the world, or whether he
in the end sees goodness as a production of our

‘gilding and staining’ of a

previously value-free world.

If goodness

‘gilds and stains’, then we

Wiggins,

‘Categorical Requirements’, 308.

David Hume, An Enquiry Concerning the Principles of Morals (Oxford: Oxford University

Press, 1998), 163.

NATURAL GOODNESS AND

‘

SECOND NATURE

’ 153

cannot evade the questions I pressed against Blackburn, Gibbard, and
Korsgaard.

Firstly, the account seems to tie

‘truth’ to our contingent sentiments in

an inappropriate way. (This is the objection I raised against Blackburn and
Gibbard in Chapter 3.) In

‘A Sensible Subjectivism?’ Wiggins considers

this objection, and says it

would be dead right if the subjectivist were saying that

‘x is good’ may be

paraphrased as

‘x is such as to arouse or make appropriate [a certain] sentiment of

approbation

’, and if he were saying that this paraphrase could then be intersub-

stituted with

‘good’ salvo sensu (or more or less salvo sensu). But the subjectivist need

not be saying that. His distinctive claim is rather that x is good if and only if x is the
sort of thing that calls forth or makes appropriate a certain sentiment of approbation
given the range of propensities that we actually have to respond in this or that way; or
generalising a little

. . . his claim is that, for each value predicate ç (or for a very large

range of such) there is an attitude or response of subjects belonging to a range of
propensities that we actually have such that an object has the property

ç stands for if

and only if the object is

ﬁtted by its characteristics to bring down that extant

attitude or response upon it and bring it down precisely because it has those
characteristics.

Whatever other bene

ﬁt it has, this formulation does not help Wiggins’

subjectivism avoid the objections I pressed against Blackburn and Gibbard.
Wiggins writes as if our

‘propensities’ are static, but a key part of moral

ﬂection involves considering where they might be misguided. We bring

to bear a variety of techniques such as deductive argument, inference to
the best explanation and

‘critical theory’ (by which I mean a the process of

identifying and correct ideological biases which was discussed at 1.4.6),
along with the less ratiocinative activities of

‘moralists, satirists and other

analogisers

’ (cf. Wiggins’ earlier remarks) to reappraise these ‘propensities’

and consider whether they are indeed

‘appropriate’.

Wiggins

’ account seeks to combine the insights of both objectivism

and subjectivism, and it should be clear that I am in sympathy with this as
an epistemological project. But I am far less clear how this will help him
out of the ontological and explanatory issues I have been pressing against
secular moral theories. Wiggins talks of a spectrum of <property, response>
pairs, some of which (such as comedy and disgustingness) seem wholly
subjective, and others of which earn the predicate of

‘truth’. But, so far as

Wiggins,

‘A Sensible Subjectivism?’, 206.

154

FROM MORALITY TO METAPHYSICS

moral predicates are concerned, it seems that at each stage in this spectrum,
one or other of the objections I have been pressing will apply. Either the
property is so rooted in our subjective response that when we come to
revise our

‘propensities’ whatever we collectively determine to

be appropriate will turn out to be so (collapsing objectivity into mere
intersubjectivity, and leading to the same problems I urged against moral
quasi-realism in Chapter 3) or the

‘appropriateness’ of our responses is

something potentially distinct from the communal consensus, in which
case the

‘explanatory gap’ recurs.

Secondly, it is not clear that our moral sentiments can provide us with an

adequate motivation for action if they are taken to be

‘a new creation’

rather than a response to something absolute. (This is the question I raised
in the Introduction and pressed against Korsgaard and the early Scanlon in
Chapter 4.) Korsgaard tells us that if we violate our self-authored maxims
we are no longer able to think of ourselves under the description under
which we value ourselves,

ﬁnd our lives to be worth living and our actions

to be worth undertaking. Considering why John ought to give his son
bread, Wiggins says that our moral sentiments demand it, and without
these

‘we should not recognise ourselves’. The ‘bedrock’ both philoso-

phers offer for moral action seems curiously, and unsatisfactorily, self-
referential. It gets things the wrong way round. The reason why we
cannot live with ourselves, or consider life worth living, if we contradict
our moral sentiments is because we take them to be the claims of some-
thing absolute upon us. Our moral sentiments do not represent themselves
as speaking to us about a reality of which they are the components. They
have the force they do precisely because they present themselves to us as
representations of the way things are. If truth were a predicate to be
‘achieved’ by the ‘new creation’ which is morality, then it is hard to see
how morality could be

‘the claim of something absolute upon us’.

6.6 Conclusion

The secular accounts considered in Part II faced a common challenge: could
any of them. steer between the Scylla of insuf

ﬁcient objectivity (that is, a

position on which moral truth turned out to vary with our changing
subjective responses) and the Charybdis of the explanatory gap (that is, a
position on which it was impossible to see how human beings had acquired
belief-forming and belief-revising capacities which track the truth)?

NATURAL GOODNESS AND

‘

SECOND NATURE

’ 155

Part II began with the positions least equipped to avoid these twin

perils

—either (as with Blackburn, Gibbard, and Korsgaard) because of

their rejection of moral objectivism or (as with the later Scanlon) because
insuf

ﬁcient care was taken to explain how human capacities might come

to track reality.

The position considered in this chapter, in which John McDowell

’s

account of

‘second nature’ built on the more limited naturalism of Philippa

Foot, has provided us with the most convincing secular attempt to
negotiate safe passage between these two objections.

It should be clear (especially from the arguments advanced in Chapter 1)

how much my account is indebted to the position McDowell develops,
and to David Wiggins

’ advocacy of a combination of subjectivism and

objectivism in meta-ethics. I hope it is now also clear why I

ﬁnd McDo-

well

’s account of ‘second nature’ problematic. It cannot explain how we

acquire our capacity to detect genuine moral realities rather than selec-
tively advantageous imitations. While McDowell is generally suspicious of
requests for

‘sideways-on’ accounts of how our capacities match reality,

I have been at pains to show that this request for explanation is quite
legitimate. It is not a request for some vindication of our logos from an
Archimedean point, but for the kind of explanation supplied by evolu-
tionary theory in the case of many of our other capacities. When such an
explanation is available for our capacity to track truth in theoretical
reasoning, why is it unacceptable to seek an explanation for our capacity
to track truth in the moral realm?

This chapter completes my discussion of secular accounts of moral

knowledge. I have sought to do more than critique the work of particular
philosophers. Rather, I have explained why secular accounts face a general
problem. We can see this with reference to the Figure 2 at the start of Part
II. Each of the accounts between the dotted lines either responds to the pull
of ontological parsimony, in which case it falls foul of the argument from
deliberative indispensability made in Chapter 1 or it responds to the pull of
objectivism and falls foul of the explanatory gap. Our discussion of
McDowell makes this dilemma particularly vivid. The only way to read
his position which might avoid the explanatory gap is to interpret him as
following Wiggins in ascribing to morality a more tentative objectivity. As
I have argued, this more tentative objectivity then falls foul of the criti-
cisms advanced against Blackburn and Korsgaard in Chapters 3 and 4.

156

FROM MORALITY TO METAPHYSICS

In Part III, I will advance an alternative, non-secular position, which

closes the

‘explanatory gap’ by recognizing ascribing a purpose to the

universe. Having concurred with so much that Foot and McDowell
have had to say about the importance of Aristotle for moral epistemology,
I will now argue that we cannot simply ignore the fact that for Aristotle
this went with a purposive view of the universe as a whole. It is to such
purposive accounts that we now turn.

NATURAL GOODNESS AND

‘

SECOND NATURE

’ 157

This page intentionally left blank

PART III

Theism

This

ﬁnal section of the book will defend the claim that non-secular (i.e.

purposive) accounts of the universe can explain why our cognitive facul-
ties are capable of tracking moral truth. It will then consider what support
this gives to apologetic arguments for theism.

Chapter 7 will provide a positive argument for theism; arguing that it is

capable (in a way secular accounts are not) of bridging the explanatory gap.
I will then consider the implications of this

ﬁnding for the overall credence

that should be given to theism. Chapter 8 will consider two rival forms of
purposive explanation, the axiarchic account given by John Leslie and
Hugh Rice and a more modest form of Neoplatonism. I will offer a
general argument against axiarchism as a valid form of explanation, and
will argue other forms of Neoplatonism offer less a satisfying explanation
of the human capacity for moral cognition than classical theism. (I am
using the term

‘classical theism’ to describe belief in a self-subsistent,

omnipotent, omniscient, benevolent, and personal deity.)

Finally, in a brief Conclusion, I will consider the part that this kind of

argument plays in a wider Christian apologetic.

This page intentionally left blank

From Goodness to God

Closing the Explanatory Gap

7.1 Revisiting Teleological Explanation

At 2.4, I argued that any satisfying explanation of the human capacity for
moral cognition will be teleological rather than nomothetic (an explana-
tion given in terms of causal laws). Teleological explanation is invoked
rather less in our own age than in previous centuries, and one reason for
this is the attractiveness of natural selection. Evolutionary biology and
socio-biology offer explanations with many of the advantages of the
teleological form (accounting for a chain of events in terms of the state
of affairs which result) whilst remaining wholly secular (in that they do not
have to postulate either an agent, or any inherent purpose in the natural
order).

At 2.4.2, we considered a case in point, in our discussion of the

displacement of William Paley

’s religious account of the human capacity

for perceiving physical objects by the evolutionary explanation given by
Dan-Erik Nilsson and Susanne Pelger. Nilsson and Pelger

’s mathematical

modelling showed that, assuming only random genetic mutations, selec-
tive pressures were likely to result in a light-sensitive patch on an animal
mutating over time into an eye. Thus, they are able to explain why a
system will tend towards a telos without either Paley

’s benevolent designer

or natural teleology of the kind we see in Aristotelian biology.

The

ﬁrst two parts of this book have shown why natural selection

cannot offer an explanation of our capacity for moral cognition, and
why nomothetic explanation is also unsuited to this task. Only two

Cf. the citation at 6.2.3 from Aristotle

’s On the Parts of Animals.

alternatives remain: accepting that the phenomenon in question cannot be
explained or turning to a more authentically teleological account.

As Charles Taylor reminded us:

‘To offer a teleological explanation of

some event or class of events

. . . is . . . to account for it by laws in terms of

which an event

’s occurring is held to be dependent on that event’s being

required for some end.

’

In demonstrating that a system has a tendency of

this kind we have not yet explained either why it has that tendency or why
we should allow the bare fact of the tendency to be accepted as the
endpoint of the explanatory process. Because there are differing explana-
tions of the tendency, and indeed differing attitudes to whether all
tendencies need explanation, modern writers have tended to distinguish
natural teleology (of the kind Aristotle invokes in On the Parts of Animals)
from intentional action.

In the case of intentional action, the agent is the

‘system’ with the

relevant

‘tendency’, and the agent’s tendency is explained in terms of

their desire to achieve a certain goal, G. At 2.4.1 I argued that a compre-
hensive intentional explanation also needed to explain the agent

’s desire

for G and her ability to do X which she knows will bring about G. In the
simplest case, a goal G is of objective worth, the agent knows this to be so,
the agent pursues G because of its value, and the agent has the power
through X-ing to bring G about.

What, then of teleological explanations which make no reference to an

agent? Rowland Stout

’s view is that all genuinely teleological explanation

must postulate an agent. Without such intentional action, we have no
explanation of why the value of the goal leads to it being realized.

ﬁrst

blush this seems a plausible claim: whereas natural selection provides us
with a mechanism by which the value (to a species) of a goal brings about
those states of affairs which promote that goal, it does not seem illuminating
to say simply to explain a species

’ possession of teeth by the contribution

teeth make to the

ﬂourishing of the species.

However (as we saw at 6.2.3) Aristotelian biology explained states of

affairs in precisely the way Stout excludes.

I will consider these two alternatives in turn:

ﬁrstly (in this chapter)

developing and evaluating an explanation of our moral cognition which is

Charles Taylor, The Explanation of Behaviour (London: Routledge & Kegan Paul, 1964), 9.

Rowland Stout, Things That Happen Because They Should: A Teleological Approach to Action

(Oxford: Oxford University Press, 1996), 83.

162

FROM MORALITY TO METAPHYSICS

agent-based (theism) and then (in Chapter 8) considering axiarchic
explanation, which invokes genuine teleology but need not involve any
agent or personal intention.

7.2 Theism as an Agent Explanation

At 2.4.1, I offered a schema for a fully satisfying intentional explanation of
a state of affairs Y:

(1)

Agent X knew the state of affairs Y would be objectively valuable

therefore

(2)

Agent X had good reason to bring about Y

which explains why

(3)

Agent X wanted to bring about Y.

(3), taken together with

(4)

Agent X knew that the best means to bring about Y involved
actions {a, b, c

. . . }

and

(5)

Agent X had the power to take actions {a, b, c

. . . }

explains why

(6)

Agent X took actions {a, b, c

. . . }

which explains, given (4), why

(7)

Y occurred.

I called this the

‘classic’ form of intentional explanation because any

explanation that omits (1) and (2) is incomplete. Agents are not merely
pushed and pulled by brute desires. Rather, agents

’ desires can be revised

and reformed in the light of their understanding of that which is truly
worthwhile. An agent might want to bring about Y without having good
reason to, and arguably an agent might have good reason to bring about Y
even if it was not objectively valuable. But either of these would require
further explanation, whereas on the

‘Apollonian’ account of ethics which

FROM GOODNESS TO GOD

163

I have defended, no further explanation is required for the moves from (1)
to (2) and from (2) to (3).

The phenomenon which concerns us in this chapter is the presence in

humans of belief-generating and belief-evaluating capacities which track
the objective moral truth. If we term this phenomenon

‘Y ’, then the

explanation offered by classical theism will be as follows:

)

God knew the occurrence of Y would be (objectively) valuable

therefore

)

God had good reason to bring about Y

which explains why

)

God wanted to bring about Y.

), taken together with

)

God knew that the best means to bring about Y involved actions
{a, b, c

. . . }

and

)

God had the power to take actions {a, b, c

. . . }

explain why

)

God took actions {a, b, c

. . . }

which explains, given (4

’), why

)

Y occurred.

This chapter will not offer a general defence of classical theism. Its claim is
much more limited: that the position can avoid the

‘explanatory gap’. If

we take classical theism to be a coherent position (an issue which is also
beyond the scope of this book), there remain two potential problems with
(1

) to (7

) as a potential explanation.

The

ﬁrst problem concerns (1

). It is not immediately obvious that

it makes sense to assert God knows an occurrence to be

‘(objectively)

valuable

’. If the statement is more than trivially true, it suggests that

value is independent of God

’s will. This might be thought an unacceptable

164

FROM MORALITY TO METAPHYSICS

encroachment on the omnipotence which theism usually ascribes to God.
Such an objection is put forward by Hugh Rice, who argues that those
who believe in a personal God face an unpalatable dilemma: either they
must accept a (morally repugnant) divine command theory or they end up
with a deity who is something less than the omnipotent God of orthodox
Christianity, Judaism, or Islam.

The second problem lies in the assumption, made in the transition from

) to (3

), that God will always will the good. Why should agent

explanation lead us to ascribe any greater degree of goodness to a creator
than we see manifest in the universe? If the universe is morally ambiguous,
do we not have reason to ascribe the same ambiguity to its maker? This
objection is made by David Hume, and more recently re-presented by
Stephen Law.

In the sections which follow, I will consider each of these objections

in turn.

7.2.1 God and Objective Goodness

The

ﬁrst objection, which takes issue with (1

), claims that any satisfactory

account of a personal deity must take objective value to be dependent
upon the will of God. The attraction of such a divine command theory of
ethics is that it ensures the omnipotence of God is in no way compro-
mised. Its disadvantage is that it generates some deeply unattractive
counter-factuals.

An account which grounds moral truth entirely in God

’s will is com-

mitted to the claim that:

(8)

If God had willed that babies ought to be tortured, then it would
have been good that babies be tortured.

It is hopeless for the advocate of an unquali

ﬁed divine command theory to

reply that God would never will such an evil thing. (I will consider Robert
Adams

’ more qualiﬁed version of the theory later in this section.) The

torture of babies is, on this account, only evil because of what God has

Hugh Rice, God and Goodness (Oxford: Oxford University Press, 2000), 64

–72.

David Hume, Enquiry Concerning Human Understanding, in his Enquiries, ed. L. A. Selby-Bigge,

3rd edn (Oxford: Oxford University Press, 1975), 137; Stephen Law,

‘The Evil God Challenge’,

Religious Studies 46 (2010): 353

–73.

FROM GOODNESS TO GOD

165

willed. If God had willed something else, there would be a corresponding
change in the moral truth.

As Hugh Rice observes, a possible line of response for the divine

command theorist is to accept the counter-factuals and to point out that
they only apply because of what God has in fact willed:

It is awful just because God wills that such a thing should not happen; and it seems
awful because God has made it seem so to our minds. What we fail to notice is that
God could have made it all right, and could have made it seem so to us.

I think he is right to dismiss this move:

It is not just that it seems to us that it is awful to torture babies: it seems to us that it
could not have been otherwise. And this is inconsistent with what the reply claims.
What I am appealing to, of course, is the necessity of basic facts about value

. . . And

the objection to the reply is that it represents a supposed basic fact about value as
contingent

—contingent on God’s will.

This parallels the argument offered in Chapter 3: the unquali

ﬁed divine

command theory generates the same kind of (morally) unacceptable
counter-factuals as quasi-realism. In one respect, the divine command
theory is on stronger ground than quasi-realism. Our primary reason for
rejecting quasi-realism was that it did not give moral truths the kind of
independence from human deliberation that I have argued (in Chapter 1)
to be deliberatively indispensable. By contrast, the divine command theo-
ry does give us the objectivity we require for practical deliberation. For
human beings engaged in moral deliberation, there is (on that theory) still
a

‘fact of the matter’ to deliberate about, independent of their sentiments

or beliefs.

However, the kind of fact these divine commands generate is not going

to provide a motivation to moral action. That an unloving, misanthropic
deity should command us to torture babies does not provide a moral
reason to torture babies, even if his power provides a prudential reason
to do so. It is hard to see how the unquali

ﬁed divine command theory is

anything more than a theory on which we must obey an omnipotent
tyrant.

Those who advocate a straightforward divine command theory will

have to reckon with a further consequence of the position. If the divine

Rice, God and Goodness, 69.

166

FROM MORALITY TO METAPHYSICS

command theory is correct, God

’s actual decision to will that baby-killing

is wrong cannot be attributed to his goodness. Whatever God had willed,
even if it was that baby-torture was obligatory, would have quali

ﬁed

equally well for these same attributes. In consequence, the language
worshippers use to praise God turns out to be no more than well-disguised
obeisance before a deity whose fundamental character is power

—for God

is no more good than he would have been had he loved torture and
cruelty. Even more than quasi-realism, this theory deserves Stig Alstrup
Rasmussen

’s verdict: no nice people should endorse it, because it makes

moral truth dependent on inappropriate contingencies.

As I indicated above, Hugh Rice wants to press the following dilemma

against a personal conception of God: either it leads to the (morally
repugnant) divine command theory or it implies a deity who is something
less than the omnipotent God of orthodox theism. It should now be clear
why I agree with the

ﬁrst half of his claim. The unqualiﬁed divine

command theory is indeed repugnant, and goes against some of our
most fundamental moral commitments.

What of the second half of the claim? If moral value does not depend on

the will of God, is his omnipotence thereby limited? Richard Swinburne
offers an argument to the effect that the independence of the moral order
is no threat to divine omnipotence. This is so, he claims, precisely because
no power could make moral truths otherwise than they are. Whether or
not they believe that morality is independent of God

’s will, theists must

accept that there are other kinds of truth which are free-standing, most
obviously the truths of logic. As he writes:

It may be objected that in order to be truly omnipotent, a person should be able to
do not merely the logically possible, but the logically impossible as well. This
objection

. . . arises from regarding a logically impossible action as an action of one

kind on a par with an action of another kind, the logically possible. But it is not.
A logically impossible action is not an action. It is what is described by a form of
words which purport to describe an action, but do not describe anything which it is
coherent to suppose could be done.

As we saw at 2.5.4, Swinburne makes the claim that moral truths are
analytic truths. In that section, I explained why I

ﬁnd that unsatisfactory. It

Stig Rasmussen,

‘Quasi-Realism and Mind-Dependence’, Philosophical Quarterly 35

(1989): 187.

R. G. Swinburne, The Coherence of Theism (Oxford: Clarendon Press, 1993), 149.

FROM GOODNESS TO GOD

167

is more plausible to take moral truths to be metaphysically, but not
logically, necessary. The problem with this revision of Swinburne

’s views

is that it may seem as if the existence of necessary truths which are
independent of God

’s will presents a more substantial limitation on divine

omnipotence than the laws of logic. As Swinburne argued above, the laws
of logic do not rule out anything that it would be coherent to suppose. By
contrast, the converse of a necessary truth may well be coherent. (Indeed,
at 2.5.4 my argument at was precisely that the converse of a moral truth is
often coherent, and hence moral truths cannot be analytic.)

Hugh Rice argues that the claim moral truths are necessary, non-

analytic and independent of God

’s will places an unsatisfactory limit on

divine omnipotence:

The trouble with this answer

. . . is, one might think, that it does not so much solve

the problem [of reconciling God

’s sovereignty with the independence of moral

truths] as declare it insoluble. It amounts to saying that we have no choice but to
concede that God

’s role would be bound to be secondary in importance to the facts

about good or bad. His role is bound to be that of mediator only; he could not be
sovereign. It amounts to saying that one of the key features in many people

’s

conception of God is incoherent.

This seems an unduly harsh assessment. How do we judge the

‘key features in

many people

’s conception of God’? Until people think deeply about the

implications of different views of divine omnipotence, they may not have
particularly detailed views about what it involves. In my experience, many
believers think God created the laws of logic until the implications of this view
are pointed out. Their view then changes without great worry or excitement.

The argument I am advancing here does not rule out all versions of the

divine command theory. Precisely because of the objections we have been
considering above, Robert Adams has developed a

‘modiﬁed divine

command theory

’. On this account, ‘ethical wrongness is (i.e. is identical

with) being contrary to the commands of a loving God.

’ Adams justiﬁes

this modi

ﬁcation as follows:

If wrongness is simply contrariety to the commands of God, it is necessarily so,
which implies that it would be wrong to disobey God even if he were so unloving
as to command the practice of cruelty for its own sake. That consequence is
unacceptable. I am not prepared to adopt the negative attitude towards possible

Hugh Rice, God and Goodness, 65.

168

FROM MORALITY TO METAPHYSICS

disobedience in that situation that would be involved in identifying wrongness
simply with contrariety to God

’s commands. The loving character of the God who

issues them seems to me therefore to be a metaethically relevant feature of divine
commands.

Adams

’ account is broadly compatible with the schema for intentional

explanation which I am defending. We would simply have to recast (1

)

and (2

) as follows:

)

God knew that Y was the most loving action to perform

and

)

God is loving, and so performs the most loving actions.

If central aspects of God

’s character are necessary (that is to say, they are

expressive of his nature rather than determined by his volition) then it may
be possible to ground morality in God

’s character rather than his will, and

so avoid the sense that moral truth is somehow an independent constraint
on God.

It seems plausible to argue that benevolence is at least as

‘key [a]

feature in many people

’s conception of God’ as omnipotence or omni-

science. If the moral order is grounded in the divine character rather than
the divine will, and if the benevolence of God is a necessary truth, then
classical theism will avoid Rice

’s dilemma. For this position both afﬁrms the

ontological dependence of objective value on God and denies that the
moral order would be substantially different in any possible world.

There are, then, two distinct ways of avoiding Rice

’s objection. The

ﬁrst is to accept that moral truths are necessary and are ontologically
distinct from God. As we have seen, this may seem at

ﬁrst blush to be at

variance with most believers

’ conceptions of God. However, I have ar-

gued that it is fully compatible with an af

ﬁrmation of divine omnipotence.

Robert Adams,

‘Divine Command Metaethics Modiﬁed Again’, in The Virtue of Faith

and Other Essays in Philosophical Theology (Oxford: Oxford University Press, 1987), 139

–40.

Cf. also the discussion in John Hare, God and Morality: A Philosophical History (Oxford:
Blackwell, 2007), 260f.

In Finite and In

ﬁnite Goods: A Framework for Ethics (Oxford: Oxford University Press,

1999), Robert Adams defends the view that goodness is constituted by resemblance to a
loving God, while rightness is constituted by obedience to what such a God commands.

Robert Adams discusses the notion of metaphysical necessity as applied to God in

‘Divine Necessity’, in his The Virtue of Faith, 209–20. Rice rejects the notion that the
existence of God, let alone his loving character, might be metaphysically necessary in his
Finite and In

ﬁnite Goods, 70: ‘it seems, if true, contingently true.’

FROM GOODNESS TO GOD

169

Omnipotence can only involve doing anything that is possible, and
changing necessary truths is (by de

ﬁnition) impossible. The second way

to avoid Rice

’s objection is to ground morality in the loving nature

of God.

It is beyond the scope of this book to determine which of these accounts

is to be preferred. What matters for the purposes of our argument is that
either account enables the classical theist to vindicate our fundamental
moral commitments and to explain the ability of humans to cognize the
objective moral order.

7.2.2 Agent explanation and divine goodness

Let us turn to the second objection, which concerns the transition in our
explanation from propositions (1

) to (3

). This transition depends on the

assumption that God is good. Our second objector will ask why we should
ascribe a greater degree of goodness to a creator than is manifest in the universe.

This line of attack receives its classic formulation in David Hume

’s

Enquiry Concerning Human Understanding:

Allowing, therefore, the gods to be the authors of the existence or order of the
universe; it follows, that they possess that precise degree of power, intelligence and
benevolence, which appears in their workmanship; but nothing farther can ever be
proved, except we call in the assistance of exaggeration and

ﬂattery to supply the

defects of argument and reasoning

. . .

The knowledge of the cause being derived solely from the effect, they must be

exactly adjusted to each other; and the one can never refer to anything farther, or
be the foundation of any new inference and conclusion.

The general objection being made here is that a morally ambivalent deity
provides a more obvious explanation of a morally ambiguous creation than
the benevolent deity of classical theism. Even if theodicies can render an
ambiguous creation compatible with a benevolent deity, Hume

’s argument

is that it is not the best explanation.

In a recent paper, Stephen Law builds on Hume

’s argument.

advances the

‘evil god hypothesis’—arguing that, given the morally

David Hume, Enquiry Concerning Human Understanding, 137.

Stephen Law,

‘The Evil God Challenge’, acknowledges a debt not only to Hume but to

C. Madden and P. Hare, Evil and the Concept of God (Spring

ﬁeld, IL: C. Thomas, 1968);

S. Cahn,

‘Cacodaemony’, Analysis 37 (1976); E. Stein, ‘God, the Demon, and the Status of

Theologies

’, American Philosophical Quarterly 27 (1990); and C. New, ‘Antitheism’, Ratio

6 (1993).

170

FROM MORALITY TO METAPHYSICS

ambiguous qualities of his creation, there is as much evidence for the
hypothesis that the creator is evil as that he is good.

Suppose that the universe has a creator. Suppose also that this being is omnipotent
and omniscient. But suppose he is not maximally good. Rather, imagine that he is
maximally evil. His depravity is without limit. His cruelty knows no bounds. There
is no other god or gods

—just this supremely wicked being. Call this the evil god

hypothesis.

How reasonable is the evil god hypothesis?

. . . [C]ertainly in their simplest

versions, many of the most popular arguments for the existence of God fail to
provide any clue as to our creator

’s moral character. In which case, to the extent

that they support the good god hypothesis (i.e. not very much), they also support
the evil god hypothesis.

Law argues that all the ways in which traditional theodicies reconcile the
existence of evil with a benevolent deity have mirror-images by which the
existence of good can be reconciled with the

‘evil god hypothesis.’ His

claim is that an evil god has the same explanatory power as the benevolent
God of classical theism. Just as theodicies can reconcile the evil in the
world with a benevolent deity, Law claims, his

‘reverse theodicies’ can

reconcile the goodness and beauty in the world with a malevolent deity.

If the reverse theodicies are as plausible as the classical theistic ones (and

hence the hypothesis of a malevolent creator has the same explanatory
power as that of a benevolent deity), Law takes this as a reason for
dismissing the latter hypothesis, not entertaining the former one.

How persuasive are our

. . . reverse theodicies? Intuitively, not at all. Rather than

being taken seriously, they usually provoke amusement among theists and non-
theists alike. But this raises the question: if the reverse theodicies are highly
unpersuasive

—if, indeed, we consider them something of a joke—why should

we be expected to take the standard theodicies anymore seriously?

Much of Law

’s argument falls outside the scope of this book. Our concern

is with any adverse impact Law

’s argument might have on the validity of

our inference from (1

) through (7

)

—and the threat is to the move from

) via (6

) to (7

It is only if God is good that we can infer from

)

God knew the occurrence of Y would be (objectively) valuable

Stephen Law,

‘The Evil God Challenge’, 363.

Stephen Law,

‘The Evil God Challenge’, 5.

FROM GOODNESS TO GOD

171

that

)

God had good reason to bring about Y

and, given (2

), that

)

God wanted to bring about Y.

Law

’s argument is that an evil God might have malevolent reasons to give

human beings truth-tracking capacities with respect to ethics:

While it might be true that only a supernatural being is capable of furnishing us
with a moral sense, an evil god might well also have an interest in providing it. By
providing us with both free will and knowledge of good and evil, an evil god can
allow for the very great evil of our freely performing evil actions in the full
knowledge that they are, indeed, evil.

As with his wider argument, Law advances this as a reductio ad absurdum. He
takes it to undercut the argument of this book, for an evil God would
explain our truth-tracking moral faculties every bit as well as a benevolent
one.

I think Law overstates his case when he claims explanatory symmetry

between classical theism and the

‘evil god hypothesis’. For there is a

signi

ﬁcant asymmetry here: on the assumption of a benevolent deity,

it is very clear that we would expect our belief-generating and belief-
evaluating faculties to track the truth. That is to say, it is evident why a
good God would want his creatures to know right from wrong. By
contrast, if there were a malevolent deity, it is far less clear what faculties
he would have reason to bestow upon his creatures. We have no reason to
think that such a god would enjoy the

‘great evil of our freely performing

evil actions

’ any more than, say, the great evil of us believing evil things to

be good.

At 3.1.1 we considered Simon Blackburn

’s contrast between Apolloni-

an and Dionysian accounts of moral psychology. In the rest of that chapter,
and in the discussion of reason and desire at 4.5.2, I made the case against
the (Dionysian) claim that

‘we talk of reasons to reﬂect the fact that we

already care.

’ I argued that the opposite is true: we care because there are

reasons to care.

This suggests an explanatory asymmetry between

(9)

God brought about Y because it was good

172

FROM MORALITY TO METAPHYSICS

and

(10)

God brought about Y because it was evil.

On the Apollonian account I have defended, there is an immediate
intelligibility about any agent, human or divine, seeking to bring about
Y because it is good. For the very fact that Y is good gives an agent a reason
to do it. By contrast, the fact Y is evil gives an agent a reason not to do it. If
we ask of any agent, human or divine,

‘Why did the agent bring about Y?’

an answer which invokes Y

’s goodness and the agent’s knowledge of its

goodness leaves no further mystery about motivation. By contrast, if Y was
known by the agent to be evil, a further level of explanation is required.
For in saying that it was evil, we have identi

ﬁed a factor which makes it less

explicable that the agent should have performed the action.

This is not, of course, to say that agents may not deliberately perform

evil acts. It is merely to note that there is an explanatory asymmetry. The
fact that an act is known to be good provides a motive for its performance,
whereas the fact that it is known to be evil means we need a further
explanation for its performance, perhaps in terms of the agent

’s appetites.

In 7.2.1, we considered a range of accounts of the relationship between

God and moral value. Whichever one the classical theist adopts, there is a
clear prima facie reason why God would do what is good. Either this is
because God

’s will (or nature) is deﬁnitive of what is good or because there

is an independent standard of moral value which God (like us) has reason to
obey. If the former is true, the

‘evil god hypothesis’ is incoherent. If the

latter is true, the hypothesis has more limited explanatory power than
classical theism. For if moral value is independent of God, then in saying
an act is evil, we are giving a prima facie reason why God would not perform
it. It may be that a deity performs evil actions because they achieve some
other end he has

—they might, for example, entertain him. But this hypoth-

esis seems rather ad hoc, whereas (on the Apollonian) account, the fact that
God would delight in the good requires no further explanation.

None of this shows that the

‘evil god hypothesis’ is, more generally,

incoherent. The hypothesis that the deity is (or deities are) morally
ambiguous is one that a signi

ﬁcant proportion of the human race has,

for substantial periods, accepted. And there are versions of agent-expla-
nation other than classical theism which would succeed in closing the
explanatory gap. However, for the reasons I have given, classical theism

FROM GOODNESS TO GOD

173

seems particularly well-placed to do so. So, having noted the possibility of
more morally ambiguous forms of theism, in the remainder of the discussion
I will focus on accounts on which God (or an impersonal creative force) is
benevolent.

7.2.2 Agent explanation and natural selection

Having defended the explanation given in (1

) to (7

) against these two

objections, I need to address a further question. What is the relationship
between the kind of explanation I have outlined and the account an
evolutionary biologist would give of our moral knowledge? It may seem
that the explanation I have been advancing is incompatible with the
evolutionary explanations discussed in Chapter 2.

It is worth recalling where those accounts left us. At 2.5.3, I argued that

(on the secular account) the only bit of our moral thought whose accuracy
was explicable was that which tracked collective survival and replication.
From the theory of natural selection (whether in its

‘Panglossian’ adapta-

tionist form, or in the more modest version outlined by Gould and
Lewontin) we have no explanation for the verisimilitude of those of our
intuitions and principles which con

ﬂict with the maximization of species-

replication. A theistic explanation of the emergence of moral knowledge
need not con

ﬂict with a version of the theory of natural selection. The

theist need not deny that the processes described by evolutionary biology
explain the generation of human convictions about ethics, and their
capacity to reason about these convictions and re

ﬁne them. All that the

theist needs to add to the account given by evolutionary biology is the
claim that the world is providentially ordered so that the interaction of
the quasi-teleological process of natural selection and of the spandrel-like
features it generates yield an outcome which enables human beings to
apprehend that which is of objective value.

There is a direct parallel between this explanation and the explanation

which classical theism gives of the consciousness friendly nature of the
universe

’s fundamental physical constants. It appears that, for a number of

fundamental physical constants {a, b, c

. . . }, only an extremely restricted

range of values of {a, b, c

. . . } would permit the existence of living

conscious beings. (Planck

’s constant is the most famous example of this

phenomenon.) The existence of a benevolent God is held by many theists
to explain the harmony between the actual values of {a, b, c

. . . } and the

174

FROM MORALITY TO METAPHYSICS

values necessary to sustain conscious life.

Such theists are not offering an

account of the origin of the universe which need compete with that of
physics. Rather, they are claiming that the physical system has surprising
properties which are better accounted for if the scienti

ﬁc account is

complemented by a higher-order account in which a benevolent creator
orders the world so as to achieve certain objectively valuable outcomes.

I am making this same claim with respect to the development of belief-

generating and belief-evaluating capacities in human beings. The scienti

ﬁc

story will not explain why our moral capacities track the truth. The theistic
story is a higher-level account which explains why our capacities have this
additional property.

7.3 The Explanatory Gap: An Argument

for Theism?

I indicated (at 2.1.3) that this book would assume explanatory realism.
That is to say, I would be assuming that there was a fact of the matter about
the truth or falsehood of

(11)

The obtaining of state of affairs Y would be a good explanation
for phenomenon X,

and also, quite separately about the truth or falsehood of

(12)

The obtaining of state of affairs Y is the reason for X.

The phenomenon for which we are seeking explanation here is the
capacity of human beings to cognize an objective moral order. The secular
moral theories considered in Part II were unable to explain this phenom-
enon, and I have advanced reasons for doubting whether any future
secular account will do any better. In this chapter I have argued that
classical theism can provide a good explanation for this phenomenon.
(As I have noted, it is not the only form of purposive explanation. In
Chapter 8 will argue classical theism is also preferable to teleological
accounts which lack an intentional agent.)

B. J. Carr and M. Rees,

‘The Anthropic Cosmological Principle and the Structure of the

Physical World

’, Nature 278 (12 April 1979): 605–12; and R. Collins, ‘Evidence for Fine-

Tuning

’, in Neil A. Manson (ed.), God and Design: The Teleological Argument and Modern Science

(London: Routledge, 2003), 178

–99.

FROM GOODNESS TO GOD

175

Let us suppose that a secular thinker accepts the entire argument of this

book. That is to say, he agrees that secular accounts of ethics cannot avoid
the problems discussed in Chapter 2, and he shares our misgivings about
the other forms of purposive explanation. What reason does that give him
for accepting the doctrines of classical theism?

It is open to the secular theorist both to accept that classical theism

provides the best explanation for the human acquisition of truth-tracking
moral capacities and to deny that any such explanation is true. The secular
theorist may think the objections to all possible purposive accounts on
offer are so signi

ﬁcant that it is better to accept (i) an unlikely accidental

correlation between our belief-generating and belief-evaluating capacities
and the objective moral order than (ii) classical theism, (iii) the hypothesis
of a morally ambivalent deity, or (iv) a causally ef

ﬁcacious, non-personal

Good.

As we have seen, there is an analogy between the moral

‘explanatory

gap

’ and challenge of explaining the universe’s fundamental physical

constants. Discussing the case of the fundamental physical constants,
and its impact on theistic apologetics, John Hawthorne claims the key
issue is the balance of

‘credences’. (It is hard to make sense of the notion

of the probability of there being a God

—especially as many classical theists

would assert that God is a necessary being. But it does make sense to
talk of the credence which it is rational to give the hypothesis that there is
a God.)

Hawthorne invites us to

‘[t]hink of someone’s pattern of credences as a

pie, with the size of credence in some proposition corresponding to the
size of some associated piece (relative to the whole pie

’ New evidence,

which rules out certain possibilities, knocks out those slices of the pie.
‘Standard updating is tantamount to expanding the pieces that are left
into a whole pie in a way that preserves the ratios between the pieces that
are left.

’

Hawthorne then imagines a rational agent who is about to be thrust into

our universe without knowing anything speci

ﬁc about it. This agent does

know enough about the dependence of physical layout on physical law to
know about the various dependences of this or that structure on this or
that setting of constants. How much credence should the agent in question
give to the possibility that the universe was created by a God concerned to
allow consciousness to emerge?

Hawthorne answers as follows:

176

FROM MORALITY TO METAPHYSICS

Certainly, the piece of pie corresponding to that credence will be entirely
contained within the piece of pie corresponding to the credence in consciousness
friendly constants. But the key issue is the appropriate ratio between the former
piece and the latter piece. Suppose the proportion is rationally mandated to be very
low. Then, when the being learns about consciousness friendly constants, its
credence in a consciousness-obsessed manipulator will go up but will remain
very low.

There are two points we should note about Hawthorne

’s remarks. Firstly,

it is extremely hard to get a sense of how likely or unlikely it is that the
fundamental physical constants of the universe

‘could have been different’.

By contrast, if the argument of this book is correct, we can assert with
some con

ﬁdence that it is highly unlikely our belief-generating and belief-

evaluating capacities {m, n, o

. . . } would track the objective moral order in

a purposeless universe. Secondly, if there are a whole series of

‘explanatory

gaps

’ in the secular account, this will have a signiﬁcant cumulative effect

on the slice of pie that represents the rationally mandated credence of
purposive accounts of the universe.

In other words, someone who assigns a low initial credence to a

purposive account may, through a series of rounds of standard updating
after new evidence, be rationally compelled to assign a much higher
credence to purposive accounts of the universe. If he regards classical
theism as the most plausible of these purposive accounts, then he will
take these pieces of evidence to constitute a signi

ﬁcant cumulative apolo-

getic case. The claim here is not (as Anthony Flew once joked) that

‘ten

leaky buckets

’ somehow add up to a watertight argument for theism.

Rather, a series of explanatory gaps for secular accounts of the universe add
up to a case for giving theism signi

ﬁcantly greater credence. The argument

of this book is one part of such a cumulative argument for theism.

John Hawthorne,

‘Religious Knowledge’, p. 3 (at <www.philosophy.ox.ac.uk/mem-

bers/john_hawthorne>).

Antony Flew, God and Philosophy (London: Hutchinson, 1966), 63.

FROM GOODNESS TO GOD

177

Purpose Without Theism?

Axiarchism and Neoplatonism

8.1 Introducing Axiarchism

Theism (whether classical or not) is only one of the possible forms a
teleological explanation might take. This chapter will consider a much
more direct form of teleology, namely the axiarchic account advanced by
John Leslie (in Value and Existence, Universes, and Immortality Defended ) and
by Hugh Rice (in God and Goodness).

Rice

’s position is the most directly challenging to classical theism. He

argues that a more economical and elegant account of both the existence
of the universe and of many of its properties (including the human capacity
for moral knowledge) would explain these phenomena by the sheer fact
of their objective value, rather than through the mediating activity of a
personal God.

As Rice acknowledges, we have reason to accept classical theism if we

take objective value to be causally impotent. For if that is so, objective
value can only make a difference to any other state of affairs via intentional
action

—and the only agent would be a personal, non-‘abstract’ God.

As we saw in Chapter 7, Rowland Stout argues that teleological

explanations need to explain how the goodness of a goal leads to its
realization. That is to say, that the sheer goodness of an event or a state
of affairs cannot, in and of itself, explain its occurrence. Rice argues against
Stout

’s claim, in the following terms:

John Leslie, Value and Existence (Oxford: Blackwell, 1979), Universes (London: Routle-

dge, 1989), and Immortality Defended (Oxford: Blackwell, 2007); and Hugh Rice, God and
Goodness (Oxford: Oxford University Press, 2000). A form of non-theistic teleology is being
defended by Thomas Nagel in his forthcoming Mind and Cosmos: Why the Materialist Neo-
Darwinian Conception of Nature Is Almost Certainly False.

If one is really committed to the idea of objective value, of value which is not
constituted by facts about minds, it is quite unobvious why one should take such a
view.

Rice

’s argument is that a number of our ordinary commitments only make

sense if we take there to be a benevolent harmony between our beliefs and
reality. Our scienti

ﬁc enquiries commit us to believing in an orderly,

comprehensible world. They also commit us (as I argued in Chapter 2)
to the belief that our faculties of theoretical reasoning can track the truth.
Likewise, if we take our moral beliefs to be capable of attaining to truth,
we must believe that

‘objective value can make a difference to the way we

think.

’ If this is so, he argues,

it is not at all clear why it should not in principle be possible for [objective value] to
affect the way things are more generally.

Rice argues that his position has a number of explanatory advantages over
classical theism. One such advantage is economy. Why hypothesize a
‘mediating God’ who knows and brings about that which is objectively
valuable, when you could simply ascribe causal power to objective value?
Another advantage which Rice claims for his account is that it sidesteps the
issues considered at 7.3.1. Whereas theism has to wrestle with the rela-
tionship between an omnipotent God and a (perhaps distinct) source of
objective value, his

‘abstract conception of God’ has no such problem.

God simply is the objective, ultimate Good.

In this chapter, I will explain why I do not take axiarchism to provide a

satisfying form of explanation. I will then consider whether a non-
axiarchic Neoplatonism might offer an alternative to classical theism.

8.2 Axiarchism and the Existence of God

Axiarchism does not preclude the existence of a personal deity. John Leslie
is considerably more accommodating of a

‘mediating God’ than Rice,

arguing that theism is in fact made more plausible if we accept the principle
that objective value itself is creative:

Rice, God and Goodness, 50.

Rice, God and Goodness, 51.

PURPOSE WITHOUT THEISM

179

God is often held to be

‘the source of all explanations,’ but it strikes me as no

compliment to interpret this as meaning that he himself just happens to exist.
Absolute inexplicability is no valid ground for worship, although tradition may
tend to have treated it as one.

Leslie thinks his position also makes more sense of the necessity of God,
which is af

ﬁrmed by classical theism:

It would seem better for believers in God-conceived-as-a-person to adopt [A. C.]
Ewing

’s suggestion that God would have an existence ‘necessary not because there

would be any internal contradiction in denying it but because it was supremely
good that God should exist

’.

If we restrict ourselves to forms of theism which assume a benevolent
deity, there are four different forms a purposive explanation could take, in
terms of their attitude to axiarchic explanation and the existence of a
personal God. At this stage, we need to consider the potential of each of
to provide the explanation this book is seeking. I have mapped out the
four possibilities in Figure 3.

I take Hugh Rice to be wrong with respect to both questions. In 8.3,

I will focus most of my attention on the question of whether axiarchism is
plausible. I will argue that we do not have reason to accept the sheer
goodness of a state of affairs or event as an explanation of it. At 8.4, I will

AXIARCHIC EXPLANATION VALID

Goodness of world

as full explanation

(Rice)

God explained by

by goodness of

his existence (Ewing)

NO PERSONAL

PERSONAL

GOD

Self-subsistent

benevolent God

(classical theism)

Causal power of

goodness as brute fact

(non-axiarchic

Neoplatonism)

AXIARCHIC EXPLANATION INVALID

Figure 3 Axiarchic explanation and the existence of a personal god

Leslie, Universes, 168. Cf. A. C. Ewing, Value and Reality (London: Allen and Unwin,

1973).

180

FROM MORALITY TO METAPHYSICS

then go on to argue (against non-axiarchic Neoplatonism) that classical
theism provides the best non-axiarchic explanation.

8.2.1 Axiarchism and a personal God

Before I begin this critique of axiarchism, I want to note another point of
disagreement with Rice. Even if axiarchic explanations were satisfying,
I think Rice is far too quick to dismiss what he calls a

‘mediating God’ in

favour of his

‘abstract conception of God’ (in which the deity is straight-

forwardly identi

ﬁed with objective value). Although Rice talks of such a

God

‘bringing things about’ and ‘working through us’ (via our knowledge

of, and response to, objective value), his account lacks any room for the
kind of personal relationship with God envisioned in Christianity, Juda-
ism, or Islam. The kind of deity possible within an axiarchic framework
would, it is true, be different from the God to whom classical theism bears
witness: not only (pace Rice

’s argument above) will such a deity be

constrained by an independent moral order. On the axiarchic account,
any deity will not be self-subsistent, but will owe their being to its
objective value.

If one accepts the rest of Rice

’s account, it might be good that such a

(limited) deity should exist. After all, philosophers should beware of
treating the deity merely as an explanatory hypothesis. Even if what
Rice calls a

‘mediating God’ is not necessary to explain the existence of

the world, and even if such a deity would lack key attributes ascribed to
God by classical theism, the world might be a better place if there was such
a being. For communion

—the possibility of human beings entering into a

personal and loving relationship with the divine

—might itself have value.

The kind of communion humans could have with Rice

’s ‘abstract’ God

falls signi

ﬁcantly short of this. In short, the goodness of the deity might lie

more in their capacity for loving communion with human beings than in
their usefulness as an explanation for the rest of the universe.

8.3 Evaluating Axiarchism

The central issue between my own position and axiarchism concerns its
claim that

(1)

Y happened because it was good that it was so

PURPOSE WITHOUT THEISM

181

constitutes an adequate and complete explanation of Y. Here, it seems, we
have come to bedrock

—to a clash of intuitions about the most fundamen-

tal principles of theoretical reasoning. The mainstream view (with which
I concur) is that (1) does not explain Y.

I have already argued (at 1.3.2) that our fundamental principles of theo-

retical reasoning are discerned through a process of re

ﬂective equilibrium.

Therefore, it is a signi

ﬁcant prima facie objection to axiarchism that the

goodness of a state of affairs or event is generally taken to be a quality
which cries out for explanation, rather than an explanation in itself.

In 2.4, I considered the competing answers given by William Paley

(theism) and Nillson and Pelger (natural selection) to the following
question:

(2)

Why do humans and other animals have eyes which are so well

ﬁtted to

their needs and

ﬂourishing?

It did not occur to Paley, still less to Nilsson and Pelger, to say that
goodness of this state of affairs explained its occurrence. It is this very
feature

—the ﬁttingness of the eye for the needs and ﬂourishing of its

possessors

—which all three take to require explanation. This suggests that

the canons of Inference to the Best Explanation which come naturally to
most reasoners (and which, I have argued alongside Enoch, must be taken
to be prima facie authoritative) do not regard (1) as a good explanation.

Another thought-example might reinforce the point. Let us suppose a

volcano on a particular island emits highly toxic fumes every time it
erupted, and that the lava

ﬂow always runs away from islanders’ homes.

Let us also suppose the winds blow the fumes away within a day. So long as
islanders are indoors when the volcano erupts, and stay indoors for a day,
they will be safe

—but even a short trip out of doors will be fatal.

Now let us suppose that, an hour before each eruption, the clouds

above the island spell the words

‘STAY INDOORS’ in the local language.

Would we take that phenomenon to require explanation, or to be ex-
plained by its goodness

—and its contribution to the goodness of the

islanders

’ wellbeing?

Perhaps that example seems too loaded. One might think that the

linguistic element implies a person communicating, and begs the question
a little too much in favour of theism. So what if we simply say the islanders
get a strong sense of foreboding, which moves them to stay indoors, an
hour before each eruption?

182

FROM MORALITY TO METAPHYSICS

In both cases, it seems quite wrong to suggest the goodness of the

warning is itself an explanation. Rather, it seems that the warnings cry
out for explanation. When one examines each case more deeply, non-
theistic explanations for these phenomena may be found. In the

ﬁrst case,

perhaps there is a scienti

ﬁc reason why a certain cloud formation on the

island always precedes the volcano

—and perhaps it is this that has led the

islanders to choose that shape for the words

‘STAY INDOORS’ in their

script. In the second case, there may be a characteristic change in the
environment on the island which always precedes the volcano, and it is this
which the islanders note subconsciously, and which generates a feeling of
foreboding. Either of these would count as a perfectly good explanation of
the phenomena in question

——and in the absence of either, one might

expect a greater preponderance of theists on the island (and perhaps a
conviction that they were a people specially blessed by God).

Each of these examples reveals the counter-intuitive nature of axiar-

chism. Very clearly, it is not part of the current practice of Inference to the
Best Explanation (IBE) to take the goodness of the

‘STAY INDOORS’

message or the sense of foreboding to offer an alternative explanation.

Leslie suggests this is something of a metaphysical prejudice, and defends

his

‘Neoplatonic’ axiarchism as follows:

It is felt that [the advocate of axiarchism] would have to identify some mechanism
whereby ethical needs, requirements, grounds for the existence of this or that,
produced their supposed effects. But this misses Neoplatonism

’s central point.

Neoplatonism is the view that some ethical needs are themselves creatively effective,
unaided by any mechanism. You might just as well ask for a mechanism which
made misery intrinsically evil or a mechanism ensuring that two and two made
four. Either some consistent set of ethical needs are creatively effective, unaided by
any act of divine will, or else they aren

’t.

[All italics in this and subsequent

quotations are Leslie

’s own.]

In terms of Figure 3, Leslie is making clear that

‘Neoplatonism’ is in the

top half of the diagram, not in the bottom left quadrant. This is an
important distinction: for there is a conceptual space for a non-personal
conception of God which does not claim the goodness of states of affairs
itself explains them, but rather that a divine being (of a different kind from
that witnessed to by the Abrahamic faiths) brings them about. However,
Leslie and Rice want to argue

—in contrast to this position and to classical

Leslie, Universes, 171.

PURPOSE WITHOUT THEISM

183

theism

—that the sheer goodness of a state of affairs does explain it. There is

no need for any further

‘mechanism’ of the kind these non-axiarchic

accounts would offer.

Leslie defends this position as follows:

Either some consistent sets of ethical needs are creatively effective, unaided by any
act of divine will, or else they aren

’t. Now, these alternatives are equally simple.

Why? Well if ethical needs, alias requirements, are never themselves creatively
effective, then presumably this isn

’t just a matter of chance. But it isn’t a matter

of logic either, like the fact that bachelors are never married: there is no logical
contradiction in supposing that some ethical requirement for the existence of such-
and-such is actually able to produce its existence.

Because axiarchism does not involve any logical contradiction, Leslie
suggests that the issue between it and rival views cannot be settled by
conceptual analysis:

‘The matter is not a logical, conceptual matter. But

the actual existence of a universe (and of one, what

’s more, which seems

ﬁne tuned for Life) can suggest its rightness.’

Both here and in Rice

’s argument for axiarchism two issues are repeat-

edly elided. The

ﬁrst issue is whether there is evidence of a harmony

between (i) the way the universe is and (ii) certain objective ethical values.
In Leslie

’s case, the harmony in question is that which obtains between the

universe

’s fundamental constants and the goodness of their being human

life (cf. my discussion of John Hawthorne in 7.5), whereas as we saw
above, Rice focuses on the harmony we have to suppose exists between
our cognitive faculties and the universe

’s order—both its moral order (the

theme of this book) and its more general intelligibility and rule-governed
nature.

However, as the thought-examples above are designed to emphasise,

from the fact that there are certain harmonies (or pace Rice and the
argument of this book that we have to suppose such harmonies to exist)
nothing at all follows about what we should count as a good explanation
of said harmonies. The question of what a good explanation of these
harmonies would be is a second, and quite distinct, one.

As I argued at 1.3, principles of Inference to the Best Explanation (IBE)

are not

‘read off ’ the phenomena for which we seek to account. Rather,

the principles are of necessity a priori, and we come to know them

Leslie, Universes, 171.

184

FROM MORALITY TO METAPHYSICS

through the re

ﬂective equilibrium process. They tell us what sorts of things

to count as good explanations for different sorts of phenomena.

The fact that our universe appears to be

ﬁne-tuned, or the fact that we

must suppose a harmony between our cognitive capacities and the world,
does not constitute evidence for axiarchism, or against the well-established
deliverances of the re

ﬂective equilibrium process with respect to IBE. The

canons of explanation which the vast majority of scientists and philoso-
phers seem to follow determine that the sheer goodness of states of affairs
or events (such as the functioning of the eye, the islanders

’ sense of

foreboding, the

ﬁne-tuning of the universe, and the harmony between

our cognitive capacities and wider physical and moral realities) does not
offer an explanation of those occurrences. Rather, the occurrence of states
of affairs or events which systematically promote human

ﬂourishing and

other kinds of objective goodness itself cries out for explanation.

The

‘reﬂective equilibrium’ process is precisely that—a process which is

responsive to new arguments, or changes in the wider landscape of
metaphysical and epistemological orthodoxies. If Leslie is to secure a
more sympathetic hearing for axiarchism, much more will need to be
offered in its defence of the position than the observation that it is

‘just as

simple

’ as the alternatives, or the (rather question-begging) assertion that it

renders otherwise inexplicable harmonies intelligible.

There a one further point to be made, which is friendlier to axiarchism.

As the

‘reﬂective equilibrium’ on any issue is inﬂuenced by the wider

philosophical landscape, and the accounts of metaphysics that seem plau-
sible, might it be that Leslie

’s Neoplatonism only seems counter-intuitive

today because of the dominance of subjectivist accounts of ethics, and
secular accounts of the universe? As these are both accounts which this
book has argued against, might this not point towards a more generous
assessment of axiarchism?

Parts I and II of this book seek to remove some fallacious arguments

against axiarchism. This is made clear if we consider J. L. Mackie

’s discussion

of axiarchism in The Miracle of Theism. For Mackie, the key objection to the
ascription of a creative role of ethical requirements is precisely that he takes
moral objectivism to be false.

If the argument of Chapter 1 is successful,

J. L. Mackie, The Miracle of Theism: Arguments For and Against the Existence of God (Oxford:

Oxford University Press, 1982), ch. 13.

PURPOSE WITHOUT THEISM

185

his objection to axiarchism is removed. However, serious inadequacies in
the position remain.

From the fact that the moral order has the kind of

‘queerness’ Mackie

rejects, it in no way follows that it can have other, even more metaphysically
perplexing, qualities. Leslie may be right to say there is no logical contradic-
tion involved in his suggestion that the objective moral order has the capacity
in and of itself to bring about states of affairs and events. This does not mean
the suggestion is metaphysically unproblematic. It surely counts against
axiarchism that we have no reason to believe any other kind of objective
imperatives have an impact on the external world in this direct way.

As I argued in Chapter 1, we have experience of imperatives in theo-

retical and practical reasoning having an impact on the world via agents
who have a (separately explained) capacity both for cognition and for
action. To apprehend an imperative

—in theoretical or moral reasoning—

is to have a ground for action. In that sense, moral norms (like theoretical
ones) can feature in explanations of states of affairs and events.

The idea

that they might feature more directly, through their own intrinsic capacity
to affect the physical world, is much more problematic.

Although it is not a logical contradiction to say that a moral or theoreti-

cal imperative could create a universe, it does seem to be a category mistake.
In that sense, the

ﬂaw in axiarchism is one which we uncover via concep-

tual analysis, although the method we must use for determining whether
the analysis is correct is re

ﬂective equilibrium thinking. Using that meth-

od, we seem well-justi

ﬁed in asserting that norms are not the kind of thing

that can have a direct causal impact on the world; a norm of reasoning
cannot (in Leslie

’s words) ‘act creatively’.

The only things that can act

purposively are agents. It is only through such agents

’ cognition of norms,

and the powers those agents have to act in the world, that we have any
reason to think objective value can have a causal impact on the universe.

8.4 Non-Axiarchic Neoplatonism

The Figure 3 indicated two possible forms of purposive explanation
that were not axiarchic. One was classical theism

—whose ability to

Cf. Nicholas Sturgeon,

‘Moral Explanations’, in Geofrey Sayre-McCord (ed.), Essays on

Moral Realism (Cornell University Press, 1988).

Leslie uses this turn of phrase in his Universes, 173.

186

FROM MORALITY TO METAPHYSICS

explaining the human capacity for moral cognition was set out in
Chapter 7. The other possibility was termed

‘non-axiarchic Neoplato-

nism

’. On this account, the claim is not that our canons of IBE render

(1)

Y happened because it was good that it was so

an adequate and complete explanation of Y. Its more limited claim is that
there is a force or power that brings about good things

—and that the

question

‘Why does such a force or power exist?’, while not a foolish or

redundant one, has no further answer. The distinction between this and
Rice and Leslie

’s axiarchism is subtle but important. Whereas Rice and

Leslie think objective value might in and of itself be creative, and that the
goodness of states of affairs might completely explain them, this non-
axiarchic Neoplatonism posits a creative force that brings about states of
affairs which are good. One might call this force an

‘abstract’, non-personal

deity, worthy of gratitude and even worship. The demand for an explana-
tion of the existence of this deity would not be redundant, because (unlike
Rice and Leslie) this account does not claim that all

‘why?’ questions have

been answered. However, its defender might argue, this constitutes a
legitimate terminus for explanation because, as I argued at 6.4.3, the chain
of explanation must stop somewhere. In that discussion, I suggested as a
plausible principle that

(3)

All phenomena either require an explanation or an account of why it would
be a mistake to seek a further level of explanation

Classical theism clearly ful

ﬁls (3). Even if it is not foolish to ask ‘Why is

there a loving divine being?

’ then, if we accept the arguments above

against axiarchism, we can see that the question could only be answered
as follows:

(4)

God is eternal, omnipotent, and self-subsistent, and hence could not be
caused by anything else. The alternative, a lesser god created by a greater
one, or perhaps a whole series of more and more powerful creative gods, is
obviously far more complex and this complexity comes with no additional
explanatory power.

Does non-axiarchic Neoplatonism also ful

ﬁl (3)? I think it does, but that it

offers a less satisfying explanation than classical theism. As we saw in
Chapter 7 (and particularly at 7.2.2), agent explanation renders it intelligi-
ble why good states of affairs are brought about by God (in a way that

PURPOSE WITHOUT THEISM

187

Stephen Law

’s ‘Evil God Hypothesis’ did not render it intelligible why bad

states of affairs would be brought about by a malevolent deity). It is surely
uncontroversial to claim that, ceteris paribus, explanations are better when
they render the explanandum more intelligible. On that basis, classical
theism has a good claim to be a better explanation of our cognition of
moral truths than this modi

ﬁed form of Neoplatonism.

8.5 Conclusion

I have been at pains to stress the limited scope of the argument being made
in this book. Part III has not advanced a general argument for classical
theism, but simply argued that it is the most satisfying explanation of the
human capacity for moral cognition. It has done so by showing,

ﬁrstly that

classical theism can explain the phenomenon, and then arguing that all
other forms of purposive explanation are much less satisfactory.

If the argument of 8.3 is correct, it counts against axiarchism in

general

—not simply as an explanation of the human capacity for moral

cognition. The position seems to have some more fundamental

ﬂaws,

which may explain its relative obscurity. It is none the less important to
engage with it, rather than dismissing it on the grounds that it is unfash-
ionable. If that were a ground for dismissal, the argument of this book
would have very little chance of success.

The argument I have made against non-axiarchic Neoplatonism is less

general in its scope. Although I have shown the position to provide a less
satisfying explanation of the human capacity for moral cognition, there
might be other grounds on which to favour it above classical theism. I will
argue below, these more detailed disagreements might not be resolved by
philosophical argument alone. Such knowledge as humans can have of the
nature of their Creator may

ﬂow from revelation as well as reason.

188

FROM MORALITY TO METAPHYSICS

Conclusion

Many people are understandably wary of religious philosophers. To those
who have a faith, their arguments can seem like hubristic attempts to know
the mind of God. To non-believers, they can seem rather contrived
attempts to justify a faith whose real motivation lies elsewhere.

This book has not sought to advance a general argument for the

existence of God. Its more limited aim has been to identify a serious
weakness in secular accounts of the nature of morality and the human
capacity for moral cognition. I want to

ﬁnish with a few remarks about the

place of this argument in the life of faith.

In his encyclical on faith and reason, Pope John Paul II wrote of the

legitimacy of

‘a philosophy completely independent of the Gospel’s Rev-

elation

’. He wrote that this was

the stance adopted by philosophy as it took shape in history before the birth of the
Redeemer and later in regions as yet untouched by the Gospel. We see here
philosophy

’s valid aspiration to be an autonomous enterprise, obeying its own rules

and employing the powers of reason alone. Although seriously handicapped by the
inherent weakness of human reason, this aspiration should be supported and
strengthened. As a search for truth within the natural order, the enterprise of
philosophy is always open

—at least implicitly—to the supernatural.

This book is written out of my conviction that classical theism

—and, more

speci

ﬁcally, Christianity—is open to apologetic arguments which take this

form. I have sought to make a contribution to that enterprise by entering
into the

‘search for truth within the natural order’, and arguing that even

when engaged with on its own terms, it points beyond itself. There is a gap
in what secular philosophy can explain. This is not the kind of

‘gap’ that

will be

ﬁlled by a new scientiﬁc discovery. Our most fundamental ethical

Pope John Paul II, Faith and Reason

—Fides et Ratio (Catholic Truth Society, 1998), 109.

commitments point towards a

‘supernatural’ source for our knowledge of

what is good and right.

The implications of these arguments for individual readers will depend

on many other factors. Some will doubt that classical theism is coherent

—

arguing that there is some incoherence in the cluster of properties it
ascribes to the deity. Others will take the sheer quantity and intensity of
suffering in the world to be a decisive objection to belief in a benevolent
God. Those issues lie beyond the scope of this book, which simply argues
that there is at least one central feature of human life which classical theism
is uniquely well-equipped to explain. Combined with other apologetic
arguments (such as the need for an explanation of the consciousness
friendly nature of the universe

’s fundamental physical constants) a signiﬁ-

cant cumulative case may emerge.

Even if this cumulative case is intellectually compelling, it does not

follow that such apologetic arguments are the cause of religious belief. (In
that sense, the non-believers

’ scepticism about religious philosophy has a

point

—but not one that undermines the practice.) The journey of faith is

one of

‘heart speaking to heart’.

But philosophy has a signi

ﬁcant part to

play

—in helping us respond to the important and legitimate worry that

this journey may be an exercise in wish ful

ﬁlment rather than a response to

a genuine reality. Philosophy can create the intellectual space for an
encounter of the heart.

What of the religious believer

’s worry about religious philosophy? Is it a

hubristic attempt to know the mind of God? The case made in this book is
part of a family of apologetic arguments which engage in the

‘wholly

autonomous enquiry

’ commended by John Paul II. They suggest that,

unless our thought is open to the supernatural, there are a number of
correlations which are, by its own lights, inexplicable. To say this is not to
suggest that philosophy can enable us to know the mind of God. Instead,
these arguments remind us of our dependence upon the Creator; they call
us to humility rather than hubris. Our ability to discern and love objective
goodness is a gift

—bestowed on us by the One from Whom all good

things come.

John Henry Newman

’s motto (Cor ad Cor Loquitur) was chosen by Benedict XVI as the

theme for his Papal Visit to the United Kingdom in 2010.

190

F R O M M O R A L I T Y T O M E T A P H Y S I C S

Bibliography

Adams, Robert M.,

‘Motive Utilitarianism’, Journal of Philosophy 73 (1976).

——, The Virtue of Faith and Other Essays in Philosophical Theology (Oxford: Oxford

University Press,

1987).

——, Finite and Inﬁnite Goods: A Framework for Ethics (Oxford University Press,

1999).

Anscombe, G. E. M.,

‘Modern Moral Philosophy’, Philosophy 33 (1958).

——, Ethics, Religion and Politics: The Collected Philosophical Papers of

G. E. M. Anscombe (Oxford: Blackwell,

1981).

Aristotle, On the Parts of Animals, trans. J. G. Lennox (Oxford: Oxford University

Press,

2001).

Audi, Robert,

‘Moderate Intuitionism and the Epistemology of Moral Judgment’,

Ethical Theory and Moral Practice

1(1) (1998): 15–44.

Ayer, A. J., Language, Truth and Logic (Penguin Classics,

2001).

Blackburn, Simon, Spreading the Word (Oxford: Oxford University Press,

1984).

——, Essays in Quasi-Realism (Oxford: Oxford University Press, 1993).
——, Ruling Passions (Oxford: Oxford University Press, 1998).
——, ‘Am I Right?’, The New York Times, 21 February 1999.
Brink, David O., Moral Realism and the Foundations of Ethics (Cambridge: Cambridge

University Press,

1989).

Carr, B. J. and Rees, M.,

‘The Anthropic Cosmological Principle and the Structure

of the Physical World

’, Nature 278 (12 April 1979).

Carroll, Lewis,

‘What the Tortoise said to Achilles’, Mind 4 (1895).

Child, William,

‘Pears’ Wittgenstein: Rule-Following, Platonism and Naturalism’,

in David Charles and William Child (eds), Wittgensteinian Themes: Essays in
Honour of David Pears (Oxford: Oxford University Press,

2001).

Collins, R.,

‘Evidence for Fine-Tuning’, in Neil A. Manson (ed.) God and Design:

The Teleological Argument and Modern Science (London: Routledge,

2003).

Crisp, Roger,

‘Hedonism Reconsidered’ (delivered to the Minnesota Interdisci-

plinary Workshop on Well-Being,

24–6 October 2003; online at <http://www.

utilitarianism. com/hedonism-reconsidered.html>).

——, ‘Value, Reasons and the Structure of Justiﬁcation: How to Stop Passing the

Buck

’, Analysis 65(285) (2005).

——, Reasons and the Good (Oxford University Press, 2006).
Dawkins, Richard,

‘Where D’you Get Those Peepers?’, New Statesman and Society,

16 June 1995.

Dawkins, Richard, The God Delusion (New York: Random House,

2006).

Dennett, Daniel C.,

‘The Scope of Natural Selection’, Boston Review (October/

November

1996).

Dworkin, Ronald,

‘Objectivity and Truth: You’d Better Believe It’, Philosophy and

Public Affairs

25(2) (1996).

Elster, Jon, Sour Grapes: Studies in the Subversion of Rationality (Cambridge: Cam-

bridge University Press,

1983).

——, Making Sense of Marx (Cambridge University Press, 1985).
Enoch, David,

‘How is Moral Disagreement a Problem for Realism?’, Journal of

Ethics

13(1) (2009).

——, ‘An Outline of an Argument for Robust Metanormative Realism’, in Russ

Shafer-Landau (ed.), Oxford Studies in Metaethics, Vol. II (Oxford: Oxford
University Press,

2007).

Ewing, A. C., Value and Reality: The Philosophical Case for Theism (London: Allen

and Unwin,

1973).

Flew, Anthony, God and Philosophy (Hutchinson,

1966).

Foot, Philippa,

‘Utilitarianism and the Virtues’, Mind 94(374) (1985).

——, ‘Morality as a System of Hypothetical Imperatives’, in Philippa Foot, Virtues

and Vices (Oxford: Oxford University Press,

2002).

——, Natural Goodness (Oxford: Oxford University Press, 2003).
Foucault, Michel, Power/Knowledge: Selected Interviews and Other Writings

1972–1977,

ed. and trans. Colin Gordon (London: Longman,

1980).

Gaita, Raymond, Good and Evil: An Absolute Conception (London: Routledge,

2004).

——, webcast discussion on torture at <http://philosophybites.com/applie-

d_ethics/page/

2/>

Geach, P. T.,

‘Assertion’, Philosophical Review 74(4) (1965).

——, The Virtues (Cambridge University Press, 1977).
Gibbard, Allan,

‘Critical Notice of Essays in Quasi-Realism’, Mind 105(418) (1996).

——, Thinking How to Live (Cambridge, MA: Harvard University Press, 2003).
Gollwitzer, H., Kuhn, K., and Schneider, R. (eds), Dying We Live: The Final

Messages and Records of Some Germans who De

ﬁed Hitler (London: Collins, 1960).

Hare, John, The Moral Gap: Kantian Ethics, Human Limits and God

’s Assistance

(Oxford: Oxford University Press,

1997).

——, God and Morality: A Philosophical History (Oxford: Blackwell, 2007).
Hare, R. M., The Language of Morals (Oxford: Clarendon Press,

1952).

Harman, Gilbert,

‘The Inference to the Best Explanation’, The Philosophical Review

74 (1965).

Hawthorne, John,

‘Practical Realism?’, Philosophy and Phenomenological Research 64

(

2002).

192

B I B L I O G R A P H Y

——, ‘Religious Knowledge’, at <www.philosophy.ox.ac.uk/members/john_

hawthorne>

Hempel, Carl,

‘Explanation in Science and in History’, in David-Hillel Ruben

(ed.), Explanation (Oxford: Oxford University Press,

1993).

Hume, David, Enquiry Concerning Human Understanding, in his Enquiries, ed.

L. A. Selby-Bigge,

3rd edn (Oxford: Oxford University Press, 1975).

——, An Enquiry Concerning the Principles of Morals (Oxford: Oxford University

Press,

1998).

Hursthouse, Rosalind, Lawrence, Gavin, and Quinn, Warren (eds), Virtues and

Reasons: Philippa Foot and Moral Theory (Oxford: Clarendon Press,

1996).

Jackson, Frank, From Metaphysics to Ethics: A Defence of Conceptual Analysis (Oxford:

Oxford University Press,

2000).

Kitcher, Philip, In Mendel

’s Mirror: Philosophical Reﬂections on Biology (Oxford:

Oxford University Press,

2003).

Korsgaard, Christine M., The Sources of Normativity, ed. Onora O

’Neill with

responses by G. A. Cohen, Raymond Geuss, Thomas Nagel, and Bernard
Williams (Cambridge: Cambridge University Press,

1996).

Kymlicka, Will, Multicultural Odysseys: Navigating the New International Politics of

Diversity (Oxford: Oxford University Press).

John Paul II, Faith and Reason

—Fides et Ratio (London: Catholic Truth Society,

1998).

Law, Stephen,

‘The Evil God Challenge’, Religious Studies 46 (2010): 353–73.

Leslie, John, Value and Existence (Oxford: Blackwell,

1979).

——, Universes (London: Routledge, 1989).
——, Immortality Defended (Oxford: Blackwell, 2007).
Lindgaard, Jakob (ed.), John McDowell: Experience, Norm and Nature (Oxford:

Blackwell,

2008).

Lovibond, Sabina, Realism and Imagination in Ethics (Oxford: Blackwell,

1983).

Lucas, J. R.,

‘The Lesbian Rule’, Philosophy, 30 (1955).

Lumsden, Charles and Wilson, E. O., Promethean Fire (Cambridge, MA: Harvard

University Press,

1983).

Mackie, J. L., Ethics: Inventing Right and Wrong (Harmondsworth: Penguin Books,

1977).

——, The Miracle of Theism: Arguments For and Against the Existence of God (Oxford:

Oxford University Press,

1982).

Mawson, T. J.,

‘God’s Creation of Morality’, Religious Studies 38 (2002).

McDowell, John,

‘Wittgenstein on Following a Rule’, Synthese 54 (1984).

——, Mind and World (Cambridge, MA: Harvard University Press, 1994).
——, Mind, Value and Reality (Cambridge, MA: Harvard University Press, 1998).

B I B L I O G R A P H Y

193

McDowell, John, The Engaged Intellect: Philosophical Essays (Cambridge, MA:

Harvard University Press,

2009).

Meyerson, Denise, False Consciousness (Oxford: Clarendon Press,

1991).

Nagel, Thomas, Mind and Cosmos: Why the Materialist Neo-Darwinian Conception of

Nature Is Almost Certainly False (Oxford: Oxford University Press, forthcoming).

Neurath, Otto, Empiricism and Sociology, ed. Marie Neurath and Robert S. Cohen

(Dordrecht: Reidel,

1973).

Nilsson, Dan-Erik and Pelger, Susanne,

‘A Pessimistic Estimate of the Time

Required for an Eye to Evolve

’, Proceedings of the Royal Society of London,

Biological Sciences

256 (1994).

Nozick, Robert, Anarchy, State, and Utopia (Oxford: Blackwell,

1974).

Nussbaum, Martha, Love

’s Knowledge: Essays on Philosophy and Literature (Oxford:

Oxford University Press,

1992).

——, Upheavals of Thought: The Intelligence of the Emotions (Cambridge: Cambridge

University Press,

2001).

Paley, William, Natural theology: Or, Evidence of the Existence and Attributes of the

Deity, Collected from the Appearances of Nature, Oxford World

’s Classics edn, ed.

Matthew Eddy and David M. Knight (Oxford: Oxford University Press,

2006).

Pears, David, The False Prison (Oxford: Clarendon Press,

1988).

Plantinga, Alvin, Warrant and Proper Function (Oxford: Oxford University Press,

1993).

Putnam, Hilary, Reason, Truth and History (Cambridge: Cambridge University

Press,

1981).

Rasmussen, Stig Alstrup,

‘Quasi-Realism and Mind-Dependence’, Philosophical

Quarterly

35(139) (1985).

Rawls, John, A Theory of Justice (Cambridge, MA: Belknap Press,

1971).

Rice, Hugh, God and Goodness (Oxford: Oxford University Press,

2000).

Ritchie, Angus W.,

‘Realism, Ideology, Truth: An Examination of Richard

Rorty

’s Critique of Metaphysical Realism’ (Unpublished B.Phil. Thesis, 1996).

Rommen, Heinrich, Die ewige Wiederkehr des Naturrechts (Hegner,

1936); in English

as The Natural Law: A Study in Legal and Social History and Philosophy, trans.
Thomas R. Hanley (Indianapolis: Liberty Fund,

1998).

Rorty, Richard, Philosophy and the Mirror of Nature (Oxford: Blackwell,

1980).

Sayre-McCord, Geoffrey, Essays on Moral Realism (Ithaca, NY: Cornell University

Press,

1988).

Scanlon, T. M.,

‘Moral Theory: Understanding and Disagreement’, Philosophy and

Phenomenological Research

55(2) (1995).

——, What We Owe to Each Other (Cambridge, MA: Belknap Harvard, 1998).
——, The Difﬁculty of Tolerance: Essays in Political Philosophy (Cambridge: Cam-

bridge University Press,

2003).

——, The John Locke Lectures 2009: Being Realistic About Reasons (online at <http://

www.philosophy.ox.ac.uk/lectures/john_locke_lectures>).

Searle, John R.,

‘Meaning and Speech Acts’, Philosophical Review 71(4) (1962).

194

B I B L I O G R A P H Y

Segerstråle, Ullica, Defenders of the Truth: The Sociobiology Debate (Oxford: Oxford

University Press,

2000).

Sen, Amartya, Development as Freedom (Oxford: Oxford University Press,

1999).

Singer, Peter, The Expanding Circle: Ethics and Sociobiology (New York: Farrar,

Straus and Giroux,

1981).

Stout, Rowland, Things That Happen Because They Should: A Teleological Approach to

Action (Oxford: Oxford University Press,

1996).

Street, Sharon,

‘A Darwinian Dilemma for Realist Theories of Value,’ Philosophical

Studies

127, (1) (2006).

Suikkanen, Jussi,

‘Buck-Passing Accounts of Value’, Philosophy Compass 4(5) (2009).

Swinburne, R. G.,

‘Duty and the Will of God’, Canadian Journal of Philosophy

4 (1974).

——, Responsibility and Atonement (Oxford: Clarendon Press, 1989).
——, The Coherence of Theism (Oxford: Clarendon Press, 1993).
Taylor, Charles, The Explanation of Behaviour (London: Routledge & Kegan Paul,

1964).

Thompson, Michael,

‘Apprehending Human Form’, in Anthony O’Hear (ed.),

Modern Moral Philosophy: Royal Institute of Philosophy Supplement

54 (Cambridge:

Cambridge University Press,

2004).

Voorhoeve, Alex,

‘The Grammar of Goodness: An Interview with Philippa Foot’,

The Harvard Review of Philosophy

11 (2003).

Walker, Ralph C. S., The Coherence Theory of Truth: Realism, Anti-Realism, Idealism